From bugzilla-daemon at llvm.org  Thu Apr  1 07:35:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 07:35:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6755] New: CMake configured make install misses
 DiagnosticCommonKinds.inc
Message-ID: <bug-6755-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6755

           Summary: CMake configured make install misses
                    DiagnosticCommonKinds.inc
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: vr086731 at studenti.univr.it
                CC: llvmbugs at cs.uiuc.edu


I am using this rep:

  http://llvm.org/svn/llvm-project/llvm/branches/release_27

and I have configured the build with cmake.

The problem is that the file

  clang/Basic/DiagnosticCommonKinds.inc

is not copied by 'make install'.
This file is needed to compile applications that include clang headers, which
otherwise fail with the message like the following:

include/clang/Basic/Diagnostic.h:63:49: error:
clang/Basic/DiagnosticCommonKinds.inc: No such file or directory

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 11:55:27 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 11:55:27 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6755] CMake configured make install misses
 DiagnosticCommonKinds.inc
In-Reply-To: <bug-6755-206@http.llvm.org/bugs/>
References: <bug-6755-206@http.llvm.org/bugs/>
Message-ID: <20100401165527.853F72A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6755

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-01 11:55:27 CDT ---
krj fixed this in r 100105

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 12:09:58 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 12:09:58 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6756] New: exception object not destructed
Message-ID: <bug-6756-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6756

           Summary: exception object not destructed
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: hhinnant at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


This test should pass:

#include <stdio.h>
#include <assert.h>

struct A
{
  static int constructed;
  int data_;

  explicit A(int data = 0) : data_(data)
  {
      ++constructed;
  }

  ~A()
  {
      --constructed;
  }

  A(const A& a)
      : data_(a.data_)
  {
      ++constructed;
  }

private:

  A& operator=(const A&);
};

int A::constructed = 0;


int main()
{
  try
  {
      throw A(5);
      assert(false);
  }
  catch (A& a)
  {
      assert(a.data_ == 5);
      assert(A::constructed == 1);
  }
  printf("J: %d\n", A::constructed);
  assert(A::constructed == 0);
}

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 12:54:54 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 12:54:54 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6757] New: ternary operator: incompatible operand
 types ('Foo const' and 'Foo')
Message-ID: <bug-6757-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6757

           Summary: ternary operator: incompatible operand types ('Foo
                    const' and 'Foo')
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: evan at chromium.org
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4621)
 --> (http://llvm.org/bugs/attachment.cgi?id=4621)
test case as found in bug

This came up in WebCore.
Simplified test case follows, will also attach.

I believe this may have worked in the past, which would make this a regression.
 But it's also possible the code in WebCore changed since the last time I
compiled this area.


build with 
clang++ -c test.cc
test.cc:12:16: error: incompatible operand types ('Foo const' and 'Foo')
  Foo x = true ? Bar() : Foo();
               ^ ~~~~~   ~~~~~


struct Foo {
};

struct Bar {
  operator const Foo&() const;
};

void f() {
  Foo x = true ? Bar() : Foo();
}

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 13:33:14 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 13:33:14 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6563] accepts invalid template argument
In-Reply-To: <bug-6563-206@http.llvm.org/bugs/>
References: <bug-6563-206@http.llvm.org/bugs/>
Message-ID: <20100401183314.5F5382A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6563

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-01 13:33:14 CDT ---
Fixed in r100125.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 13:33:23 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 13:33:23 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6749] clang allows references to fulfill reference
 non-type parameters
In-Reply-To: <bug-6749-206@http.llvm.org/bugs/>
References: <bug-6749-206@http.llvm.org/bugs/>
Message-ID: <20100401183323.67C892A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6749

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-01 13:33:23 CDT ---
Fixed in r100125.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 15:42:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 15:42:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6758] New: JIT doesn't work on PowerPC: relocation
	fails
Message-ID: <bug-6758-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6758

           Summary: JIT doesn't work on PowerPC: relocation fails
           Product: new-bugs
           Version: unspecified
          Platform: Macintosh
               URL: https://wwws.clamav.net/bugzilla/show_bug.cgi?id=1921
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


On Darwin 9.8.0 Power Macintosh powerpc:

The JIT aborts with this error:
Assertion failed: (ResultPtr >= -(1 << 23) && ResultPtr < (1 << 23) &&
"Relocation out of range!"), function relocate, file
llvm/lib/Target/PowerPC/PPCJITInfo.cpp, line 399.

backtrace:
Program received signal SIGABRT, Aborted.
0x96191b50 in __kill ()
(gdb) bt
#0  0x96191b50 in __kill ()
#1  0x9622cc00 in abort ()
#2  0x9621fc10 in __assert_rtn ()
#3  0x012bf304 in llvm::PPCJITInfo::relocate ()
#4  0x00e0c904 in (anonymous namespace)::JITEmitter::finishFunction ()
#5  0x01286634 in (anonymous namespace)::PPCCodeEmitter::runOnMachineFunction
()
#6  0x00fd61b8 in llvm::FPPassManager::runOnFunction ()
#7  0x00fd63e8 in llvm::FunctionPassManagerImpl::run ()
#8  0x00fd66ac in llvm::FunctionPassManager::run ()
#9  0x00dfc590 in llvm::JIT::runJITOnFunctionUnlocked ()
#10 0x00dfd670 in llvm::JIT::getPointerToFunction ()


This happens when the JIT client (ClamAV) has allocated some memory, and then
tries to use the JIT, it doesn't happen when only trying to use the JIT.

If needed I can probably write up a testcase for LLI that allocates some memory
first, and then tries to JIT, but I think the abort message is pretty clear
about what the problem is.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 16:04:08 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 16:04:08 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6759] New: "friend class" implies forward
	declaration "class"?
Message-ID: <bug-6759-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6759

           Summary: "friend class" implies forward declaration "class"?
           Product: clang
           Version: trunk
          Platform: Macintosh
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: jdd at efn.org
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


The following example will compile in g++ but not in clang++:

#include <stdio.h>

/* If you uncomment the forward declaration, then this compiles OK
   (but g++ does not require the forward declaration):
*/
// class FooImpl;

class Foo_connection {
    private:
        friend class FooImpl;
        Foo_connection(FooImpl&);
};


int main() {
  int i = 1;
  printf("%d \n",++i);
}

-----
clang++     testcase.cc   -o testcase
testcase.cc:9:31: error: expected ')'
        Foo_connection(FooImpl&);
                              ^

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 16:48:20 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 16:48:20 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6759] "friend class" implies forward declaration
	"class"?
In-Reply-To: <bug-6759-206@http.llvm.org/bugs/>
References: <bug-6759-206@http.llvm.org/bugs/>
Message-ID: <20100401214820.33D2E2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6759

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-01 16:48:19 CDT ---
Clang is correct here, per C+ [11.4]p9:

  For a friend class declaration, if there is no prior declaration, the class
that is specified belongs to the innermost enclosing non-class scope, but if it
is subsequently referenced, its name is not found by name lookup until a
matching declaration is provided in the innermost enclosing nonclass scope.

GCC is wrong to make FooImpl visible in the friend class declaration.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 17:21:58 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 17:21:58 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6760] New: The -globalopt pass assertions on the
	given code
Message-ID: <bug-6760-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6760

           Summary: The -globalopt pass assertions on the given code
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: Interprocedural Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: criswell at uiuc.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4622)
 --> (http://llvm.org/bugs/attachment.cgi?id=4622)
Reduced test case from 164.gzip

The Global Variable optimization pass asserts when run on the attached LLVM
bitcode.  It occurs when using the release_27 branch.

opt -f -o /dev/null -globalopt bugpoint-reduced-simplified.bc 
Assertion failed: (GEPI->getNumOperands() >= 3 &&
isa<ConstantInt>(GEPI->getOperand(2)) && "Unexpected GEPI!"), function
RewriteHeapSROALoadUser, file
/Users/criswell/src/llvm27/lib/Transforms/IPO/GlobalOpt.cpp, line 1197.
0  opt               0x000000010041c557 PrintStackTrace(void*) + 38
1  opt               0x000000010041ca4e SignalHandler(int) + 336
2  libSystem.B.dylib 0x00007fff83053eaa _sigtramp + 26
3  libSystem.B.dylib 0x00000001013f5000 _sigtramp + 2117734768
4  libSystem.B.dylib 0x00007fff830cfe74 __pthread_markcancel + 0
5  opt               0x0000000100083a33
RewriteHeapSROALoadUser(llvm::Instruction*, llvm::DenseMap<llvm::Value*,
std::vector<llvm::Value*, std::allocator<llvm::Value*> >,
llvm::DenseMapInfo<llvm::Value*>, llvm::DenseMapInfo<std::vector<llvm::Value*,
std::allocator<llvm::Value*> > > >&, std::vector<std::pair<llvm::PHINode*,
unsigned int>, std::allocator<std::pair<llvm::PHINode*, unsigned int> > >&) +
574
6  opt               0x0000000100083dfc
RewriteUsesOfLoadForHeapSRoA(llvm::LoadInst*, llvm::DenseMap<llvm::Value*,
std::vector<llvm::Value*, std::allocator<llvm::Value*> >,
llvm::DenseMapInfo<llvm::Value*>, llvm::DenseMapInfo<std::vector<llvm::Value*,
std::allocator<llvm::Value*> > > >&, std::vector<std::pair<llvm::PHINode*,
unsigned int>, std::allocator<std::pair<llvm::PHINode*, unsigned int> > >&) +
109
7  opt               0x0000000100084c3c
PerformHeapAllocSRoA(llvm::GlobalVariable*, llvm::CallInst*, llvm::Value*,
llvm::TargetData*) + 2457
8  opt               0x0000000100085e5d
TryToOptimizeStoreOfMallocToGlobal(llvm::GlobalVariable*, llvm::CallInst*,
llvm::Type const*, llvm::ilist_iterator<llvm::GlobalVariable>&,
llvm::TargetData*) + 1173
9  opt               0x000000010008660d
OptimizeOnceStoredGlobal(llvm::GlobalVariable*, llvm::Value*,
llvm::ilist_iterator<llvm::GlobalVariable>&, llvm::TargetData*) + 309
10 opt               0x0000000100087bcd (anonymous
namespace)::GlobalOpt::ProcessInternalGlobal(llvm::GlobalVariable*,
llvm::ilist_iterator<llvm::GlobalVariable>&) + 2081
11 opt               0x0000000100087dbe (anonymous
namespace)::GlobalOpt::OptimizeGlobalVars(llvm::Module&) + 360
12 opt               0x0000000100087e75 (anonymous
namespace)::GlobalOpt::runOnModule(llvm::Module&) + 137
13 opt               0x00000001003ab5fe
llvm::MPPassManager::runOnModule(llvm::Module&) + 376
14 opt               0x00000001003ad4d1
llvm::PassManagerImpl::run(llvm::Module&) + 111
15 opt               0x00000001003ad533 llvm::PassManager::run(llvm::Module&) +
33
16 opt               0x000000010006090f main + 3130
17 opt               0x00000001000568cc start + 52
18 opt               0x0000000000000006 start + 4294612846

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 17:46:39 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 17:46:39 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6759] Poor diagnostic with missing type name in
 constructor parameter
In-Reply-To: <bug-6759-206@http.llvm.org/bugs/>
References: <bug-6759-206@http.llvm.org/bugs/>
Message-ID: <20100401224639.D6E832A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6759

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |quality-of-implementation
             Status|RESOLVED                    |REOPENED
         Resolution|INVALID                     |
            Summary|"friend class" implies      |Poor diagnostic with
                   |forward declaration         |missing type name in
                   |"class"?                    |constructor parameter

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 18:34:02 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 18:34:02 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6761] New: optimizer generates an infinite loop
 with floating-point counter.
Message-ID: <bug-6761-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6761

           Summary: optimizer generates an infinite loop with
                    floating-point counter.
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nanki at dotswitch.net
                CC: llvmbugs at cs.uiuc.edu


I tested the following code which prints 'H' 10 times.

-- h.ll
; ModuleID = '<stdin>'

define void @main() nounwind {
; <label>:0
  br label %1

; <label>:1                                       ; preds = %1, %0
  %2 = phi double [ 9.000000e+00, %0 ], [ %4, %1 ] ; <double> [#uses=1]
  %3 = call i32 @putchar(i8 72)                   ; <i32> [#uses=0]
  %4 = fsub double %2, 1.000000e+00               ; <double> [#uses=2]
  %5 = fcmp oge double %4, 0.000000e+00           ; <i1> [#uses=1]
  br i1 %5, label %1, label %6

; <label>:6                                       ; preds = %1
  ret void
}

declare i32 @putchar(i8) nounwind
--

$ cat h.ll | llvm-as | lli
HHHHHHHHHH

But with optimization, it turns out an infinite loop.

$ cat h.ll | llvm-as | opt -O1 | lli
HHHHHHHHHHHHH...

* Using i32 counter works well with optimization.
* This problem seems to be happened when the delta is 1.0, 2.0, 3.0...
* See also http://llvm.org/bugs/show_bug.cgi?id=3299

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 18:52:57 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 18:52:57 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6762] New: clang++ trigger ADL on va_list
Message-ID: <bug-6762-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6762

           Summary: clang++ trigger ADL on va_list
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
            Blocks: 5511


Testcase:
#define a_list __builtin_va_list
extern a_list l;
extern int f (a_list arg);
namespace n {
int f(a_list arguments);
void y() {
  f(l);
}
}

On x86-64, gcc accepts this, clang rejects it.  Affects Firefox build.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 20:49:30 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 20:49:30 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6289] Use of ostringstream in c++ code causes
 resulting binary to segfault
In-Reply-To: <bug-6289-206@http.llvm.org/bugs/>
References: <bug-6289-206@http.llvm.org/bugs/>
Message-ID: <20100402014930.EBB582A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6289

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME

--- Comment #5 from Douglas Gregor <dgregor at apple.com> 2010-04-01 20:49:30 CDT ---
This is working for me now. Please try the latest Clang from Subversion and
re-open this bug if it isn't fixed for you.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 20:51:42 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 20:51:42 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6763] New: Using True/False as identifiers is not
	portable
Message-ID: <bug-6763-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6763

           Summary: Using True/False as identifiers is not portable
           Product: new-bugs
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: mtsmith at synopsys.com
                CC: llvmbugs at cs.uiuc.edu


The definition of areInvalidOperands in llvm/Instructions.h is not very
friendly. It uses True and False as identifiers, which in my code - and
probably a lot of older C code - are defined to 0 and 1.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  1 22:02:15 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  1 Apr 2010 22:02:15 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6763] Using True/False as identifiers is not
	portable
In-Reply-To: <bug-6763-206@http.llvm.org/bugs/>
References: <bug-6763-206@http.llvm.org/bugs/>
Message-ID: <20100402030215.26D642A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6763

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-01 22:02:14 CDT ---
We can't avoid every possible crazy macro :)

If these are coming from system headers, I recommend doing a #undef before the
#include of the llvm header.  Alternatively, #include the llvm headers before
anything else.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  2 00:15:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  2 Apr 2010 00:15:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6764] New: ../../llvm-gcc/gcc/crtstuff.c:378:
 internal compiler error: Bus error
Message-ID: <bug-6764-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6764

           Summary: ../../llvm-gcc/gcc/crtstuff.c:378: internal compiler
                    error: Bus error
           Product: Build scripts
           Version: 2.6
          Platform: DEC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Makefiles
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dot618point at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4623)
 --> (http://llvm.org/bugs/attachment.cgi?id=4623)
internal compiler error: Bus error

/home/zheng/llvm/obj/./gcc/xgcc -B/home/zheng/llvm/obj/./gcc/
-B/usr/local/alphaev6-unknown-linux-gnu/bin/
-B/usr/local/alphaev6-unknown-linux-gnu/lib/ -isystem
/usr/local/alphaev6-unknown-linux-gnu/include -isystem
/usr/local/alphaev6-unknown-linux-gnu/sys-include -O2 -O2 -g -O2  -mieee
-DIN_GCC    -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes
-Wold-style-definition  -isystem ./include  -I. -I. -I../../llvm-gcc/gcc
-I../../llvm-gcc/gcc/. -I../../llvm-gcc/gcc/../include
-I../../llvm-gcc/gcc/../libcpp/include  -I../../llvm-gcc/gcc/../libdecnumber
-I../libdecnumber -I/usr/local/include  -g0 -finhibit-size-directive
-fno-inline-functions -fno-exceptions -fno-zero-initialized-in-bss
-fno-toplevel-reorder   \
          -c ../../llvm-gcc/gcc/crtstuff.c -DCRT_BEGIN \
          -o crtbegin.o
../../llvm-gcc/gcc/crtstuff.c:378: internal compiler error: Bus error
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.
make[3]: *** [crtbegin.o] Error 1
make[3]: Leaving directory `/home/zheng/llvm/obj/gcc'
make[2]: *** [all-stage1-gcc] Error 2
make[2]: Leaving directory `/home/zheng/llvm/obj'
make[1]: *** [stage1-bubble] Error 2
make[1]: Leaving directory `/home/zheng/llvm/obj'
make: *** [all] Error 2

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  2 00:18:30 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  2 Apr 2010 00:18:30 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6765] New:
	../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp error
Message-ID: <bug-6765-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6765

           Summary: ../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp error
           Product: compiler-rt
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: compiler-rt
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dot618point at gmail.com
                CC: llvmbugs at cs.uiuc.edu


make[3]: Entering directory `/home/zheng/llvm/obj/gcc'
c++ -c -g -DIN_GCC -W -Wall -Wwrite-strings -Wmissing-format-attribute
-fno-common -DHAVE_CONFIG_H -Wno-unused
-DTARGET_NAME=\"alphaev6-unknown-linux-gnu\" -frandom-seed=0 -I. -I.
-I../../llvm-gcc/gcc -I../../llvm-gcc/gcc/. -I../../llvm-gcc/gcc/../include
-I../../llvm-gcc/gcc/../libcpp/include  -I../../llvm-gcc/gcc/../libdecnumber
-I../libdecnumber -I/usr/local/include -DENABLE_LLVM -I/usr/local/include 
-D_DEBUG -mieee -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS  
-I. -I. -I../../llvm-gcc/gcc -I../../llvm-gcc/gcc/.
-I../../llvm-gcc/gcc/../include -I../../llvm-gcc/gcc/../libcpp/include 
-I../../llvm-gcc/gcc/../libdecnumber -I../libdecnumber -I/usr/local/include \
                ../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp -o llvm-alpha.o
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp: In member function 'bool
TreeToLLVM::TargetIntrinsicLower(tree_node*, unsigned int, const MemRef*,
llvm::Value*&, const llvm::Type*, std::vector<llvm::Value*,
std::allocator<llvm::Value*> >&)':
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:108: error: no matching function
for call to 'llvm::IntegerType::get(int)'
/usr/local/include/llvm/DerivedTypes.h:108: note: candidates are: static const
llvm::IntegerType* llvm::IntegerType::get(llvm::LLVMContext&, unsigned int)
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:109: error: no matching function
for call to 'llvm::IntegerType::get(int)'
/usr/local/include/llvm/DerivedTypes.h:108: note: candidates are: static const
llvm::IntegerType* llvm::IntegerType::get(llvm::LLVMContext&, unsigned int)
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:111: error: 'Int64Ty' is not a
member of 'llvm::Type'
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:112: error: 'Int64Ty' is not a
member of 'llvm::Type'
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:120: error: 'Int64Ty' is not a
member of 'llvm::Type'
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:121: error: 'Int8Ty' is not a
member of 'llvm::Type'
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:122: error: 'Int64Ty' is not a
member of 'llvm::Type'
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:123: error: 'Int8Ty' is not a
member of 'llvm::Type'
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:126: error: 'Int64Ty' is not a
member of 'llvm::Type'
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:127: error: 'Int64Ty' is not a
member of 'llvm::Type'
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:147: error: 'Int64Ty' is not a
member of 'llvm::Type'
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:149: error: 'Int64Ty' is not a
member of 'llvm::Type'
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:162: error: 'Int64Ty' is not a
member of 'llvm::Type'
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:163: error: 'Int64Ty' is not a
member of 'llvm::Type'
../../llvm-gcc/gcc/config/alpha/llvm-alpha.cpp:165: error: 'Int64Ty' is not a
member of 'llvm::Type'
make[3]: *** [llvm-alpha.o] Error 1
make[3]: Leaving directory `/home/zheng/llvm/obj/gcc'
make[2]: *** [all-stage1-gcc] Error 2
make[2]: Leaving directory `/home/zheng/llvm/obj'
make[1]: *** [stage1-bubble] Error 2
make[1]: Leaving directory `/home/zheng/llvm/obj'
make: *** [all] Error 2

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  2 04:06:47 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  2 Apr 2010 04:06:47 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6766] New: clang crash with a complex C99 structure
 initialisation
Message-ID: <bug-6766-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6766

           Summary: clang crash with a complex C99 structure
                    initialisation
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: quickslyver at free.fr
                CC: llvmbugs at cs.uiuc.edu


It is a reduction from gcc torture test "c99-init-1.c"

$cat b.c
typedef __WCHAR_TYPE__ wchar_t;

struct K {
  wchar_t L[6];
  int M;
} l =  { { L"foo" }, 1 };

$ clang -cc1 b.c -emit-llvm
clang: llvm/tools/clang/lib/CodeGen/CGExprConstant.cpp:51:
bool<unnamed>::ConstStructBuilder::AppendField(const clang::FieldDecl*,
uint64_t, const clang::Expr*): Assertion `NextFieldOffsetInBytes <=
FieldOffsetInBytes && "Field offset mismatch!"' failed.
0  clang 0x08f6b498
Stack dump:
0.    Program arguments: clang -cc1 b.c -emit-llvm 
1.    <eof> parser at end of file
2.    b.c:3:8: LLVM IR generation of declaration 'K'
3.    b.c:6:3: Generating code for declaration 'l'

Aborted

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  2 04:38:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  2 Apr 2010 04:38:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6767] New: Assertion failed: (isVector() &&
 "Invalid vector type!"), function getVectorNumElements,
 file /data/home/rdivacky/llvm/include/llvm/CodeGen/ValueTypes.h, line 550.
Message-ID: <bug-6767-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6767

           Summary: Assertion failed: (isVector() && "Invalid vector
                    type!"), function getVectorNumElements, file
                    /data/home/rdivacky/llvm/include/llvm/CodeGen/ValueTyp
                    es.h, line 550.
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4625)
 --> (http://llvm.org/bugs/attachment.cgi?id=4625)
test case

pes delta$ clang -O2 write.c

  ... warnings ...

Assertion failed: (isVector() && "Invalid vector type!"), function
getVectorNumElements, file
/data/home/rdivacky/llvm/include/llvm/CodeGen/ValueTypes.h, line 550.
Stack dump:
0.      Program arguments: /usr/local/bin/clang -cc1 -triple
x86_64-unknown-freebsd8.0 -S -disable-free -main-file-name write.c
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -resource-dir /usr/local/lib/clang/1.5 -O2
-fmessage-length 152 -fgnu-runtime -fdiagnostics-show-option
-fcolor-diagnostics -o /tmp/cc-9NHkV8.s -x c write.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'X86 DAG->DAG Instruction Selection' on function
'@write_contents'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

this is on X86-64, I can provide the BC but I cant reproduce it with it

pes delta$ clang -c -emit-llvm write.c

  ... warnings ...

pes delta$ llc -O2 write.o
pes delta$

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  2 09:14:35 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  2 Apr 2010 09:14:35 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6768] New: clang crash on function call with vla
	argument
Message-ID: <bug-6768-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6768

           Summary: clang crash on function call with vla argument
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: quickslyver at free.fr
                CC: llvmbugs at cs.uiuc.edu


testcase reduced from vla-4.c gcc torture test

$cat b.c
int foo(int n);
int foo6(int a, int b[a][a]);

int main() {
  int b[10][10];
  foo6(10, b);
  return 0;
}

$ clang -cc1 -emit-llvm c.c
clang: llvm/tools/clang/lib/CodeGen/CGExpr.cpp:1084: clang::CodeGen::LValue
clang::CodeGen::CodeGenFunction::EmitDeclRefLValue(const clang::DeclRefExpr*):
Assertion `V && "DeclRefExpr not entered in LocalDeclMap?"' failed.
0  clang 0x08f6b498
Stack dump:
0.    Program arguments: clang -cc1 -emit-llvm c.c 
1.    <eof> parser at end of file
2.    c.c:4:5: LLVM IR generation of declaration 'main'
3.    c.c:4:5: Generating code for declaration 'main'
4.    c.c:4:12: LLVM IR generation of compound statement ('{}')
Aborted

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  2 13:00:12 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  2 Apr 2010 13:00:12 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6769] New: Codegen crashes emitting code for static
 member function of local class within a static member function
Message-ID: <bug-6769-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6769

           Summary: Codegen crashes emitting code for static member
                    function of local class within a static member
                    function
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chandlerc at gmail.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


% cat t.cc
struct X {
  static void f();
};

void X::f() {
  static int *i;
  {
    struct Y {
      static void g() {
        i = new int();
      }
    };
    (void)Y::g();
  }
  (void)i;
}

% ./bin/clang -c -emit-llvm t.cc
clang: /home/chandlerc/src/llvm/trunk/tools/clang/lib/CodeGen/CGExpr.cpp:1085:
clang::CodeGen::LValue clang::CodeGen::CodeGenFunction::EmitDeclRefLValue(const
clang::DeclRefExpr*): Assertion `V && "DeclRefExpr not entered in
LocalDeclMap?"' failed.
Stack dump:
0.      Program arguments: /home/chandlerc/src/llvm/trunk/build/bin/clang -cc1
-triple x86_64-unknown-linux-gnu -emit-llvm-bc -disable-free -main-file-name
t.cc -mrelocation-model static -mdisable-fp-elim -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -resource-dir
/home/chandlerc/src/llvm/trunk/build/lib/clang/1.5 -fmessage-length 179
-fexceptions -fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics -o t.o
-x c++ t.cc 
1.      <eof> parser at end of file
2.      Per-file LLVM IR generation
3.      t.cc:9:19: Generating code for declaration 'X::f()::Y::g'
4.      t.cc:9:23: LLVM IR generation of compound statement ('{}')
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  2 13:28:46 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  2 Apr 2010 13:28:46 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6757] ternary operator: incompatible operand types
 ('Foo const' and 'Foo')
In-Reply-To: <bug-6757-206@http.llvm.org/bugs/>
References: <bug-6757-206@http.llvm.org/bugs/>
Message-ID: <20100402182846.1721E2A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6757

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Douglas Gregor <dgregor at apple.com> 2010-04-02 13:28:45 CDT ---
Fixed in r100196.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  2 14:36:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  2 Apr 2010 14:36:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6767] Assertion failed: (isVector() && "Invalid
 vector type!"), function getVectorNumElements,
 file /data/home/rdivacky/llvm/include/llvm/CodeGen/ValueTypes.h, line 550.
In-Reply-To: <bug-6767-206@http.llvm.org/bugs/>
References: <bug-6767-206@http.llvm.org/bugs/>
Message-ID: <20100402193656.D9DB52A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6767

Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Evan Cheng <evan.cheng at apple.com> 2010-04-02 14:36:56 CDT ---
memset of undef value.
Fixed: 100208.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  2 15:48:16 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  2 Apr 2010 15:48:16 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6770] New: clang::Sema::ActOnTag(...): Assertion
 `Previous.isSingleResult()' failed
Message-ID: <bug-6770-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6770

           Summary: clang::Sema::ActOnTag(...): Assertion
                    `Previous.isSingleResult()' failed
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: evan at chromium.org
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4627)
 --> (http://llvm.org/bugs/attachment.cgi?id=4627)
reduction of problem

Hit while trying to -fsyntax-only Chromium.
clang version 1.5 (trunk 100211)

template <class T>
void f() {
  friend class f;
}


$ clang++ -c reduce.cc
reduce.cc:3:3: error: 'friend' used outside of class
  friend class f;
  ^
clang: SemaDecl.cpp:4912: virtual clang::OpaquePtr<0>
clang::Sema::ActOnTag(clang::Scope*, unsigned int, clang::Action::TagUseKind,
clang::SourceLocation, const clang::CXXScopeSpec&, clang::IdentifierInfo*,
clang::SourceLocation, clang::AttributeList*, clang::AccessSpecifier,
clang::ASTMultiPtr<&clang::ActionBase::DeleteTemplateParams>, bool&, bool&):
Assertion `Previous.isSingleResult()' failed.
0  clang           0x00000000011a1e8f
1  clang           0x00000000011a291a
2  libpthread.so.0 0x00007f92166b4190
3  libc.so.6       0x00007f92159ba4b5 gsignal + 53
4  libc.so.6       0x00007f92159bdf50 abort + 384
5  libc.so.6       0x00007f92159b3481 __assert_fail + 241
6  clang           0x00000000006328fc
7  clang           0x0000000000936525
8  clang           0x0000000000929df4
9  clang           0x000000000092e67b
10 clang           0x000000000092e7d6
11 clang           0x0000000000915c79
12 clang           0x00000000009169c6
13 clang           0x0000000000917101
14 clang           0x00000000009225a0
15 clang           0x00000000009504b9
16 clang           0x0000000000950921
17 clang           0x0000000000950bab
18 clang           0x000000000092e8ad
19 clang           0x0000000000922d70
20 clang           0x0000000000922f7a
21 clang           0x00000000005f2beb
22 clang           0x00000000004103a9
23 clang           0x000000000040294c
24 clang           0x00000000004057e1 main + 2417
25 libc.so.6       0x00007f92159a5abd __libc_start_main + 253
26 clang           0x00000000004002a9
Stack dump:
0.    Program arguments: /work/chrome/llvm/Release/bin/clang -cc1 -triple
x86_64-unknown-linux-gnu -S -disable-free -main-file-name reduce.cc
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -resource-dir
/work/chrome/llvm/Release/lib/clang/1.5 -fmessage-length 129 -fexceptions
-fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics -o /tmp/cc-vZAHfq.s
-x c++ reduce.cc 
1.    reduce.cc:3:17: current parser token ';'
2.    reduce.cc:2:10: parsing function body 'f'
3.    reduce.cc:2:10: in compound statement ('{}')
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  2 23:28:49 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  2 Apr 2010 23:28:49 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6748] missed optimization for vtables
In-Reply-To: <bug-6748-206@http.llvm.org/bugs/>
References: <bug-6748-206@http.llvm.org/bugs/>
Message-ID: <20100403042849.31C002A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6748

Rafael ?vila de Esp?ndola <rafael.espindola at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Rafael ?vila de Esp?ndola <rafael.espindola at gmail.com> 2010-04-02 23:28:48 CDT ---
fixed in r100266.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr  3 00:25:06 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat,  3 Apr 2010 00:25:06 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6524] opt -internalize breaks C++ initialization
 (compiling with clang)
In-Reply-To: <bug-6524-206@http.llvm.org/bugs/>
References: <bug-6524-206@http.llvm.org/bugs/>
Message-ID: <20100403052506.567592A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6524

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |INVALID

--- Comment #7 from Chris Lattner <clattner at apple.com> 2010-04-03 00:25:05 CDT ---
internalize is working correctly.  The likely problem here is that it is
internalizing _ZTISo, whic is the typeinfo for basic_ostream.  Since libstdc++
defines its own version of this symbol and expects pointer comparisons between
them to work, all sorts of badness breaks out.  However, this is exactly what
internalize is supposed to do: seal off the edges of a fully contained program.
 Because you haven't statically linked in libstdc++ and haven't provided a
correct export list, you lose.

However, there is a bug here that doesn't manifest in this testcase:
available_externally functions should not be externalized.  I fixed this
theoretical issue in r100269.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr  3 01:30:28 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat,  3 Apr 2010 01:30:28 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6761] optimizer generates an infinite loop with
 floating-point counter.
In-Reply-To: <bug-6761-206@http.llvm.org/bugs/>
References: <bug-6761-206@http.llvm.org/bugs/>
Message-ID: <20100403063028.475EE2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6761

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-03 01:30:27 CDT ---
Fixed in a series of patches leading up to 100282, however, there are still
bugs which I'll iron out in a few subsequent patches.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr  3 02:23:50 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat,  3 Apr 2010 02:23:50 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6771] New: Quoted -D macros break static analyzer
Message-ID: <bug-6771-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6771

           Summary: Quoted -D macros break static analyzer
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: alexei.svitkine at gmail.com
                CC: llvmbugs at cs.uiuc.edu


If a macro in the form of: -DFOO='"BAR"' is specified on the command-line to
scan-build, it will issue the following error:

<command line>:10:25: error: missing terminating '"' character
#define FOO \"BAR\"

Preventing it from working. This does not happen if clang or gcc is given that
command-line option.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr  3 07:33:27 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat,  3 Apr 2010 07:33:27 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6772] New: Not an error for gcc -- error: default
 initialization of an object of const type '...' requires a user-provided
 default constructor
Message-ID: <bug-6772-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6772

           Summary: Not an error for gcc -- error: default initialization
                    of an object of const type '...' requires a
                    user-provided default constructor
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: nms+llvm at otdel-1.org
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Script started on Sat Apr  3 16:29:45 2010
$ cat ccd.cpp
struct A {
        A() {}
};

class B {
        A a_;
};

int
foo()
{
        const B b;
        return 0;
}
$ g++ --version
g++ (GCC) 4.2.1 20070719  [FreeBSD]
Copyright (C) 2007 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

$ g++ -c ccd.cpp
$ clang++ --version
clang version 1.5 (trunk)
Target: i386-portbld-freebsd9.0
Thread model: posix
$ clang++ -c ccd.cpp
ccd.cpp:12:10: error: default initialization of an object of const type 'B
const' requires a user-provided default constructor
        const B b;
                ^
1 diagnostic generated.
$ ^D
Script done on Sat Apr  3 16:30:29 2010

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr  3 18:49:45 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat,  3 Apr 2010 18:49:45 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6773] New: Add Instruction Fold for Masked Merge
Message-ID: <bug-6773-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6773

           Summary: Add Instruction Fold for Masked Merge
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: me22.ca at gmail.com
                CC: llvmbugs at cs.uiuc.edu


When experimenting with SHA, I was pleased to notice that LLVM will fold the
"Maj" function (x & y) ^ (x & z) ^ (y & z) down to ((z ^ y) & x) ^ (y & z).

The "Ch" function, though, doesn't fold.  (x & y) | (~x & z) should become ((y
^ z) & x) ^ z, as mentioned on
http://graphics.stanford.edu/~seander/bithacks.html#MaskedMerge (as should (x &
y) ^ (~x & z), the version used in the SHA standard).

(If you're wondering at the similarity, Maj(x,y,z) is equivalent to Ch(x, y|z,
y&z).  Using that implementation with the optimized Ch again gives optimal code
from LLVM as it knows to fold (y|z)^(y&z) down to y^z.)


LLVM IR from Bitter Melon:

define i32 @Ch(i32 %x, i32 %y, i32 %z) nounwind readnone {
entry:
  %0 = and i32 %y, %x                             ; <i32> [#uses=1]
  %not = xor i32 %x, -1                           ; <i32> [#uses=1]
  %1 = and i32 %z, %not                           ; <i32> [#uses=1]
  %2 = xor i32 %1, %0                             ; <i32> [#uses=1]
  ret i32 %2
}

define i32 @Maj(i32 %x, i32 %y, i32 %z) nounwind readnone {
entry:
  %0 = xor i32 %z, %y                             ; <i32> [#uses=1]
  %1 = and i32 %0, %x                             ; <i32> [#uses=1]
  %2 = and i32 %z, %y                             ; <i32> [#uses=1]
  %3 = xor i32 %1, %2                             ; <i32> [#uses=1]
  ret i32 %3
}

define i32 @Ch2(i32 %x, i32 %y, i32 %z) nounwind readnone {
entry:
  %0 = xor i32 %z, %y                             ; <i32> [#uses=1]
  %1 = and i32 %0, %x                             ; <i32> [#uses=1]
  %2 = xor i32 %1, %z                             ; <i32> [#uses=1]
  ret i32 %2
}

define i32 @Maj2(i32 %x, i32 %y, i32 %z) nounwind readnone {
entry:
  %0 = and i32 %z, %y                             ; <i32> [#uses=1]
  %1 = xor i32 %z, %y                             ; <i32> [#uses=1]
  %2 = and i32 %1, %x                             ; <i32> [#uses=1]
  %3 = xor i32 %2, %0                             ; <i32> [#uses=1]
  ret i32 %3
}


>From the following C source:

int Ch(int x, int y, int z) {
   return (x&y) ^ (~x&z);
}
int Maj(int x, int y, int z) {
   return (x&y) ^ (x&z) ^ (y&z);
}

int Ch2(int x, int y, int z) {
   return z ^ ((y ^ z) & x);
}
int Maj2(int x, int y, int z) {
   return Ch2(x, y|z, y&z);
}

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr  4 04:15:35 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun,  4 Apr 2010 04:15:35 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6772] Not an error for gcc -- error: default
 initialization of an object of const type '...' requires a user-provided
 default constructor
In-Reply-To: <bug-6772-206@http.llvm.org/bugs/>
References: <bug-6772-206@http.llvm.org/bugs/>
Message-ID: <20100404091535.4E5572A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6772

Chandler Carruth <chandlerc at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |chandlerc at gmail.com
         Resolution|                            |INVALID

--- Comment #1 from Chandler Carruth <chandlerc at gmail.com> 2010-04-04 04:15:34 CDT ---
(In reply to comment #0)
> Script started on Sat Apr  3 16:29:45 2010
> $ cat ccd.cpp
> struct A {
>         A() {}
> };
> 
> class B {
>         A a_;
> };
> 
> int
> foo()
> {
>         const B b;
>         return 0;
> }
> $ g++ --version
> g++ (GCC) 4.2.1 20070719  [FreeBSD]
> Copyright (C) 2007 Free Software Foundation, Inc.
> This is free software; see the source for copying conditions.  There is NO
> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
> 
> $ g++ -c ccd.cpp
> $ clang++ --version
> clang version 1.5 (trunk)
> Target: i386-portbld-freebsd9.0
> Thread model: posix
> $ clang++ -c ccd.cpp
> ccd.cpp:12:10: error: default initialization of an object of const type 'B
> const' requires a user-provided default constructor

Clang is correct: C++'03 [dcl.init] p9: If no initializer is specified for an
object, and the object is of (possibly cv-qualified) non-POD class type (or
array thereof), the object shall be default-initialized; if the object is of
const-qualified type, the underlying class type shall have a user-declared
default constructor.

The user declared default constructor for A makes A non-POD. The A member in B
makes B non-POD, and thus this rule kicks in, and the code provided is invalid.

You could re-open as a request for a GCC-compatibility setting to allow such
code, but I don't personally see a lot of value in it. Perhaps others do, so
feel free to re-open if you want to pursue that path.

>         const B b;
>                 ^
> 1 diagnostic generated.
> $ ^D
> Script done on Sat Apr  3 16:30:29 2010

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr  4 06:18:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun,  4 Apr 2010 06:18:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6758] JIT doesn't work on PowerPC: relocation fails
In-Reply-To: <bug-6758-206@http.llvm.org/bugs/>
References: <bug-6758-206@http.llvm.org/bugs/>
Message-ID: <20100404111818.7EAC32A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6758

T?r?k Edwin <edwintorok at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #8 from T?r?k Edwin <edwintorok at gmail.com> 2010-04-04 06:18:17 CDT ---
Marking this as dup of PR5201.

*** This bug has been marked as a duplicate of bug 5201 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr  4 06:50:47 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun,  4 Apr 2010 06:50:47 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6774] New: Too much stack manipulation with
	-tailcallopt on x86
Message-ID: <bug-6774-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6774

           Summary: Too much stack manipulation with -tailcallopt on x86
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: batterseapower at hotmail.com
                CC: llvmbugs at cs.uiuc.edu


The GHC LLVM backend generates code with lots of tail calls, and we use
-tailcallopt to ensure that these are turned into actual "jmp" instructions.

Unfortunately, the -tailcallopt option seems to lead to excessive manipulation
of the stack pointer. Here is some an example IR fragment generated from some
relatively simple Haskell code:

{{{
define cc10 void @Toy_toy_entry(i32 %stg_terei_baseArg, i32
%stg_terei_spArg, i32 %stg_terei_hpArg, i32 %stg_terei_r1Arg) nounwind
{
cdx:
  %ndz = add i32 %stg_terei_spArg, -4             ; <i32> [#uses=3]
  %ndB = add i32 %stg_terei_baseArg, 84           ; <i32> [#uses=1]
  %ndC = inttoptr i32 %ndB to i32*                ; <i32*> [#uses=1]
  %ndD = load i32* %ndC                           ; <i32> [#uses=1]
  %ndE = icmp ult i32 %ndz, %ndD                  ; <i1> [#uses=1]
  br i1 %ndE, label %cdG, label %ndH

ndH:                                              ; preds = %cdx
  %ndJ = inttoptr i32 %stg_terei_spArg to i32*    ; <i32*> [#uses=2]
  %ndK = load i32* %ndJ                           ; <i32> [#uses=1]
  %ndP = inttoptr i32 %ndz to i32*                ; <i32*> [#uses=1]
  store i32 add (i32 ptrtoint (%Toy_toy1_closure_struct*
@Toy_toy2_closure to i32), i32 1), i32* %ndP
  store i32 ptrtoint (%sbo_info_struct* @sbo_info to i32), i32* %ndJ
  tail call cc10 void @stg_ap_p_fast(i32 %stg_terei_baseArg, i32 %ndz,
i32 %stg_terei_hpArg, i32 %ndK) nounwind
  ret void

cdG:                                              ; preds = %cdx
  %ne1 = add i32 %stg_terei_baseArg, -4           ; <i32> [#uses=1]
  %ne2 = inttoptr i32 %ne1 to i32*                ; <i32*> [#uses=1]
  %ne3 = load i32* %ne2                           ; <i32> [#uses=1]
  %ne4 = inttoptr i32 %ne3 to void (i32, i32, i32, i32)* ; <void (i32,
i32, i32, i32)*> [#uses=1]
  tail call cc10 void %ne4(i32 %stg_terei_baseArg, i32
%stg_terei_spArg, i32 %stg_terei_hpArg, i32 ptrtoint
(%Toy_toy_closure_struct* @Toy_toy_closure to i32)) nounwind
  ret void
}
}}}

With -tailcallopt, llc generates the following x86:

{{{
_Toy_toy_entry:                         ## @Toy_toy_entry
## BB#0:                                ## %cdx
    subl    $12, %esp
    leal    -4(%ebp), %eax
    cmpl    84(%ebx), %eax
    jb    LBB2_2
## BB#1:                                ## %ndH
    movl    (%ebp), %esi
    movl    $_Toy_toy2_closure+1, -4(%ebp)
    movl    $_sbo_info, (%ebp)
    movl    %eax, %ebp
    addl    $12, %esp
    jmp    _stg_ap_p_fast  # TAILCALL
LBB2_2:                                 ## %cdG
    movl    -4(%ebx), %eax
    movl    $_Toy_toy_closure, %esi
    addl    $12, %esp
    jmpl    *%eax  # TAILCALL

    .globl    ___stginit_Toy_
    .align    4, 0x90
}}}

Note the subl of %esp at the end point and corresponding addl to %esp at both
exit points. %esp is not used in the body of the function, so this is
pointless.

Compiling without -tailcallopt removes this extra stack manipulation.

As far as I can see, the two tail calls meet the requirements for "sibling
calls" - and indeed llc without -tailcallopt still turns the calls into "jmp"s.
Is it possible for LLVM to try this approach first, and only fall back on the
full tail-call optimisation if it can't apply the simpler sibling-call
optimisation?

(See also my post to the GHC mailing list at
http://article.gmane.org/gmane.comp.lang.haskell.cvs.ghc/38323)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr  4 11:07:11 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun,  4 Apr 2010 11:07:11 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6775] New: clang should support -use-gold-plugin
Message-ID: <bug-6775-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6775

           Summary: clang should support -use-gold-plugin
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Driver
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Clang should support -use-gold-plugin, the same way llvm-gcc supports it.

It should also build & install libLLVMgold.so along with clang.

Alternatively it should strip flags unknown to llvm-ld, and invoke it, as
described here: http://llvm.org/bugs/show_bug.cgi?id=4155
However 'nm' would still need the gold plugin to work with LTO objects.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr  4 16:55:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun,  4 Apr 2010 16:55:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6776] New: false positive : bad bound guessing
Message-ID: <bug-6776-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6776

           Summary: false positive : bad bound guessing
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: castet.matthieu at free.fr
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4632)
 --> (http://llvm.org/bugs/attachment.cgi?id=4632)
testcase

In the attached test case, type can only be between 0 and 63 and c is always
set to 2.
This is because type come from (uint32_t)>>(32-(6))).

But analyser fails to see that :
test_ana.c:27:2: warning: Undefined or garbage value returned to caller

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr  4 18:35:46 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun,  4 Apr 2010 18:35:46 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6777] New: Code generators (both llvmc and Jit) get
 stuck when dealing circular CFG (svn r100296)
Message-ID: <bug-6777-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6777

           Summary: Code generators (both llvmc and Jit) get stuck when
                    dealing circular CFG (svn  r100296)
           Product: compiler-rt
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: compiler-rt
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bugspynet at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The following arbitrary example makes the code generators to get stuck
(llx won't return from command line, Jit won't return from function
call)
Basically it is a CFG with circles, (the function will return by
comparing a branch counter to a threshold on runtime - see the
"Brancher" blocks below)
Any idea what goes wrong? Is it a bug in llvm ?

; ModuleID = 'TestModule!'

define double @fn1(i8*, i64) {
Entry:
  %BranchCounter = alloca i32                     ; <i32*> [#uses=7]
  store i32 0, i32* %BranchCounter
  %Vars = alloca double, i32 3                    ; <double*> [#uses=3]
  %DoubleVar = getelementptr inbounds double* %Vars, i32 0 ; <double*>
[#uses=6]
  store double 0xC031AD5A40000000, double* %DoubleVar
  %DoubleVar1 = getelementptr inbounds double* %Vars, i32 1 ;
<double*> [#uses=1]
  store double 0xC055A9E9E0000000, double* %DoubleVar1
  %DoubleVar2 = getelementptr inbounds double* %Vars, i32 2 ;
<double*> [#uses=4]
  store double 0xC0495B8A80000000, double* %DoubleVar2
  br label %Brancher

Exit:                                             ; preds =
%Brancher7, %Brancher4, %Brancher
  %2 = load double* %DoubleVar                    ; <double> [#uses=1]
  ret double %2

Body:                                             ; preds = %Brancher
  %3 = srem i64 30, %1                            ; <i64> [#uses=1]
  %4 = getelementptr inbounds i8* %0, i64 %3      ; <i8*> [#uses=1]
  %5 = load i8* %4                                ; <i8> [#uses=1]
  %PeekResult = uitofp i8 %5 to double            ; <double> [#uses=0]
  %6 = load double* %DoubleVar2                   ; <double> [#uses=1]
  store double 0x40619C4E80000000, double* %DoubleVar
  %7 = srem i64 178, %1                           ; <i64> [#uses=1]
  %8 = getelementptr inbounds i8* %0, i64 %7      ; <i8*> [#uses=1]
  %9 = load i8* %8                                ; <i8> [#uses=1]
  %PeekResult9 = uitofp i8 %9 to double           ; <double> [#uses=0]
  store double %6, double* %DoubleVar
  %10 = load double* %DoubleVar2                  ; <double> [#uses=0]
  %11 = load double* %DoubleVar                   ; <double> [#uses=0]
  br i1 false, label %Brancher7, label %Brancher4

Brancher:                                         ; preds = %Entry
  %12 = load i32* %BranchCounter                  ; <i32> [#uses=1]
  %"BrCounter++" = add i32 %12, 1                 ; <i32> [#uses=2]
  store i32 %"BrCounter++", i32* %BranchCounter
  %13 = icmp ult i32 %"BrCounter++", 10           ; <i1> [#uses=1]
  br i1 %13, label %Body, label %Exit

Body3:                                            ; preds = %Brancher4
  %14 = srem i64 40, %1                           ; <i64> [#uses=1]
  %15 = getelementptr inbounds i8* %0, i64 %14    ; <i8*> [#uses=1]
  %16 = load i8* %15                              ; <i8> [#uses=1]
  %PeekResult10 = uitofp i8 %16 to double         ; <double> [#uses=2]
  %17 = fsub double 0xC058173FC0000000, %PeekResult10 ; <double> [#uses=1]
  store double %PeekResult10, double* %DoubleVar2
  %18 = fsub double 0x406AB76488000000, %17       ; <double> [#uses=1]
  %19 = fdiv double %18, 0x4068C87680000000       ; <double> [#uses=0]
  br label %Brancher4

Brancher4:                                        ; preds = %Body6,
%Body3, %Body
  %20 = load i32* %BranchCounter                  ; <i32> [#uses=1]
  %"BrCounter++5" = add i32 %20, 1                ; <i32> [#uses=2]
  store i32 %"BrCounter++5", i32* %BranchCounter
  %21 = icmp ult i32 %"BrCounter++5", 10          ; <i1> [#uses=1]
  br i1 %21, label %Body3, label %Exit

Body6:                                            ; preds = %Brancher7
  %22 = load double* %DoubleVar                   ; <double> [#uses=3]
  %23 = srem i64 7118, %1                         ; <i64> [#uses=1]
  %24 = getelementptr inbounds i8* %0, i64 %23    ; <i8*> [#uses=1]
  %25 = load i8* %24                              ; <i8> [#uses=1]
  %PeekResult11 = uitofp i8 %25 to double         ; <double> [#uses=0]
  %26 = fsub double %22, %22                      ; <double> [#uses=0]
  %27 = fcmp oeq double 0xC04FBB2E40000000, %22   ; <i1> [#uses=1]
  br i1 %27, label %Brancher4, label %Brancher7

Brancher7:                                        ; preds = %Body6, %Body
  %28 = load i32* %BranchCounter                  ; <i32> [#uses=1]
  %"BrCounter++8" = add i32 %28, 1                ; <i32> [#uses=2]
  store i32 %"BrCounter++8", i32* %BranchCounter
  %29 = icmp ult i32 %"BrCounter++8", 10          ; <i1> [#uses=1]
  br i1 %29, label %Body6, label %Exit
}

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr  4 19:07:21 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun,  4 Apr 2010 19:07:21 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6778] New: llvm-gcc-4.2 from release_2.7 branch
 fails to build f951 on x86_64-apple-darwin10
Message-ID: <bug-6778-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6778

           Summary: llvm-gcc-4.2 from release_2.7 branch fails to build
                    f951 on x86_64-apple-darwin10
           Product: new-bugs
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: release blocker
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: howarth at bromo.msbb.uc.edu
                CC: llvmbugs at cs.uiuc.edu


When attempting to build llvm-gcc-4.2 from the release_2.7 branch at r99913,
the build fails at...

c++   -g -O2 -mdynamic-no-pic -DIN_GCC   -W -Wall -Wwrite-strings
-Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long
-Wno-variadic-macros -Wno-overlength-strings -Wold-style-definition
-Wmissing-format-attribute   -mdynamic-no-pic -DHAVE_CONFIG_H  -o f951 \
        fortran/arith.o fortran/array.o fortran/bbt.o fortran/check.o
fortran/data.o fortran/decl.o fortran/dump-parse-tree.o fortran/error.o
fortran/expr.o fortran/interface.o fortran/intrinsic.o fortran/io.o
fortran/iresolve.o fortran/match.o fortran/matchexp.o fortran/misc.o
fortran/module.o fortran/openmp.o fortran/options.o fortran/parse.o
fortran/primary.o fortran/resolve.o fortran/scanner.o fortran/simplify.o
fortran/st.o fortran/symbol.o fortran/convert.o fortran/dependency.o
fortran/f95-lang.o fortran/trans.o fortran/trans-array.o fortran/trans-common.o
fortran/trans-const.o fortran/trans-decl.o fortran/trans-expr.o
fortran/trans-intrinsic.o fortran/trans-io.o fortran/trans-openmp.o
fortran/trans-stmt.o fortran/trans-types.o llvm-main.o  libbackend.a
../libcpp/libcpp.a /sw/lib/llvm/lib/libLLVMBitReader.a
/sw/lib/llvm/lib/libLLVMipo.a /sw/lib/llvm/lib/libLLVMBitWriter.a
/sw/lib/llvm/lib/libLLVMX86Disassembler.a
/sw/lib/llvm/lib/libLLVMX86AsmParser.a /sw/lib/llvm/lib/libLLVMMCParser.a
/sw/lib/llvm/lib/libLLVMX86AsmPrinter.a /sw/lib/llvm/lib/libLLVMX86CodeGen.a
/sw/lib/llvm/lib/libLLVMSelectionDAG.a /sw/lib/llvm/lib/libLLVMAsmPrinter.a
/sw/lib/llvm/lib/libLLVMCodeGen.a /sw/lib/llvm/lib/libLLVMScalarOpts.a
/sw/lib/llvm/lib/libLLVMInstCombine.a /sw/lib/llvm/lib/libLLVMTransformUtils.a
/sw/lib/llvm/lib/libLLVMipa.a /sw/lib/llvm/lib/libLLVMAnalysis.a
/sw/lib/llvm/lib/libLLVMTarget.a /sw/lib/llvm/lib/libLLVMMC.a
/sw/lib/llvm/lib/libLLVMCore.a /sw/lib/llvm/lib/libLLVMX86Info.a
/sw/lib/llvm/lib/libLLVMSupport.a /sw/lib/llvm/lib/libLLVMSystem.a attribs.o
stub-objc.o stub-c.o -L/sw/lib -lmpfr -lgmp ../libcpp/libcpp.a -lintl -L/sw/lib
-liconv  ../libiberty/libiberty.a ../libdecnumber/libdecnumber.a
-L/sw/lib/llvm/lib  -lpthread -lm 
Undefined symbols:
  "_gt_ggc_r_gt_darwin_c_h", referenced from:
      _gt_ggc_rtab in f95-lang.o
ld: symbol(s) not found
collect2: ld returned 1 exit status
make[3]: *** [f951] Error 1
make[2]: *** [all-stage2-gcc] Error 2
make[1]: *** [stage2-bubble] Error 2
make: *** [all] Error 2

when built with...

# ../llvm-gcc4.2-2.7.source/configure   --prefix=/sw
--prefix=/sw/lib/llvm-gcc-4.2 --mandir=/sw/share/man --infodir=/sw/share/info
--with-gmp=/sw --with-libiconv-prefix=/sw --with-system-zlib
--x-includes=/usr/X11R6/include --x-libraries=/usr/X11R6/lib
--with-as=/Developer/usr/bin/as --with-ld=/Developer/usr/bin/ld
--with-nm=/Developer/usr/bin/nm --with-gxx-include-dir=/usr/include/c++/4.2.1
--enable-llvm=/sw/lib/llvm --enable-languages=c,c++,fortran

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr  4 22:04:43 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun,  4 Apr 2010 22:04:43 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6779] New: ice: library memcpy does not support
	volatile
Message-ID: <bug-6779-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6779

           Summary: ice: library memcpy does not support volatile
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4633)
 --> (http://llvm.org/bugs/attachment.cgi?id=4633)
failure-inducing C code

Sorry for the big testcase.

regehr at john-home:~/volatile/bugs/tmp309$ clang -emit-llvm small.c -c -o
small.bc -w
regehr at john-home:~/volatile/bugs/tmp309$ llc small.bc -o /dev/null
llc: SelectionDAG.cpp:3526: llvm::SDValue
llvm::SelectionDAG::getMemcpy(llvm::SDValue, llvm::DebugLoc, llvm::SDValue,
llvm::SDValue, llvm::SDValue, unsigned int, bool, bool, const llvm::Value*,
uint64_t, const llvm::Value*, uint64_t): Assertion `!isVol && "library memcpy
does not support volatile"' failed.
0  llc 0x0897c148
Stack dump:
0.    Program arguments: llc small.bc -o /dev/null 
1.    Running pass 'X86 DAG->DAG Instruction Selection' on function '@func_1'
Aborted
regehr at john-home:~/volatile/bugs/tmp309$ clang -v
clang version 1.5 (trunk 100339)
Target: i386-pc-linux-gnu
Thread model: posix

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr  4 23:04:13 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun,  4 Apr 2010 23:04:13 -0500 (CDT)
Subject: [LLVMbugs] [Bug 4269] klee cannot be built in a separate objects
	directory
In-Reply-To: <bug-4269-206@http.llvm.org/bugs/>
References: <bug-4269-206@http.llvm.org/bugs/>
Message-ID: <20100405040413.A1CE12A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=4269

Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Daniel Dunbar <daniel at zuster.org> 2010-04-04 23:04:13 CDT ---
Fixed here:
  http://llvm.org/viewvc/llvm-project?view=rev&revision=100395

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr  4 23:04:30 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun,  4 Apr 2010 23:04:30 -0500 (CDT)
Subject: [LLVMbugs] [Bug 5302] trying to build on mac snow leopard fails
In-Reply-To: <bug-5302-206@http.llvm.org/bugs/>
References: <bug-5302-206@http.llvm.org/bugs/>
Message-ID: <20100405040430.C02E72A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=5302

Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Daniel Dunbar <daniel at zuster.org> 2010-04-04 23:04:30 CDT ---
Should be fixed as of (at least):
  http://llvm.org/viewvc/llvm-project?view=rev&revision=100395

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr  4 23:05:33 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun,  4 Apr 2010 23:05:33 -0500 (CDT)
Subject: [LLVMbugs] [Bug 4295] Support arbitrary width types.
In-Reply-To: <bug-4295-206@http.llvm.org/bugs/>
References: <bug-4295-206@http.llvm.org/bugs/>
Message-ID: <20100405040533.4DBBD2A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=4295

Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Daniel Dunbar <daniel at zuster.org> 2010-04-04 23:05:33 CDT ---
This is done.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 04:18:01 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 04:18:01 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6780] New: clang++ -m32 breaks inline asm
Message-ID: <bug-6780-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6780

           Summary: clang++ -m32 breaks inline asm
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


pes delta$ cat atomicity.cc
extern          "C" {
}
typedef int     _Atomic_word;
namespace __gnu_cxx 
__attribute__((__visibility__("default")))
{
        _Atomic_word    __attribute__((__unused__)) __exchange_and_add(volatile
_Atomic_word * __mem, int __val){
                register _Atomic_word __result;
                __asm__         __volatile__("lock; xadd{l}
{%0,%1|%1,%0}":"=r"(__result), "=m"(*__mem):"0"(__val), "m"(*__mem));
                                return  __result;
        }
}
pes delta$ g++ -c -m32 atomicity.cc && echo success
success
pes delta$ clang++ -c -m32 atomicity.cc && echo success
/tmp/cc-NbWHQp.s: Assembler messages:
/tmp/cc-NbWHQp.s:22: Error: invalid character '{' in mnemonic
clang: error: assembler command failed with exit code 1 (use -v to see
invocation)
pes delta$

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 11:34:11 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 11:34:11 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6329] GNU extension: initialization of members of
 an anonymous struct
In-Reply-To: <bug-6329-206@http.llvm.org/bugs/>
References: <bug-6329-206@http.llvm.org/bugs/>
Message-ID: <20100405163411.55EB82A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6329

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Chris Lattner <clattner at apple.com> 2010-04-05 11:34:11 CDT ---
Verified, thanks.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 11:36:48 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 11:36:48 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6781] New: Template diagnostics should suppress
 instantiation contexts when there are lots
Message-ID: <bug-6781-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6781

           Summary: Template diagnostics should suppress instantiation
                    contexts when there are lots
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


This testcase:

template <int N> struct X {
    static const int value = X<N-1>::value;
};
template struct X<1000>;

compiles to this with gcc:

testsuite/9335.cc:2:38: error: template instantiation depth exceeds maximum of
1024 (use -ftemplate-depth= to increase the maximum) instantiating ?struct
X<-0x00000000000000018>?
testsuite/9335.cc:2:38:   instantiated from ?const int
X<-0x00000000000000017>::value?
testsuite/9335.cc:2:38:   instantiated from ?const int
X<-0x00000000000000016>::value?
testsuite/9335.cc:2:38:   instantiated from ?const int
X<-0x00000000000000015>::value?
testsuite/9335.cc:2:38:   instantiated from ?const int
X<-0x00000000000000014>::value?
testsuite/9335.cc:2:38:   instantiated from ?const int
X<-0x00000000000000013>::value?
testsuite/9335.cc:2:38:   [ skipping 1014 instantiation contexts ]
testsuite/9335.cc:2:38:   instantiated from ?const int X<996>::value?
testsuite/9335.cc:2:38:   instantiated from ?const int X<997>::value?
testsuite/9335.cc:2:38:   instantiated from ?const int X<998>::value?
testsuite/9335.cc:2:38:   instantiated from ?const int X<999>::value?
testsuite/9335.cc:2:38:   instantiated from ?const int X<1000>::value?
testsuite/9335.cc:4:17:   instantiated from here

Clang emits all of those instantiation contexts, which is not helpful.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 12:00:49 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 12:00:49 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6782] New: clang accepts code missing ;
Message-ID: <bug-6782-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6782

           Summary: clang accepts code missing ;
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Clang accepts this, but shouldn't.

template<class T>
         class Class1;

         class Class2 {
         }  // no ;

         typedef Class1<Class2> Type1;

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 12:15:57 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 12:15:57 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6783] New: crash on invalid: (DefinitionData &&
 "queried property of class with no definition)"
Message-ID: <bug-6783-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6783

           Summary: crash on invalid:  (DefinitionData && "queried
                    property of class with no definition)"
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


clang crashes on this:

struct Base {};
         struct X;                       // X derived from Base later but
incomplete
         here
         struct Y {};                    // Y not derived from Base

         int test1(int Base::* p2m, X* object)
        {
             return object->*p2m;        // fails because X is incomplete
         }

        struct X : Base
        {
       };

        int test2(int Base::* p2m, X* object)
        {
            return object->*p2m;        // OK
        }

        int test3(int Base::* p2m, Y* object)
        {
            return object->*p2m;        // fails because Y is not derived from
A
        }

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 12:44:58 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 12:44:58 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6784] New: C++98/03: Template arguments cannot be
	unnamed types
Message-ID: <bug-6784-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6784

           Summary: C++98/03: Template arguments cannot be unnamed types
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


For example, the following is ill-formed in C++98/03:

     1    struct A {};
     2    int foo(A);
     3    template <typename T> int foo(void *, T);
     4    enum {e};
     5    int i = foo(0, e);

but well-formed in C++0x. Clang does not diagnose the error in C++98/03 mode.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 13:04:25 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 13:04:25 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6663] MultiSource/Benchmarks/Bullet/bullet fails on
 x86-32 linux when using -fPIC
In-Reply-To: <bug-6663-206@http.llvm.org/bugs/>
References: <bug-6663-206@http.llvm.org/bugs/>
Message-ID: <20100405180425.D7B392A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6663

Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #15 from Duncan Sands <baldrick at free.fr> 2010-04-05 13:04:25 CDT ---
In the assembler generated by llc for bullet.bad.bc, ebp is used to store the
stack pointer

        pushl   %ebp
        movl    %esp, %ebp

but is also used as a scratch register

        movl    8(%ebp), %ebp
        movl    %ebp, (%esp)

which is fatal since ebp is used to restore the stack pointer

        movl    %ebp, %esp
        popl    %ebp
        ret

This looks like a duplicate of PR6696.  One difference is that there are
no vectors here, only operations on int and float.

*** This bug has been marked as a duplicate of bug 6696 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 13:18:41 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 13:18:41 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6782] clang accepts code missing ;
In-Reply-To: <bug-6782-206@http.llvm.org/bugs/>
References: <bug-6782-206@http.llvm.org/bugs/>
Message-ID: <20100405181842.012462A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6782

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-05 13:18:41 CDT ---
Fixed in r100443

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 13:44:47 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 13:44:47 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6780] clang++ -m32 breaks inline asm
In-Reply-To: <bug-6780-206@http.llvm.org/bugs/>
References: <bug-6780-206@http.llvm.org/bugs/>
Message-ID: <20100405184447.E3CE72A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6780

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Chris Lattner <clattner at apple.com> 2010-04-05 13:44:47 CDT ---
Fixed here, thanks:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20100405/029000.html

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 14:17:54 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 14:17:54 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6777] Code generators (both llvmc and Jit) get
 stuck when dealing circular CFG (svn r100296)
In-Reply-To: <bug-6777-206@http.llvm.org/bugs/>
References: <bug-6777-206@http.llvm.org/bugs/>
Message-ID: <20100405191755.070DA2A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6777

Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |gohman at apple.com
         Resolution|                            |FIXED

--- Comment #2 from Dan Gohman <gohman at apple.com> 2010-04-05 14:17:54 CDT ---
Fixed in 100455.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 15:11:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 15:11:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6779] ice: library memcpy does not support volatile
In-Reply-To: <bug-6779-206@http.llvm.org/bugs/>
References: <bug-6779-206@http.llvm.org/bugs/>
Message-ID: <20100405201159.D717B2A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6779

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-05 15:11:59 CDT ---
Fixed in r100457, thanks!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 15:15:26 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 15:15:26 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6785] New: MultiSource/Applications/spiff needs
 1.2G virtual memory, crashes with ulimit 300
Message-ID: <bug-6785-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6785

           Summary: MultiSource/Applications/spiff needs 1.2G virtual
                    memory, crashes with ulimit 300
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


RunSafely sets a 300MB ulimit -v, and this causes
MultiSource/Applications/spiff to crash on x86-64.

If I change the limit to 1200000, the test passes.
This is not a regression if we look at the behavior: in LLVM 2.6 it has run out
of memory too on all configurations, including native:
Out of Memory
exit 2

The difference is that now we crash when running out of memory, so I don't
think this is a release blocker.

Something should be done for llvm 2.8, either make this test use less memory,
or increase the limit in RunSafely.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 15:42:17 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 15:42:17 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6786] New: New parameter to assume function
 parameters are nonnull by default
Message-ID: <bug-6786-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6786

           Summary: New parameter to assume function parameters are
                    nonnull by default
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: tss at iki.fi
                CC: llvmbugs at cs.uiuc.edu


The nonnull attribute is kind of nice in theory and static analyzer gives
useful output with them, but since in typical code most parameters are nonnull,
it's annoying to specify it in 99% of function declarations.

So what if there was instead a new parameter that changes the default to assume
that all parameters are nonnull, unless overridden? The overriding could happen
in two ways:

a) nonnull attribute specified for the function and doesn't list some specific
parameter

b) nonnull attribute doesn't work if all parameters can be null, so there would
also have to be a new null attribute that works opposite to what nonnull does.

There are of course them potential problems with false positive warnings with
libc and other external libraries, but I'm hoping they won't cause too much
trouble.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 16:20:58 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 16:20:58 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6787] New: Unable to step through a program
 compiled with clang in gdb
Message-ID: <bug-6787-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6787

           Summary: Unable to step through a program compiled with clang
                    in gdb
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: lech.lorens+llvm at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4637)
 --> (http://llvm.org/bugs/attachment.cgi?id=4637)
source of the program

I compile a simple program displaying integers from 0 to 2. If I try to single
step through the program in gdb, I am unable to do so: gdb steps over the main
function. This does not happen if the same program is compiled with gcc.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 16:32:13 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 16:32:13 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6788] New: Missing tail-call optimization
Message-ID: <bug-6788-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6788

           Summary: Missing tail-call optimization
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: john at bass-software.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4641)
 --> (http://llvm.org/bugs/attachment.cgi?id=4641)
Code piece resulting in missing tail-optimization

Code piece:

--8<--
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-n32"
target triple = "armv7-unknown-eabi"

define arm_aapcscc i32 @test2(i32 %low, i32 %high) nounwind readnone {
entry:
  %conv = zext i32 %high to i64                   ; <i64> [#uses=1]
  %shl = shl i64 %conv, 32                        ; <i64> [#uses=1]
  %conv2 = zext i32 %low to i64                   ; <i64> [#uses=1]
  %add = or i64 %shl, %conv2                      ; <i64> [#uses=1]
  %div = udiv i64 %add, 3                         ; <i64> [#uses=1]
  %conv4 = trunc i64 %div to i32                  ; <i32> [#uses=1]
  ret i32 %conv4
}
--8<--

gets compiled as (see test-indirect.ll attachment):

--8<--
test2:
    stmdb    sp!, {r11, lr}
    mov    r2, #3
    mov    r3, #0
    bl    __udivdi3
    ldmia    sp!, {r11, pc}
--8<--

Can't this become:

--8<--
test2:
    mov    r2, #3
    mov    r3, #0
    b    __udivdi3
--8<--

Aside: in case it is important, the .ll code piece came from compiling using
clang/llvm-gcc (tried both):

--8<--
#include <stdint.h>

uint32_t
test2 (uint32_t low, uint32_t high)
{
  uint64_t sum = ((uint64_t)high << 32) + (uint64_t)low;
  return sum / 3ULL;
}
--8<--

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 18:23:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 18:23:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6789] New: clang shouldn't check for the existence
 of files specified as arguments to -Xlinker, -Xassembler, etc.
Message-ID: <bug-6789-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6789

           Summary: clang shouldn't check for the existence of files
                    specified as arguments to -Xlinker, -Xassembler, etc.
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Driver
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: mark.lacey at intel.com
                CC: llvmbugs at cs.uiuc.edu


Something like:
  clang -o hello hello.c -Xlinker -map hello.map
fails with:
  clang: error: no such file or directory: 'hello.map'
if 'hello.map' does not already exist.

Doing:
  touch hello.map
works around the issue, but it's not very convenient.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 21:46:01 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 21:46:01 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6726] [2.7] Analysis/retain-release.m fail in
	release-asserts
In-Reply-To: <bug-6726-206@http.llvm.org/bugs/>
References: <bug-6726-206@http.llvm.org/bugs/>
Message-ID: <20100406024601.7C9772A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6726

Ted Kremenek <kremenek at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #10 from Ted Kremenek <kremenek at apple.com> 2010-04-05 21:46:00 CDT ---
Zhongxing and I have investigated the bug and it does indeed appear to be the
same issue as PR 6725.

*** This bug has been marked as a duplicate of bug 6725 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr  5 22:03:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon,  5 Apr 2010 22:03:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6725] Analysis/malloc.c fail in release-asserts
In-Reply-To: <bug-6725-206@http.llvm.org/bugs/>
References: <bug-6725-206@http.llvm.org/bugs/>
Message-ID: <20100406030359.B35402A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6725

Zhongxing Xu <xuzhongxing at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #9 from Zhongxing Xu <xuzhongxing at gmail.com> 2010-04-05 22:03:59 CDT ---
Fixed in r100516. The copy constructor somehow gets elided during inlining.
This only shows up with GCC 4.4.1 Release-Asserts build.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 02:14:38 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 02:14:38 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6790] New: segmentation failure in static
	initialization
Message-ID: <bug-6790-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6790

           Summary: segmentation failure in static initialization
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: llvm-g++
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: victor.zverovich at googlemail.com
                CC: llvmbugs at cs.uiuc.edu


The following test program compiled with llvm-g++ from 2.7 prerelease 1
terminates with a segmentation failure:

#include "llvm/Support/CommandLine.h"
static llvm::cl::opt<bool>
    printStats("print-stats",
               llvm::cl::desc("Print performance metrics and statistics"),
               llvm::cl::Hidden);
int main() {}


Call stack:

#0  0x00000000 in ?? ()
#1  0x08054fac in __static_initialization_and_destruction_0 (__initialize_p=1,
__priority=65535) at /usr/local/src/llvm-2.7/lib/Support/Debug.cpp:41
#2  0x0805517e in global constructors keyed to Debug.cpp () at
/usr/local/src/llvm-2.7/lib/Support/Debug.cpp:128
#3  0x08076ae2 in __do_global_ctors_aux ()
#4  0x0807dc94 in _init ()
#5  0x08076999 in __libc_csu_init ()
#6  0x00297af4 in __libc_start_main (main=0x80481d0 <main>, argc=1,
ubp_av=0xbffff514, init=0x8076980 <__libc_csu_init>, fini=0x8076970
<__libc_csu_fini>, 
    rtld_fini=0x11dd20 <_dl_fini>, stack_end=0xbffff50c) at libc-start.c:179
#7  0x08048161 in _start () at ../sysdeps/i386/elf/start.S:119


The generated bitcode file is attached.

The following command is used to produce an executable:

llvm-g++ -use-gold-plugin test.bc
`/usr/local/src/llvm-2.7/install/debug/bin/llvm-config --ldflags --libs core`
-o test

Versions:
llvm-g++ (GCC) 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2.7)
  llvm version 2.7svn
  Optimized build with assertions.
  Built Apr  6 2010 (06:51:08).
  Host: i386-pc-linux-gnu
  Host CPU: yonah

The segmentation failure only occurs when the program uses debug LLVM libraries
(configured with --disable-optimized).

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 02:35:15 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 02:35:15 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6791] New: rev 99053 broke string escapes for
	scan-build
Message-ID: <bug-6791-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6791

           Summary: rev 99053 broke string escapes for scan-build
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: uspoerlein at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Hi,

preprocessor defines like -DPLATFORM_CONFIG_H=\"config_freebsd.h\" are no
longer passed to the static analyzer correctly. This has been broken in rev
99053. Before, the FreeBSD world sources could be analyzed correctly, now it is
erroring out for lots of the tools, observe:


/data/src/llvm/Release/bin/ccc-analyzer -O2 -pipe 
-DPLATFORM_CONFIG_H=\"config_freebsd.h\"
-I/data/obj/data/src/freebsd-head/lib/libarchive -DWITH_OPENSSL -std=gnu99
-fstack-protector -Wsystem-headers -Werror -Wall -Wno-format-y2k -W
-Wno-unused-parameter -Wstrict-prototypes -Wmissing-prototypes -Wpointer-arith
-Wreturn-type -Wcast-qual -Wwrite-strings -Wswitch -Wshadow -Wcast-align
-Wunused-parameter -Wchar-subscripts -Winline -Wnested-externs
-Wredundant-decls -Wno-pointer-sign -c
/data/src/freebsd-head/lib/libarchive/archive_check_magic.c
In file included from <built-in>:111:
<command line>:2:28: error: missing terminating '"' character
#define PLATFORM_CONFIG_H \"config_freebsd.h\"
                           ^
In file included from
/data/src/freebsd-head/lib/libarchive/archive_check_magic.c:26:
/data/src/freebsd-head/lib/libarchive/archive_platform.h:46:10: error: expected
"FILENAME" or <FILENAME>
#include PLATFORM_CONFIG_H
         ^

I'm not sure what triggered ddunbar to change this, but it's breaking the
scan-world <-> clang interaction.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 06:13:16 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 06:13:16 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6792] New: fails to build with ICC: multiple def.
 of HandleSDNode
Message-ID: <bug-6792-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6792

           Summary: fails to build with ICC: multiple def. of HandleSDNode
           Product: new-bugs
           Version: unspecified
          Platform: PC
               URL: https://wwws.clamav.net/bugzilla/show_bug.cgi?id=1926
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


icc-11.1.069 fails to build LLVM 2.7, with lots of errors about HandleSDNode
being multiply defined:

c++/.libs/libclamavcxx.a(AsmPrinter.o): In function
`llvm::HandleSDNode::HandleSDNode(llvm::SDValue)':
llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp:(.text+0x8b70): multiple definition
of `llvm::HandleSDNode::HandleSDNode(llvm::SDValue)'
c++/.libs/libclamavcxx.a(MachineFunction.o):llvm/lib/CodeGen/MachineFunction.cpp:(.text+0x3720):
first defined here

See ClamAV build log (which builds LLVM as part of its own build):
https://wwws.clamav.net/bugzilla/attachment.cgi?id=1250

HandleSDNode is declared like this:
  // FIXME: Remove the "noinline" attribute once <rdar://problem/5852746> is
  // fixed.
#ifdef __GNUC__
  explicit __attribute__((__noinline__)) HandleSDNode(SDValue X)
#else
  explicit HandleSDNode(SDValue X)
#endif

I don't know what the rdar:// is about, but is it related to this icc issue?

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 06:41:16 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 06:41:16 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6793] New: clang: better diagnostic for member
 function declaration typo
Message-ID: <bug-6793-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6793

           Summary: clang: better diagnostic for member function
                    declaration typo
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Testcase:
class A {
  int foo();
};
int A::foo();
{
  return 4;
}

clang++ outputs:
x.cpp:4:8: error: out-of-line declaration of a member must be a definition
int A::foo();
    ~~~^
x.cpp:5:1: error: expected unqualified-id
{
^
2 diagnostics generated.

It'd be better if it'd recognize the ';' typo (or rather copy-pasto), and give
a better hint at whats wrong.

It could recognize ';' followed by '{' as a typo.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 09:12:00 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 09:12:00 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6794] New: Assertion failed: ((Result ||
 D->isInvalidDecl()) && "declaration was not instantiated in this scope!")
Message-ID: <bug-6794-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6794

           Summary: Assertion failed: ((Result || D->isInvalidDecl()) &&
                    "declaration was not instantiated in this scope!")
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ich at az2000.de
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Assertion failed: ((Result || D->isInvalidDecl()) && "declaration was not
instantiated in this scope!"), function getInstantiationOf, file Sema.h, line
3566.
0  clang             0x00f342fe main + 15916632
1  clang             0x00f34744 main + 15917726
2  libSystem.B.dylib 0x94cbd2bb _sigtramp + 43
3  libSystem.B.dylib 0xffffffff _sigtramp + 1798581615
4  libSystem.B.dylib 0x94d3123a raise + 26
5  libSystem.B.dylib 0x94d3d679 abort + 73
6  libSystem.B.dylib 0x94d323db __assert_rtn + 101
7  clang             0x003a5a27 main + 3798401
8  clang             0x0039f621 main + 3772795
9  clang             0x0037166e main + 3584456
10 clang             0x0038b7e8 main + 3691330
11 clang             0x00374868 main + 3597250
12 clang             0x0038821e main + 3677560
13 clang             0x0038d4fd main + 3698775
14 clang             0x003886cd main + 3678759
15 clang             0x0038cf57 main + 3697329
16 clang             0x0038d48b main + 3698661
17 clang             0x00388ee4 main + 3680830
18 clang             0x0038dc46 main + 3700640
19 clang             0x0038873a main + 3678868
20 clang             0x00392ca9 main + 3721219
21 clang             0x00391256 main + 3714480
22 clang             0x00391d13 main + 3717229
23 clang             0x00392228 main + 3718530
24 clang             0x003910a2 main + 3714044
25 clang             0x00374950 main + 3597482
26 clang             0x003a3bf5 main + 3790671
27 clang             0x003a3538 main + 3788946
28 clang             0x003a3d5b main + 3791029
29 clang             0x003a3538 main + 3788946
30 clang             0x003a3d85 main + 3791071
31 clang             0x003a3538 main + 3788946
32 clang             0x00227242 main + 2231708
33 clang             0x005f00e7 main + 6200385
34 clang             0x002254e3 main + 2224189
35 clang             0x0003fdb4 main + 235790
36 clang             0x0003fca4 main + 235518
37 clang             0x00020b25 main + 108159
38 clang             0x00002cbb _mh_execute_header + 7355
39 clang             0x0000658f main + 233
40 clang             0x00001b36 _mh_execute_header + 2870
Stack dump:
0.    Program arguments:
/Users/az/Programmierung/CppIDE/llvm-src/Debug/bin/clang -cc1 -triple
i386-apple-darwin9.0.0 -S -disable-free -main-file-name gusgame.cpp -pic-level
1 -mdisable-fp-elim -target-cpu yonah -g -resource-dir
/Users/az/Programmierung/CppIDE/llvm-src/Debug/lib/clang/1.5 -DDEBUG=1
-D_AI_DEBUG -D SYSTEM_DATA_DIR="/usr/share/games"
-I/Users/az/Programmierung/openlierox/./optional-includes/generated
-I/Users/az/Programmierung/openlierox/./include
-I/Users/az/Programmierung/openlierox/./src
-I/Users/az/Programmierung/openlierox/./libs/pstreams -I/usr/include/libxml2
-I/usr/local/include/libxml2
-I/Users/az/Programmierung/openlierox/./libs/breakpad/src
-I/Users/az/Programmierung/openlierox/./libs/hawknl/include
-I/Users/az/Programmierung/openlierox/./libs/libzip
-I/Users/az/Programmierung/openlierox/./libs/lua
-I/Users/az/Programmierung/openlierox/./build/Xcode/include
-I/Users/az/Programmierung/openlierox/./build/Xcode/freealut/include
-I/Library/Frameworks/SDL.framework/Headers
-I/Library/Frameworks/SDL_image.framework/Headers
-I/Library/Frameworks/SDL_mixer.framework/Headers
-I/Library/Frameworks/UnixImageIO.framework/Headers
-I/Library/Frameworks/GD.framework/Headers -F ./build/Xcode
-F/Library/Frameworks -Wall -fmessage-length 102 -pthread -fexceptions
-fdiagnostics-show-option -fcolor-diagnostics -o
/var/folders/P9/P9wWY+Z8HDeOPixVmDZ0J++++TI/-Tmp-/cc-nizA45.s -x c++
/Users/az/Programmierung/openlierox/src/gusanos/gusgame.cpp 
1.    <eof> parser at end of file
2.    /Users/az/Programmierung/openlierox/src/gusanos/OlxVariable.h:61:12:
instantiating function definition '__OlxVariable<float, getFct,
putFct>::gusVar'
3.    /Users/az/Programmierung/openlierox/src/gusanos/OlxVariable.h:42:12:
instantiating function definition '__OlxVariable<float, getFct,
putFct>::gusVar'
4.    /Users/az/Programmierung/openlierox/src/gusanos/OlxVariable.h:48:24:
instantiating function definition '__OlxVariable<float, getFct,
putFct>::gusVar(std::string const &)::GusVarWrapper::invoke'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)
make[2]: *** [CMakeFiles/openlierox.dir/src/gusanos/gusgame.o] Error 250
make[1]: *** [CMakeFiles/openlierox.dir/all] Error 2
make: *** [all] Error 2

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 09:14:06 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 09:14:06 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6795] New: clang/opt -O2 makes re2c crash on startup
Message-ID: <bug-6795-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6795

           Summary: clang/opt -O2 makes re2c crash on startup
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4647)
 --> (http://llvm.org/bugs/attachment.cgi?id=4647)
bugpoint-passinput

When compiling re2c with clang++ -O2 it crashes on startup like this:
#0  0x00000038ea66ae39 in std::locale::operator=(std::locale const&) () from
/usr/lib/libstdc++.so.6
#1  0x00000038ea66a37f in std::ios_base::_M_init() () from
/usr/lib/libstdc++.so.6
#2  0x00000038ea67b199 in std::basic_ios<char, std::char_traits<char>
>::init(std::basic_streambuf<char, std::char_traits<char
> >*) () from /usr/lib/libstdc++.so.6
#3  0x000000000041c349 in re2c::basic_ifstream_lc<char, std::char_traits<char>
>::basic_ifstream_lc() ()
#4  0x000000000041a99a in main () at main.cc:394

To reproduce: compile re2c with clang++ -O2, and run ./re2c foo

Small C++ testcase, crashes on startup with clang++ -O2, but not with clang++
-O1 or -O0:
#include <fstream>
using namespace std;
typedef unsigned int uint;
template<
    class _BaseStream>
class basic_fstream_lc
    : public _BaseStream
{
public:
    basic_fstream_lc()
    {
    }

    virtual ~basic_fstream_lc()
    {
    }

    uint get_line() const
    {
    }

protected:
    mutable void *mybuf;
};
class basic_ifstream_lc
    : public basic_fstream_lc<std::basic_istream<char> >
{
};
int main()
{
    class basic_ifstream_lc source;
}

Program received signal SIGSEGV, Segmentation fault.
0x0000003b3a475989 in *__GI___libc_free (mem=<value optimized out>) at
malloc.c:3724
3724    malloc.c: No such file or directory.
        in malloc.c
(gdb) bt
#0  0x0000003b3a475989 in *__GI___libc_free (mem=<value optimized out>) at
malloc.c:3724
#1  0x00000038ea6694bb in std::ios_base::~ios_base() () from
/usr/lib/libstdc++.so.6
#2  0x0000000000400c82 in main () at foo.cc:30

Attaching bugpoint reduced testcase:
*** Found miscompiling pass: -inline
Emitted bitcode to 'bugpoint-passinput.bc'

*** You can reproduce the problem with: opt bugpoint-passinput.bc -inline

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 11:43:08 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 11:43:08 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6789] clang shouldn't check for the existence of
 files specified as arguments to -Xlinker, -Xassembler, etc.
In-Reply-To: <bug-6789-206@http.llvm.org/bugs/>
References: <bug-6789-206@http.llvm.org/bugs/>
Message-ID: <20100406164308.7AC3F2A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6789

Mark Lacey <mark.lacey at intel.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME

--- Comment #2 from Mark Lacey <mark.lacey at intel.com> 2010-04-06 11:43:08 CDT ---
Yes, you're right. I hadn't realized that the -X options worked this way, but
it makes sense.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 11:51:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 11:51:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6796] New: Varargs on amd64 generates very strange
	code
Message-ID: <bug-6796-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6796

           Summary: Varargs on amd64 generates very strange code
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ed at 80386.nl
                CC: llvmbugs at cs.uiuc.edu


The following code generates very different code when using GCC 4.2.1/Clang
SVN. The generated code also makes little sense.

int
foo(int a, ...)
{
    int r;
    __builtin_va_list va;

    __builtin_va_start(va, a);
    r = __builtin_va_arg(va, int);
    __builtin_va_end(va);
    return (r);
}

GCC:

0000000000000000 <foo>:
   0:    48 83 ec 60              sub    $0x60,%rsp
   4:    48 8d 44 24 68           lea    0x68(%rsp),%rax
   9:    48 89 74 24 b0           mov    %rsi,0xffffffffffffffb0(%rsp)
   e:    c7 44 24 88 10 00 00     movl   $0x10,0xffffffffffffff88(%rsp)
  15:    00 
  16:    48 89 44 24 90           mov    %rax,0xffffffffffffff90(%rsp)
  1b:    48 8d 44 24 a8           lea    0xffffffffffffffa8(%rsp),%rax
  20:    48 89 44 24 98           mov    %rax,0xffffffffffffff98(%rsp)
  25:    48 83 c0 08              add    $0x8,%rax
  29:    8b 00                    mov    (%rax),%eax
  2b:    48 83 c4 60              add    $0x60,%rsp
  2f:    c3                       retq   

Clang:

0000000000000000 <foo>:
   0:    55                       push   %rbp
   1:    48 89 e5                 mov    %rsp,%rbp
   4:    48 83 ec 50              sub    $0x50,%rsp
   8:    84 c0                    test   %al,%al
   a:    74 26                    je     32 <foo+0x32>
   c:    0f 29 85 60 ff ff ff     movaps %xmm0,0xffffffffffffff60(%rbp)
  13:    0f 29 8d 70 ff ff ff     movaps %xmm1,0xffffffffffffff70(%rbp)
  1a:    0f 29 55 80              movaps %xmm2,0xffffffffffffff80(%rbp)
  1e:    0f 29 5d 90              movaps %xmm3,0xffffffffffffff90(%rbp)
  22:    0f 29 65 a0              movaps %xmm4,0xffffffffffffffa0(%rbp)
  26:    0f 29 6d b0              movaps %xmm5,0xffffffffffffffb0(%rbp)
  2a:    0f 29 75 c0              movaps %xmm6,0xffffffffffffffc0(%rbp)
  2e:    0f 29 7d d0              movaps %xmm7,0xffffffffffffffd0(%rbp)
  32:    4c 89 8d 58 ff ff ff     mov    %r9,0xffffffffffffff58(%rbp)
  39:    4c 89 85 50 ff ff ff     mov    %r8,0xffffffffffffff50(%rbp)
  40:    48 89 8d 48 ff ff ff     mov    %rcx,0xffffffffffffff48(%rbp)
  47:    48 89 95 40 ff ff ff     mov    %rdx,0xffffffffffffff40(%rbp)
  4e:    48 89 b5 38 ff ff ff     mov    %rsi,0xffffffffffffff38(%rbp)
  55:    48 8d 85 30 ff ff ff     lea    0xffffffffffffff30(%rbp),%rax
  5c:    48 89 45 f8              mov    %rax,0xfffffffffffffff8(%rbp)
  60:    48 8d 45 10              lea    0x10(%rbp),%rax
  64:    48 89 45 f0              mov    %rax,0xfffffffffffffff0(%rbp)
  68:    c7 45 ec 30 00 00 00     movl   $0x30,0xffffffffffffffec(%rbp)
  6f:    c7 45 e8 08 00 00 00     movl   $0x8,0xffffffffffffffe8(%rbp)
  76:    48 63 45 e8              movslq 0xffffffffffffffe8(%rbp),%rax
  7a:    48 83 f8 28              cmp    $0x28,%rax
  7e:    77 0f                    ja     8f <foo+0x8f>
  80:    48 89 c1                 mov    %rax,%rcx
  83:    48 03 4d f8              add    0xfffffffffffffff8(%rbp),%rcx
  87:    83 c0 08                 add    $0x8,%eax
  8a:    89 45 e8                 mov    %eax,0xffffffffffffffe8(%rbp)
  8d:    eb 0c                    jmp    9b <foo+0x9b>
  8f:    48 8b 4d f0              mov    0xfffffffffffffff0(%rbp),%rcx
  93:    48 8d 41 08              lea    0x8(%rcx),%rax
  97:    48 89 45 f0              mov    %rax,0xfffffffffffffff0(%rbp)
  9b:    8b 01                    mov    (%rcx),%eax
  9d:    48 83 c4 50              add    $0x50,%rsp
  a1:    5d                       pop    %rbp
  a2:    c3                       retq

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:00:34 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:00:34 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2438] compiling gfortran.dg/hollerith.f90 testcase
 causes llvm-gfortran compiler to bus-error
In-Reply-To: <bug-2438-206@http.llvm.org/bugs/>
References: <bug-2438-206@http.llvm.org/bugs/>
Message-ID: <20100406170034.ABE372A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2438

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-06 12:00:34 CDT ---
Only need on gfortran bug.

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:00:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:00:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2439] gfortran.dg/namelist_13.f90 causes malloc
	errors in f951
In-Reply-To: <bug-2439-206@http.llvm.org/bugs/>
References: <bug-2439-206@http.llvm.org/bugs/>
Message-ID: <20100406170051.5F12D2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2439

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #3 from Chris Lattner <clattner at apple.com> 2010-04-06 12:00:51 CDT ---
only need one gfortran bug.

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:01:05 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:01:05 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2440] gfortran.dg/optional_assumed_charlen_1.f90
 causes internal compiler error
In-Reply-To: <bug-2440-206@http.llvm.org/bugs/>
References: <bug-2440-206@http.llvm.org/bugs/>
Message-ID: <20100406170105.A9C772A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2440

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #4 from Chris Lattner <clattner at apple.com> 2010-04-06 12:01:05 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:01:16 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:01:16 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2441] gfortran.dg/pr21177.f90 causes an internal
 compiler error: Bus error
In-Reply-To: <bug-2441-206@http.llvm.org/bugs/>
References: <bug-2441-206@http.llvm.org/bugs/>
Message-ID: <20100406170116.B7EC42A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2441

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-06 12:01:16 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:01:28 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:01:28 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2442] gfortran.dg/specifics_1.f90 testcase compiled
 with -ff2c crashes the llvm-gfortran compiler
In-Reply-To: <bug-2442-206@http.llvm.org/bugs/>
References: <bug-2442-206@http.llvm.org/bugs/>
Message-ID: <20100406170128.40DC22A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2442

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #3 from Chris Lattner <clattner at apple.com> 2010-04-06 12:01:27 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:01:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:01:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2430] PR26491 test in
 gfortran.dg/actual_array_constructor_1.f90 fails above -O0.
In-Reply-To: <bug-2430-206@http.llvm.org/bugs/>
References: <bug-2430-206@http.llvm.org/bugs/>
Message-ID: <20100406170156.27F752A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2430

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-06 12:01:55 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:02:11 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:02:11 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2444] gfortran.fortran-torture/execute/nullarg.f90
 causes an internal compiler error in llvm-gfortran
In-Reply-To: <bug-2444-206@http.llvm.org/bugs/>
References: <bug-2444-206@http.llvm.org/bugs/>
Message-ID: <20100406170211.72CFF2A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2444

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #7 from Chris Lattner <clattner at apple.com> 2010-04-06 12:02:10 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:02:21 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:02:21 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2445]
 gfortran.fortran-torture/execute/equiv_init_1.f90 fails execution test when
 compiled with llvm-gfortran
In-Reply-To: <bug-2445-206@http.llvm.org/bugs/>
References: <bug-2445-206@http.llvm.org/bugs/>
Message-ID: <20100406170221.9FB832A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2445

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-06 12:02:21 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:02:30 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:02:30 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2446] gfortran.dg/g77/intrinsic-unix-bessel.f fails
 execution tests when compiled with llvm-gfortran
In-Reply-To: <bug-2446-206@http.llvm.org/bugs/>
References: <bug-2446-206@http.llvm.org/bugs/>
Message-ID: <20100406170230.636972A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2446

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-06 12:02:30 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:03:38 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:03:38 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2447] gfortran.dg/g77/980628-9.f fails execution
 test when compiled with llvm-gfortran
In-Reply-To: <bug-2447-206@http.llvm.org/bugs/>
References: <bug-2447-206@http.llvm.org/bugs/>
Message-ID: <20100406170338.BD3712A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2447

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-06 12:03:38 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:03:52 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:03:52 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2448] gfortran.dg/g77/980628-3.f fails execution
 test when compiled with llvm-gfortran.
In-Reply-To: <bug-2448-206@http.llvm.org/bugs/>
References: <bug-2448-206@http.llvm.org/bugs/>
Message-ID: <20100406170352.0D68E2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2448

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #3 from Chris Lattner <clattner at apple.com> 2010-04-06 12:03:51 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:04:02 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:04:02 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2449] gfortran.dg/g77/980628-2.f fails execution
 test when compiled with llvm-gfortran
In-Reply-To: <bug-2449-206@http.llvm.org/bugs/>
References: <bug-2449-206@http.llvm.org/bugs/>
Message-ID: <20100406170402.774122A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2449

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-06 12:04:02 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:04:12 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:04:12 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2450] gfortran.dg/g77/980628-10.f testcase fails
 execution test when compiled with llvm-gfortran
In-Reply-To: <bug-2450-206@http.llvm.org/bugs/>
References: <bug-2450-206@http.llvm.org/bugs/>
Message-ID: <20100406170412.4E33F2A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2450

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-06 12:04:11 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:04:23 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:04:23 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2457] gfortran.dg/assumed_charlen_function_5.f90
 has multiple compile failures
In-Reply-To: <bug-2457-206@http.llvm.org/bugs/>
References: <bug-2457-206@http.llvm.org/bugs/>
Message-ID: <20100406170423.BDD142A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2457

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-06 12:04:23 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:04:33 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:04:33 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2458] gfortran.dg/char_transpose_1.f90 fails to
 compile under llvm-gfortran
In-Reply-To: <bug-2458-206@http.llvm.org/bugs/>
References: <bug-2458-206@http.llvm.org/bugs/>
Message-ID: <20100406170433.5EA132A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2458

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-06 12:04:33 CDT ---
only need one gfortran bug

*** This bug has been marked as a duplicate of bug 2431 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 12:49:50 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 12:49:50 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6654] llvm-test regressions with 2.7
In-Reply-To: <bug-6654-206@http.llvm.org/bugs/>
References: <bug-6654-206@http.llvm.org/bugs/>
Message-ID: <20100406174950.714C92A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6654

Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #10 from Duncan Sands <baldrick at free.fr> 2010-04-06 12:49:49 CDT ---
The SingleSource/Regression/C/2004-03-15-IndirectGoto failure is:

lib/ExecutionEngine/JIT/JIT.h:131: virtual void*
llvm::JIT::getPointerToBasicBlock(llvm::BasicBlock*): Assertion `0 && "JIT does
not support address-of-label yet!"' failed

This is a duplicate of PR6744.  Since this was the only remaining issue tracked
here, I'm closing this bug.

*** This bug has been marked as a duplicate of bug 6744 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 14:05:23 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 14:05:23 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6797] New: Patches to port llvm to Minix 3,
	improve portability
Message-ID: <bug-6797-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6797

           Summary: Patches to port llvm to Minix 3, improve portability
           Product: new-bugs
           Version: trunk
          Platform: Other
        OS/Version: other
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: reeuwijk at few.vu.nl
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4649)
 --> (http://llvm.org/bugs/attachment.cgi?id=4649)
The patch described above

The attached patch against a recent llvm trunk version adds basic support for
Minix 3 to llvm. It adds Minix as a recognized platform and target to
configure.ac, adds configuration for Minix at a few other places, and honors
HAVE_DLFCN_H in DynamicLibrary.cpp.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 14:12:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 14:12:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6798] New: missed optimization on byval argument
	with memcpy
Message-ID: <bug-6798-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6798

           Summary: missed optimization on byval argument with memcpy
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: quickslyver at free.fr
                CC: llvmbugs at cs.uiuc.edu


in the generated code bellow, I think the memcpy is useless because the
argument of the function f is passed "byval"

------------------------------------------------------------------------------
$ cat bug.c
struct a { int v[500];};
int f(struct a);

void g()
{
    struct a x={{0,1,2,3,4,5}};
    f(x);
}

-----------------------------------------------------------------------------

$clang bug.c -S -emit-llvm -o - -O3
; ModuleID = 'bug.c'
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32-n8:16:32"
target triple = "i386-pc-linux-gnu"

@g.x = internal constant %struct.a { 
[...]

define void @g() nounwind {
entry:
  %agg.tmp = alloca %struct.a, align 4            ; <%struct.a*> [#uses=2]
  %agg.tmp.04 = bitcast %struct.a* %agg.tmp to i8* ; <i8*> [#uses=1]
  call void @llvm.memcpy.p0i8.p0i8.i32(i8* %agg.tmp.04, i8* bitcast (%struct.a*
@g.x to i8*), i32 2000, i32 4, i1 false)
  %call = call i32 @f(%struct.a* byval %agg.tmp) nounwind ; <i32> [#uses=0]
  ret void
}

declare i32 @f(%struct.a* byval)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 14:42:22 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 14:42:22 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6791] rev 99053 broke string escapes for scan-build
In-Reply-To: <bug-6791-206@http.llvm.org/bugs/>
References: <bug-6791-206@http.llvm.org/bugs/>
Message-ID: <20100406194222.E48CD2A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6791

Ted Kremenek <kremenek at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 14:43:21 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 14:43:21 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6771] Quoted -D macros break static analyzer
In-Reply-To: <bug-6771-206@http.llvm.org/bugs/>
References: <bug-6771-206@http.llvm.org/bugs/>
Message-ID: <20100406194321.14AB12A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6771

Ted Kremenek <kremenek at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |kremenek at apple.com
         Resolution|                            |FIXED

--- Comment #1 from Ted Kremenek <kremenek at apple.com> 2010-04-06 14:43:20 CDT ---
Should be now fixed in r100551.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 15:09:34 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 15:09:34 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6792] fails to build with ICC: multiple def. of
	HandleSDNode
In-Reply-To: <bug-6792-206@http.llvm.org/bugs/>
References: <bug-6792-206@http.llvm.org/bugs/>
Message-ID: <20100406200934.2AC0C2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6792

Bill Wendling <wendling at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #5 from Bill Wendling <wendling at apple.com> 2010-04-06 15:09:33 CDT ---
I made the hack more specific with 100556. Please check to see if that fixes
the icc problem.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 15:30:46 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 15:30:46 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6696] ebp wrongly clobbered - causes
 SingleSource/Regression/C/casts.c to crash
In-Reply-To: <bug-6696-206@http.llvm.org/bugs/>
References: <bug-6696-206@http.llvm.org/bugs/>
Message-ID: <20100406203046.513732A6C12F@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6696

Jim Grosbach <grosbach at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #29 from Jim Grosbach <grosbach at apple.com> 2010-04-06 15:30:44 CDT ---
Fixed in r100559.

This brings back some of the old logic, just in a more restricted form to
retain the benefits as much as possible.

I hand-inspected the generated assembly for the testcases here and in 6663 and
verified that EBP is not used as a scratch register for these functions. I also
checked some functions that do not use higher alignment registers and verified
that EBP is still being allocated normally there (the test suite has examples
of this).

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 15:51:06 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 15:51:06 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6799] New: clang -O2 emits wrong asm
Message-ID: <bug-6799-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6799

           Summary: clang -O2 emits wrong asm
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


when compiling the attached gcc produces this code:

gcc -S -O2 tree1.c:
getcnode:
.LFB6:
        movq    %rbx, -24(%rsp)
.LCFI0:
        movq    %r12, -16(%rsp)
.LCFI1:
        movq    %rsi, %r12
        movq    %r13, -8(%rsp)
.LCFI2:
        subq    $24, %rsp
.LCFI3:
        movq    %rdi, %r13
        call    getnode
        movl    $32, %edi
        movq    %rax, %rbx
        movq    %r13, 8(%rax)
        movl    $47, (%rax)
        call    tgetblk
        movq    %rax, 24(%rbx)
        movl    (%r13), %edx
        movq    %r12, %rdi
        movl    %edx, (%rax)
        movl    4(%r12), %eax
        movq    24(%rbx), %rdx
        movl    %eax, 4(%rdx)
        movq    16(%r12), %rax
        movq    24(%rbx), %rdx
        movq    %rax, 16(%rdx)
        movq    24(%r12), %rax              (1)
        movq    %rax, 24(%rdx)              (2)
        call    free
        movq    %rbx, %rax
        movq    8(%rsp), %r12
        movq    (%rsp), %rbx
        movq    16(%rsp), %r13
        addq    $24, %rsp
        ret

while clang -S -O2 tree1.c:

getcnode:
.Leh_func_begin1:
        pushq   %rbp
.Ltmp0:
        movq    %rsp, %rbp
.Ltmp1:
        pushq   %r15
        pushq   %r14
        pushq   %rbx
        subq    $8, %rsp
.Ltmp2:
        movq    %rsi, %rbx
        movq    %rdi, %r14
        callq   getnode
        movq    %rax, %r15
        movl    $47, (%r15)
        movq    %r14, 8(%r15)
        movl    $32, %edi
        callq   tgetblk
        movq    %rax, 24(%r15)
        movl    (%r14), %ecx
        movl    %ecx, (%rax)
        movq    24(%r15), %rax
        movl    4(%rbx), %ecx
        movl    %ecx, 4(%rax)
        movq    24(%r15), %rax
        movaps  16(%rbx), %xmm0           (3)
        movaps  %xmm0, 16(%rax)           (4)
        movq    %rbx, %rdi
        callq   free
        movq    %r15, %rax
        addq    $8, %rsp
        popq    %rbx
        popq    %r14
        popq    %r15
        popq    %rbp
        ret

note that the offset for (1) and (2) is 24 for gcc while 16 for clang at (3)
and (4), this makes "lint" crash

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 16:31:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 16:31:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6616] Warnings building LLVM 2.7
In-Reply-To: <bug-6616-206@http.llvm.org/bugs/>
References: <bug-6616-206@http.llvm.org/bugs/>
Message-ID: <20100406213103.CD2BD2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6616

David Greene <greened at obbligato.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #14 from David Greene <greened at obbligato.org> 2010-04-06 16:31:02 CDT ---
I'm going to resolve this FIXED in the interest of getting the release moving. 
No one is assigned to this bug.  Please let me know if you disagree.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 17:43:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 17:43:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6800] New: Missing fix-it when operator-> is
 applied to a class type
Message-ID: <bug-6800-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6800

           Summary: Missing fix-it when operator-> is applied to a class
                    type
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: quality-of-implementation
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


blackthorn:clang dgregor$ cat t.cpp 
struct X { int member; };

int f(X x) {
  return x->member;
}
blackthorn:clang dgregor$ clang++ t.cpp 
t.cpp:4:11: error: member reference type 'X' is not a pointer
  return x->member;
         ~^
1 diagnostic generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 19:02:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 19:02:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6801] New: sample clang plugin crashes
Message-ID: <bug-6801-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6801

           Summary: sample clang plugin crashes
           Product: clang
           Version: trunk
          Platform: All
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: zhanyong.wan at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Build the PrintFunctionNames sample plugin in
tools/clang/examples/PrintFunctionNames/, then run:

$ clang -cc1 -load path/to/PrintFunctionNames.so -plugin print-fns t.cc

where t.cc has content:

int f() {
  return 1;
}

and clang will crash:

Two passes with the same argument (-preverify) attempted to be registered!
UNREACHABLE executed at include/llvm/Support/PassNameParser.h:74!
0  clang                  0x00000000012975ce
1  clang                  0x00000000012974b5
2  libpthread.so.0        0x00007f9a507e3580
3  libc.so.6              0x00007f9a4ff98da5 gsignal + 53
4  libc.so.6              0x00007f9a4ff9a750 abort + 272
5  clang                  0x00000000012bd928
llvm::FoldingSetNodeID::AddPointer(void const*) + 0
6  clang                  0x0000000001213bb1
llvm::PassNameParser::passRegistered(llvm::PassInfo const*) + 229
7  clang                  0x0000000001211c36 llvm::PassInfo::registerPass() +
140
8  clang                  0x0000000000aecde3 llvm::PassInfo::PassInfo(char
const*, char const*, long, llvm::Pass* (*)(), bool, bool) + 145
9  PrintFunctionNames.so  0x00007f9a4ef3b965
10 PrintFunctionNames.so  0x00007f9a4ef3c99a
11 PrintFunctionNames.so  0x00007f9a4ef3cb19
12 PrintFunctionNames.so  0x00007f9a4ef9ae36
Stack dump:
0.    Program arguments: clang -cc1 -load path/to/PrintFunctionNames.so -plugin
print-fns t.cc 
Aborted

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 19:47:28 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 19:47:28 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6036] clang analyzer crash with signal 8
In-Reply-To: <bug-6036-206@http.llvm.org/bugs/>
References: <bug-6036-206@http.llvm.org/bugs/>
Message-ID: <20100407004728.4364C2A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6036

Ted Kremenek <kremenek at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #4 from Ted Kremenek <kremenek at apple.com> 2010-04-06 19:47:27 CDT ---
Fixed in r100594.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr  6 20:25:44 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue,  6 Apr 2010 20:25:44 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6802] New: ice: Branch should be foldable!
Message-ID: <bug-6802-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6802

           Summary: ice: Branch should be foldable!
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu, chenyang at cs.utah.edu


[regehr at gamow tmp400]$ clang -O0 -emit-llvm small.c -c -o small.bc -w
[regehr at gamow tmp400]$ opt -O1 small.bc
WARNING: You're attempting to print out a bitcode file.
This is inadvisable as it may cause display problems. If
you REALLY want to taste LLVM bitcode first-hand, you
can force output with the `-f' option.

opt: SCCP.cpp:1896: virtual bool<unnamed>::IPSCCP::runOnModule(llvm::Module&):
Assertion `BI->isConditional() && isa<UndefValue>(BI->getCondition()) &&
"Branch should be foldable!"' failed.
0  opt             0x000000000082f0af
1  opt             0x000000000082fb3a
2  libpthread.so.0 0x00007f1cde88d190
3  libc.so.6       0x00007f1cddb934b5 gsignal + 53
4  libc.so.6       0x00007f1cddb96f50 abort + 384
5  libc.so.6       0x00007f1cddb8c481 __assert_fail + 241
6  opt             0x00000000005735d2
7  opt             0x00000000007c261e
llvm::MPPassManager::runOnModule(llvm::Module&) + 494
8  opt             0x00000000007c2779 llvm::PassManagerImpl::run(llvm::Module&)
+ 137
9  opt             0x00000000004c28bc main + 2428
10 libc.so.6       0x00007f1cddb7eabd __libc_start_main + 253
11 opt             0x00000000004b6b69
Stack dump:
0.    Program arguments: opt -O1 small.bc 
1.    Running pass 'Interprocedural Sparse Conditional Constant Propagation' on
module 'small.bc'.
Aborted

[regehr at gamow tmp400]$ clang -v
clang version 1.5 (trunk 100578)
Target: x86_64-unknown-linux-gnu
Thread model: posix

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr  7 06:13:52 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed,  7 Apr 2010 06:13:52 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6704] Porting llvm to Minix
In-Reply-To: <bug-6704-206@http.llvm.org/bugs/>
References: <bug-6704-206@http.llvm.org/bugs/>
Message-ID: <20100407111352.34AB02A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6704

Kees van Reeuwijk <reeuwijk at few.vu.nl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Kees van Reeuwijk <reeuwijk at few.vu.nl> 2010-04-07 06:13:51 CDT ---
Marked as RESOLVED since I've submitted a new patch against trunk.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr  7 10:32:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed,  7 Apr 2010 10:32:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6803] New: UNREACHABLE executed at
	MachineInstr.cpp:170!
Message-ID: <bug-6803-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6803

           Summary: UNREACHABLE executed at MachineInstr.cpp:170!
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4654)
 --> (http://llvm.org/bugs/attachment.cgi?id=4654)
test case

on amd64 with clang/llvm 100612 I am getting:

pes delta$ clang -c -O2  acpi_hpet.c

.... warnings ....

Unrecognized operand type
UNREACHABLE executed at MachineInstr.cpp:170!
Stack dump:
0.      Program arguments: /usr/local/bin/clang -cc1 -triple
x86_64-unknown-freebsd8.0 -S -disable-free -main-file-name acpi_hpet.c
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -resource-dir /usr/local/lib/clang/1.5 -O2
-fmessage-length 152 -fgnu-runtime -fdiagnostics-show-option
-fcolor-diagnostics -o /tmp/cc-ZXwvTd.s -x c acpi_hpet.c 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'Control Flow Optimizer' on function '@acpi_hpet_attach'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)
pes delta$ 

this is quite recent (at most 2 weeks?)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr  7 13:05:23 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed,  7 Apr 2010 13:05:23 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6803] UNREACHABLE executed at MachineInstr.cpp:170!
In-Reply-To: <bug-6803-206@http.llvm.org/bugs/>
References: <bug-6803-206@http.llvm.org/bugs/>
Message-ID: <20100407180523.B38E92A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6803

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-07 13:05:23 CDT ---
Fixed here, thanks:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20100405/099387.html

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr  7 20:02:12 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed,  7 Apr 2010 20:02:12 -0500 (CDT)
Subject: [LLVMbugs] [Bug 3440] llvm miscompiles
	gcc.target/i386/sse4_1-mpsadbw.c
In-Reply-To: <bug-3440-206@http.llvm.org/bugs/>
References: <bug-3440-206@http.llvm.org/bugs/>
Message-ID: <20100408010212.6DD272A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=3440

Eric Christopher <echristo at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #6 from Eric Christopher <echristo at gmail.com> 2010-04-07 20:02:11 CDT ---
Fixed thusly:

http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20100405/099469.html

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  8 09:49:43 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  8 Apr 2010 09:49:43 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6804] New: alpha+linux(Debain4) BUILD ERROR
	[llvm-gcc 2.8svn]
Message-ID: <bug-6804-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6804

           Summary: alpha+linux(Debain4) BUILD ERROR [llvm-gcc 2.8svn]
           Product: new-bugs
           Version: trunk
          Platform: DEC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dot618point at gmail.com
                CC: llvmbugs at cs.uiuc.edu


/home/zheng/obj/./gcc/xgcc -B/home/zheng/obj/./gcc/
-B/usr/local/alphaev6-unknown-linux-gnu/bin/
-B/usr/local/alphaev6-unknown-linux-gnu/lib/ -isystem
/usr/local/alphaev6-unknown-linux-gnu/include -isystem
/usr/local/alphaev6-unknown-linux-gnu/sys-include  -O2  -O2 -g -O2  -mieee
-DIN_GCC    -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes
-Wold-style-definition  -isystem ./include  -fPIC -mieee -g -DHAVE_GTHR_DEFAULT
-DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED  -I. -I. -I../../llvm-gcc/gcc
-I../../llvm-gcc/gcc/. -I../../llvm-gcc/gcc/../include
-I../../llvm-gcc/gcc/../libcpp/include  -I../../llvm-gcc/gcc/../libdecnumber
-I../libdecnumber -I/usr/local/include -DL_muldi3 -fvisibility=hidden
-DHIDE_EXPORTS -c ../../llvm-gcc/gcc/libgcc2.c -o libgcc/./_muldi3.o
cc1: /home/zheng/llvm/lib/VMCore/Instructions.cpp:1762: static
llvm::BinaryOperator*
llvm::BinaryOperator::Create(llvm::Instruction::BinaryOps, llvm::Value*,
llvm::Value*, const llvm::Twine&, llvm::Instruction*): Assertion `S1->getType()
== S2->getType() && "Cannot create binary operator with two operands of
differing type!"' failed.
../../llvm-gcc/gcc/libgcc2.c: In function ?__multi3?:
../../llvm-gcc/gcc/libgcc2.c:567: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  8 10:31:36 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  8 Apr 2010 10:31:36 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6805] New: -fcatch-undefined-behavior crashes with
 signal 6, "LLVM IR generation of compound statement"
Message-ID: <bug-6805-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6805

           Summary: -fcatch-undefined-behavior crashes with signal 6,
                    "LLVM IR generation of compound statement"
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: magnus.reftel at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4657)
 --> (http://llvm.org/bugs/attachment.cgi?id=4657)
Minimized test case

I have a small test case (attached) that fails with the below error message
when building with -fcatch-undefined-behavior, but succeeds without it. I'm
using revision 100754 on cygwin.

/opt/llvm-r100754/bin/clang  -fcatch-undefined-behavior -v -c testcase.c     
clang version 1.5 (trunk 100754)
Target: i386-pc-cygwin
Thread model: posix
 "/opt/llvm-r100754/bin/clang" -cc1 -triple i386-pc-cygwin -S -disable-free
-main-file-name testcase.c -mrelocation-model static -mdisable-fp-elim
-mconstructor-aliases -target-cpu pentium4 -v -resource-dir
/opt/llvm-r100754/lib/clang/1.5 -ferror-limit 19 -fmessage-length 80
-fcatch-undefined-behavior -fgnu-runtime -fdiagnostics-show-option
-fcolor-diagnostics -o /c/DOCUME~1/qrefmag/LOCALS~1/Temp/cc-T9hI02.s -x c
testcase.c
clang -cc1 version 1.5 based upon llvm 2.8svn hosted on i386-pc-cygwin
ignoring nonexistent directory "/usr/local/include"
#include "..." search starts here:
#include <...> search starts here:
 /opt/llvm-r100754/lib/clang/1.5/include
 /usr/include
End of search list.
assertion "(i >= FTy->getNumParams() || FTy->getParamType(i) ==
Params[i]->getType()) && "Calling a function with a bad signature!"" failed:
file "Instructions.cpp", line 324, function: void
llvm::CallInst::init(llvm::Value*, llvm::Value* const*, unsigned int)
Stack dump:
0.      Program arguments: /opt/llvm-r100754/bin/clang -cc1 -triple
i386-pc-cygwin -S -disable-free -main-file-name testcase.c -mrelocation-model
static -mdisable-fp-elim -mconstructor-aliases -target-cpu pentium4 -v
-resource-dir /opt/llvm-r100754/lib/clang/1.5 -ferror-limit 19 -fmessage-length
80 -fcatch-undefined-behavior -fgnu-runtime -fdiagnostics-show-option
-fcolor-diagnostics -o /c/DOCUME~1/qrefmag/LOCALS~1/Temp/cc-T9hI02.s -x c
testcase.c
1.      <eof> parser at end of file
2.      testcase.c:6:6: LLVM IR generation of declaration 'foo'
3.      testcase.c:6:6: Generating code for declaration 'foo'
4.      testcase.c:7:1: LLVM IR generation of compound statement ('{}')
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr  8 15:26:16 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu,  8 Apr 2010 15:26:16 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6752] Assert when compiling - "Access specifier is
 AS_none inside a record decl"
In-Reply-To: <bug-6752-206@http.llvm.org/bugs/>
References: <bug-6752-206@http.llvm.org/bugs/>
Message-ID: <20100408202616.66E372A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6752

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |rjmccall at apple.com
         Resolution|                            |FIXED

--- Comment #3 from John McCall <rjmccall at apple.com> 2010-04-08 15:26:15 CDT ---
Fixed in r100806.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  9 06:40:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  9 Apr 2010 06:40:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6806] New: Codegen regression "Instruction does not
 dominate all uses!"
Message-ID: <bug-6806-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6806

           Summary: Codegen regression "Instruction does not dominate all
                    uses!"
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4659)
 --> (http://llvm.org/bugs/attachment.cgi?id=4659)
testcase .ll

The dragonegg self-host build broke recently, due to this recent regression
(occurs while building SelectionDAGBuilder).  Here's the bugpoint reduced
testcase:

$ llc domcrash.bc
Instruction does not dominate all uses!
  %uglygep78 = bitcast i8* %uglygep7 to %"struct CaseBits&" ; <%"struct
CaseBits&"> [#uses=1]
  %1 = bitcast %"struct CaseBits&" %uglygep78 to i8* ; <i8*> [#uses=2]
Instruction does not dominate all uses!
  %1 = bitcast %"struct CaseBits&" %uglygep78 to i8* ; <i8*> [#uses=2]
  %uglygep4 = getelementptr i8* %1, i64 16        ; <i8*> [#uses=1]
Instruction does not dominate all uses!
  %uglygep4 = getelementptr i8* %1, i64 16        ; <i8*> [#uses=1]
  %uglygep45 = bitcast i8* %uglygep4 to %"struct CaseBits&" ; <%"struct
CaseBits&"> [#uses=1]
Instruction does not dominate all uses!
  %uglygep45 = bitcast i8* %uglygep4 to %"struct CaseBits&" ; <%"struct
CaseBits&"> [#uses=1]
  %uglygep456 = bitcast %"struct CaseBits&" %uglygep45 to i8* ; <i8*> [#uses=1]
Instruction does not dominate all uses!
  %uglygep456 = bitcast %"struct CaseBits&" %uglygep45 to i8* ; <i8*> [#uses=1]
  %uglygep7 = getelementptr i8* %uglygep456, i64 -16 ; <i8*> [#uses=1]
Instruction does not dominate all uses!
  %uglygep7 = getelementptr i8* %uglygep456, i64 -16 ; <i8*> [#uses=1]
  %uglygep78 = bitcast i8* %uglygep7 to %"struct CaseBits&" ; <%"struct
CaseBits&"> [#uses=1]
Instruction does not dominate all uses!
  %1 = bitcast %"struct CaseBits&" %uglygep78 to i8* ; <i8*> [#uses=2]
  %uglygep = getelementptr i8* %1, i64 16         ; <i8*> [#uses=1]
Instruction does not dominate all uses!
  %uglygep = getelementptr i8* %1, i64 16         ; <i8*> [#uses=1]
  %scevgep2 = getelementptr i8* %uglygep, i64 %lsr.iv ; <i8*> [#uses=1]
Instruction does not dominate all uses!
  %scevgep2 = getelementptr i8* %uglygep, i64 %lsr.iv ; <i8*> [#uses=1]
  %scevgep23 = bitcast i8* %scevgep2 to i32*      ; <i32*> [#uses=1]
Instruction does not dominate all uses!
  %scevgep23 = bitcast i8* %scevgep2 to i32*      ; <i32*> [#uses=1]
  store i32 undef, i32* %scevgep23, align 4
Broken module found, compilation aborted!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  9 11:07:42 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  9 Apr 2010 11:07:42 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6807] New: llvm::SimplifyFortifiedLibCalls::fold
	crash
Message-ID: <bug-6807-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6807

           Summary: llvm::SimplifyFortifiedLibCalls::fold crash
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4660)
 --> (http://llvm.org/bugs/attachment.cgi?id=4660)
testcase .ll

$ llc fortified.ll
llc: llvm/lib/VMCore/Value.cpp:317: void
llvm::Value::replaceAllUsesWith(llvm::Value*): Assertion `New->getType() ==
getType() && "replaceAllUses of value with new value of different type!"'
failed.
...
llvm::SimplifyFortifiedLibCalls::fold(llvm::CallInst*, llvm::TargetData const*)
+ 2065
...
Stack dump:
0.    Program arguments: llc fortified.ll 
1.    Running pass 'Optimize for code generation' on function '@stpncpy'
Aborted (core dumped)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  9 11:08:11 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  9 Apr 2010 11:08:11 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6799] clang -O2 emits wrong asm
In-Reply-To: <bug-6799-206@http.llvm.org/bugs/>
References: <bug-6799-206@http.llvm.org/bugs/>
Message-ID: <20100409160811.18C792A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6799

Roman Divacky <rdivacky at freebsd.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #9 from Roman Divacky <rdivacky at freebsd.org> 2010-04-09 11:08:10 CDT ---
this was a bug in the original program so this bug report is invalid, closing

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  9 12:42:20 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  9 Apr 2010 12:42:20 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6778] llvm-gcc-4.2 from release_2.7 branch fails to
 build f951 on x86_64-apple-darwin10
In-Reply-To: <bug-6778-206@http.llvm.org/bugs/>
References: <bug-6778-206@http.llvm.org/bugs/>
Message-ID: <20100409174221.7F96D2A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6778

Dale Johannesen <dalej at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #28 from Dale Johannesen <dalej at apple.com> 2010-04-09 12:42:20 CDT ---
Good enough for me.  Tanya, I think this can go in 2.7, I believe the protocol
is that you're the only one who does that?  It's 100823.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  9 14:15:28 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  9 Apr 2010 14:15:28 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6806] Codegen regression "Instruction does not
 dominate all uses!"
In-Reply-To: <bug-6806-206@http.llvm.org/bugs/>
References: <bug-6806-206@http.llvm.org/bugs/>
Message-ID: <20100409191528.9F72A2A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6806

Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Dan Gohman <gohman at apple.com> 2010-04-09 14:15:28 CDT ---
Fixed in 100883 and 100884.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  9 14:46:27 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  9 Apr 2010 14:46:27 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6808] New: -Wshadow warns about shadowing
	inaccessible fields
Message-ID: <bug-6808-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6808

           Summary: -Wshadow warns about shadowing inaccessible fields
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ori at avtalion.name
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Using trunk r100808.

In the following example, the static method cannot access the non-static field.
Having a variable with the same name shouldn't trigger a warning.

$ cat a.cpp
class Foo {
    int a;

    static void b() {
        int a;
    }
};

$ clang++ -Wshadow -c a.cpp 
a.cpp:5:7: warning: declaration shadows a field of 'Foo' [-Wshadow]
                int a;
                    ^
a.cpp:2:6: note: previous declaration is here
        int a;
            ^

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  9 15:45:19 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  9 Apr 2010 15:45:19 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6797] Patches to port llvm to Minix 3,
	improve portability
In-Reply-To: <bug-6797-206@http.llvm.org/bugs/>
References: <bug-6797-206@http.llvm.org/bugs/>
Message-ID: <20100409204519.081E42A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6797

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-09 15:45:18 CDT ---
Applied in r100895, thanks!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  9 16:02:47 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  9 Apr 2010 16:02:47 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6809] New: Improve waymarking algorithm
Message-ID: <bug-6809-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6809

           Summary: Improve waymarking algorithm
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ggreif at gmail.com
                CC: llvmbugs at cs.uiuc.edu


http://llvm.org/docs/ProgrammersManual.html#UserLayout describes the waymarking
algorithm as presently used. I propose two enhancements to this scheme, both
aimed at higher speed of the Use::getUser() method.

1) When looking at the serial protocol, we see that each binary number starts
with a 1 binary digit. While visually appealing, it is clearly redundant. This
will reduce each call to getUser() by one iteration in the general case.

2) On 64-bit architectures we can allocate 3 bits for waymarking purposes. This
allows 8 bit patterns, and we could come up with clever encodings that give us
statistically optimal results for the n < N (with some reasonable fixed N)
operand counts. A naive one could be:
s
S1 (next one is the User)
S2 (two farter is the User)
S3 (three farter is the User)
00
01
10
11

It would be interesting to study the probability densities of use_iterators,
(i.e. histograms of getUser() iterations).

This algorithm could be similarly QuickChecked in a reference implementation
and transcribed into C++.

Template metaprogramming could select the right algorithm for the host
architecture at LLVM compilation time. We have a wide variety of buildbots that
would validate the implementation for both 32 and 64 bits.

3) This is applicable tweak for both ideas above: Pre-compute waymarks for
(say) 16 Uses and memcpy() it on initialization, instead of computing waymarks
for each case as currently done.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  9 17:05:39 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  9 Apr 2010 17:05:39 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6810] New: Crash on invalid in LLVM IR parser
Message-ID: <bug-6810-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6810

           Summary: Crash on invalid in LLVM IR parser
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Testcase:
%0 = = type { i32, float, float, double }

Crashes with:
llc: LLParser.cpp:325: bool llvm::LLParser::ParseUnnamedType(): Assertion
`Lex.getKind() == lltok::kw_type' failed.
0  llc             0x0000000000ce263f
1  llc             0x0000000000ce30ca
2  libpthread.so.0 0x00007f32d5bab190
3  libc.so.6       0x00007f32d4eb14b5 gsignal + 53
4  libc.so.6       0x00007f32d4eb4f50 abort + 384
5  libc.so.6       0x00007f32d4eaa481 __assert_fail + 241
6  llc             0x000000000056a3fb llvm::LLParser::ParseUnnamedType() + 1051
7  llc             0x0000000000578658 llvm::LLParser::ParseTopLevelEntities() +
264
8  llc             0x0000000000578a68 llvm::LLParser::Run() + 24
9  llc             0x000000000055cc11 llvm::ParseAssembly(llvm::MemoryBuffer*,
llvm::Module*, llvm::SMDiagnostic&, llvm::LLVMContext&) + 1617
10 llc             0x0000000000559965 main + 437
11 libc.so.6       0x00007f32d4e9cabd __libc_start_main + 253
12 llc             0x0000000000558319
Stack dump:
0.    Program arguments: /home/eli/llvmgbuild/Release/bin/llc 
Aborted

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  9 17:23:27 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  9 Apr 2010 17:23:27 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6811] New: Parenthetical messages to super can not
	be compiled
Message-ID: <bug-6811-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6811

           Summary: Parenthetical messages to super can not be compiled
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: csdavec at swan.ac.uk
                CC: llvmbugs at cs.uiuc.edu


$ cat super.m 
@interface A
{
    id isa;
}
- (void)a;
@end

@interface B : A @end
@implementation B
- (void)a
{
    [(super) a];
}
@end
$ clang -c super.m 
super.m:12:4: error: cannot compile this scalar expression yet
        [(super) a];
          ^~~~~
1 diagnostic generated.

Removing the brackets works fine.  Unfortunately, this is not possible when the
code comes from a macro, which adds the brackets.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr  9 21:07:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri,  9 Apr 2010 21:07:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6813] New: fix-it hints broken
Message-ID: <bug-6813-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6813

           Summary: fix-it hints broken
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


A few fix-it tests are failing. The problem is that 'clang -cc1 -fixit' doesn't
emit any output at all when there are additional errors in the file which fixit
can't fix. Because the tests are written to run clang -fixit | clang -Werror,
an empty output from -fixit will pass. (I'm assuming the tests are written
correctly and am working on a patch to change the behaviour of the -fixit
flag.)

In the mean time, we've accumulated the 3 broken tests:

1. test/FixIt/fixit-errors.c. The fix-it from 'get_origin->x' to
'get_origin()->x' is missing entirely. Note that this breakage would have been
caught if that line had an 'expected-note' on it.

2. test/FixIt/fixit-pmem.cpp. This test generates four errors with four fix-it
hints, but one of which is diagnosed as unfixable. Why?

3. test/FixIt/typo.m. Right in the middle, on line 61, the use of undeclared
identifier error on 'NSstring' is emitted with a did you mean 'NSString', but
no fix-it hint shows up. Why not? It shows up for other ones.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 04:29:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 04:29:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6575] Clang fails to compile member initialiser for
 fields of nested anonymous structs and unions
In-Reply-To: <bug-6575-206@http.llvm.org/bugs/>
References: <bug-6575-206@http.llvm.org/bugs/>
Message-ID: <20100410092918.D707A2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6575

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |rjmccall at apple.com
         Resolution|                            |FIXED

--- Comment #1 from John McCall <rjmccall at apple.com> 2010-04-10 04:29:18 CDT ---
Fixed in r100923.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 13:01:34 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 13:01:34 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6810] Crash on invalid in LLVM IR parser
In-Reply-To: <bug-6810-206@http.llvm.org/bugs/>
References: <bug-6810-206@http.llvm.org/bugs/>
Message-ID: <20100410180134.73F822A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6810

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-10 13:01:34 CDT ---
Fixed in r100934, thanks!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 13:04:13 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 13:04:13 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6802] ice: Branch should be foldable!
In-Reply-To: <bug-6802-206@http.llvm.org/bugs/>
References: <bug-6802-206@http.llvm.org/bugs/>
Message-ID: <20100410180413.968B62A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6802

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-10 13:04:13 CDT ---
I recently did a bunch of work on sccp which appears to have fixed this, I
can't repro it.  Please reopen if this still fails on mainline, thanks John!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 13:19:39 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 13:19:39 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6760] The -globalopt pass assertions on the given
	code
In-Reply-To: <bug-6760-206@http.llvm.org/bugs/>
References: <bug-6760-206@http.llvm.org/bugs/>
Message-ID: <20100410181939.400532A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6760

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-10 13:19:38 CDT ---
Fixed in r100936, thanks.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 13:27:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 13:27:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6743] jump threading segfault
In-Reply-To: <bug-6743-206@http.llvm.org/bugs/>
References: <bug-6743-206@http.llvm.org/bugs/>
Message-ID: <20100410182751.6994A2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6743

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-10 13:27:50 CDT ---
Fixed here, thanks John!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20100405/099581.html

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 13:28:36 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 13:28:36 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6735] Cannot split an edge from an IndirectBrInst
 failed (llvm with clang compiling gambit-c)
In-Reply-To: <bug-6735-206@http.llvm.org/bugs/>
References: <bug-6735-206@http.llvm.org/bugs/>
Message-ID: <20100410182836.5A9662A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6735

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #11 from Chris Lattner <clattner at apple.com> 2010-04-10 13:28:35 CDT ---
I can't do anything with this bug without a testcase.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 13:34:45 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 13:34:45 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6805] -fcatch-undefined-behavior crashes with
 signal 6, "LLVM IR generation of compound statement"
In-Reply-To: <bug-6805-206@http.llvm.org/bugs/>
References: <bug-6805-206@http.llvm.org/bugs/>
Message-ID: <20100410183445.A015F2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6805

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-10 13:34:45 CDT ---
Fixed here, thanks:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20100405/029185.html

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 14:15:43 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 14:15:43 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6811] [(super) foo] not rejected:'super' is
	modelled wrong
In-Reply-To: <bug-6811-206@http.llvm.org/bugs/>
References: <bug-6811-206@http.llvm.org/bugs/>
Message-ID: <20100410191543.CE18B2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6811

David Chisnall <csdavec at swan.ac.uk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from David Chisnall <csdavec at swan.ac.uk> 2010-04-10 14:15:43 CDT ---
I'm not sure that I agree.  Super is magic - there are no expressions that
evaluate to super, just the token itself inside 0 or more parentheses.  You
can't ever get super as the result of a message send, for example.  You can't
do pointer arithmetic with super.  The only thing you can do with it is send
messages to it.  

I don't see how we'd fit it into a message expr, because it's an attribute of
the receiver, rather than an attribute of the message expression, so having the
flag in the message expression would not really make sense.  You'd have to have
a flag on all expressions that told you if it evaluated to super, but it would
only ever be true in two cases: when the token is super or when the token is a
ParenExpr containing the super token.

It turns out that the fix for this is trivial - just add ->IgnoreParens() when
testing whether the receiver is super.  It actually took less time to fix than
file the bug report - I should probably do these things in the opposite
order...

Fixed in r100942.

By the way, gcc accepts super as being equivalent to self in every situation
other than message sending, so you can do things like super->isa.  This fails
in the parser in clang.  I don't really object to this, because it's a stupid
thing to be doing and has non-obvious semantics (e.g. you can do super->foo
where foo is defined in the current class).

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 14:16:09 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 14:16:09 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6811] [(super) foo] not rejected:'super' is
	modelled wrong
In-Reply-To: <bug-6811-206@http.llvm.org/bugs/>
References: <bug-6811-206@http.llvm.org/bugs/>
Message-ID: <20100410191609.B07372A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6811

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |

--- Comment #3 from Chris Lattner <clattner at apple.com> 2010-04-10 14:16:09 CDT ---
This code is invalid, please revert your patch.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 16:01:11 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 16:01:11 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6814] New: -pedantic is on by default,
 even if it's not being used
Message-ID: <bug-6814-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6814

           Summary: -pedantic is on by default, even if it's not being
                    used
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: amine48rz at gmail.com
                CC: llvmbugs at cs.uiuc.edu


-pedantic is on by default, even if it's not being used

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 17:14:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 17:14:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6815] New: [META] Compiling ReactOS with clang
Message-ID: <bug-6815-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6815

           Summary: [META] Compiling ReactOS with clang
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: amine48rz at gmail.com
                CC: llvmbugs at cs.uiuc.edu


ReactOS has a mostly C codebase, with a couple of asm and C++ modules. This bug
groups various issues that prevent it from compiling with clang/clang++.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 17:21:39 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 17:21:39 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6361] Clang complains about redeclaring function
 with a calling convention
In-Reply-To: <bug-6361-206@http.llvm.org/bugs/>
References: <bug-6361-206@http.llvm.org/bugs/>
Message-ID: <20100410222139.7294B2A6C130@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6361

amine48rz <amine48rz at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |

--- Comment #3 from amine48rz <amine48rz at gmail.com> 2010-04-10 17:21:39 CDT ---
Tested with clang r100910, issue still exists.

testcase.c(2) :  warning: attribute declaration must precede definition
void __cdecl a();
             ^
testcase.c(1) :  note: previous definition is here
void a(){}
     ^

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 10 19:01:48 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 10 Apr 2010 19:01:48 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6816] New: LLVM-GCC fails to build when configured
 for mips target
Message-ID: <bug-6816-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6816

           Summary: LLVM-GCC fails to build when configured for mips
                    target
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: mark.aldham at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4665)
 --> (http://llvm.org/bugs/attachment.cgi?id=4665)
Terminal output during attempted installation

The following set of commands is used to configure/build the llvm-gcc
installation:

LLVM_OBJ=/home/maldham/llvm/llvm-svn-obj/
cd llvm-gcc-obj
../llvm-gcc-svn-Apr-10-2010-source/configure
--prefix=/home/maldham/llvm/llvm-gcc-install --program-prefix=llvm- \
--enable-llvm=$LLVM_OBJ --enable-languages=c,c++ \
--build=i686-pc-linux-gnu --host=i686-pc-linux-gnu --target=mips-unknown-elf
make > MAKE_LOG.txt
make install

The following error occurs:
/home/maldham/llvm/llvm-gcc-obj/./gcc/xgcc
-B/home/maldham/llvm/llvm-gcc-obj/./gcc/
-B/home/maldham/llvm/llvm-gcc-install/mips-unknown-elf/bin/
-B/home/maldham/llvm/llvm-gcc-install/mips-unknown-elf/lib/ -isystem
/home/maldham/llvm/llvm-gcc-install/mips-unknown-elf/include -isystem
/home/maldham/llvm/llvm-gcc-install/mips-unknown-elf/sys-include -O2 -O2 -g -O2
 -DIN_GCC -DCROSS_DIRECTORY_STRUCTURE   -W -Wall -Wwrite-strings
-Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition  -isystem
./include  -I. -I. -I../../llvm-gcc-svn-Apr-10-2010-source/gcc
-I../../llvm-gcc-svn-Apr-10-2010-source/gcc/.
-I../../llvm-gcc-svn-Apr-10-2010-source/gcc/../include
-I../../llvm-gcc-svn-Apr-10-2010-source/gcc/../libcpp/include 
-I../../llvm-gcc-svn-Apr-10-2010-source/gcc/../libdecnumber -I../libdecnumber
-I/home/maldham/llvm/llvm-svn-obj//include
-I/home/maldham/llvm/llvm-svn-Apr-10-2010-source/include  -g0
-finhibit-size-directive -fno-inline-functions -fno-exceptions
-fno-zero-initialized-in-bss -fno-toplevel-reorder -Dinhibit_libc -G 0 \
          -c ../../llvm-gcc-svn-Apr-10-2010-source/gcc/crtstuff.c -DCRT_BEGIN \
          -o crtbegin.o
exec: 2: -G: not found


Attached is config.log, MAKE_LOG.txt, and the capture of the onscreen output in
ONSCREEN_LOG.txt

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 11 00:51:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 11 Apr 2010 00:51:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6817] New: #warning not ignored in -MM mode
Message-ID: <bug-6817-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6817

           Summary: #warning not ignored in -MM mode
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: alexei.svitkine at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The -MM mode is used to generate a Makefile-style dependency graph for the
files passed in.

In clang, #warning directives in the files passed in (and I suspect #error
directives too) will still generate diagnostics, even though the code here
isn't being compiled - only parsed for the sake of generating the header
dependency information.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 11 02:41:52 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 11 Apr 2010 02:41:52 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6816] LLVM-GCC fails to build when configured for
	mips target
In-Reply-To: <bug-6816-206@http.llvm.org/bugs/>
References: <bug-6816-206@http.llvm.org/bugs/>
Message-ID: <20100411074152.0D2C72A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6816

Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |asl at math.spbu.ru
         Resolution|                            |INVALID

--- Comment #4 from Anton Korobeynikov <asl at math.spbu.ru> 2010-04-11 02:41:51 CDT ---
(In reply to comment #0)
> -finhibit-size-directive -fno-inline-functions -fno-exceptions
> -fno-zero-initialized-in-bss -fno-toplevel-reorder -Dinhibit_libc -G 0 \
>           -c ../../llvm-gcc-svn-Apr-10-2010-source/gcc/crtstuff.c -DCRT_BEGIN \
>           -o crtbegin.o
> exec: 2: -G: not found
Apparently it doesn't seem you have cross binutils for mips-unknown-elf
installed.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 11 03:29:01 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 11 Apr 2010 03:29:01 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6811] [(super) foo] not rejected:'super' is
	modelled wrong
In-Reply-To: <bug-6811-206@http.llvm.org/bugs/>
References: <bug-6811-206@http.llvm.org/bugs/>
Message-ID: <20100411082901.834832A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6811

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #5 from Chris Lattner <clattner at apple.com> 2010-04-11 03:29:00 CDT ---
This is now fixed, this example (and many other similar bogus things) is
rejected by the parser here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20100405/029203.html

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 11 11:55:35 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 11 Apr 2010 11:55:35 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6818] New: clang++ permits use of private struct.
Message-ID: <bug-6818-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6818

           Summary: clang++ permits use of private struct.
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: markus at unixforces.net
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


clang++ accepts the following code, while g++ does not:


class List {
    private:
        struct Elem {
            Elem* n;
        };
};
int main()
{
    List::Elem a;
}


I'm not entirely sure which one is correct, but I think this should not be
accepted by clang++.

$ clang++ -v
clang version 1.5 (trunk 99761)
Target: x86_64-unknown-linux-gnu
Thread model: posix

$ g++ -v
Using built-in specs.
Target: x86_64-pc-linux-gnu
Configured with:
/var/tmp/paludis/build/sys-devel-gcc-4.4.3/work/gcc-4.4.3/configure
--prefix=/usr --host=x86_64-pc-linux-gnu --build=x86_64-pc-linux-gnu
--mandir=/usr/share/man --infodir=/usr/share/info --datadir=/usr/share
--sysconfdir=/etc --localstatedir=/var/lib --disable-dependency-tracking
--disable-silent-rules --enable-fast-install --libdir=/usr/lib64
--disable-multilib --enable-languages=c++,ada,fortran,java
--enable-libstdcxx-pch=yes --enable-nls --program-suffix=-4.4
--with-pkgversion='Exherbo gcc-4.4.3' --without-cloog --without-ppl
--with-system-zlib --disable-libssp
Thread model: posix
gcc version 4.4.3 (Exherbo gcc-4.4.3)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 11 21:19:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 11 Apr 2010 21:19:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6287] Clang doesn't support
	__attribute__((returns_twice))
In-Reply-To: <bug-6287-206@http.llvm.org/bugs/>
References: <bug-6287-206@http.llvm.org/bugs/>
Message-ID: <20100412021903.9CED32A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6287

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-11 21:19:03 CDT ---
Fixed in r101005/101006

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 11 23:57:08 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 11 Apr 2010 23:57:08 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6807] llvm::SimplifyFortifiedLibCalls::fold crash
In-Reply-To: <bug-6807-206@http.llvm.org/bugs/>
References: <bug-6807-206@http.llvm.org/bugs/>
Message-ID: <20100412045708.E7C5A2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6807

Eric Christopher <echristo at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 00:39:44 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 00:39:44 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6818] clang++ permits use of private struct.
In-Reply-To: <bug-6818-206@http.llvm.org/bugs/>
References: <bug-6818-206@http.llvm.org/bugs/>
Message-ID: <20100412053944.2DABE2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6818

Nick Lewycky <nicholas at mxc.ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |nicholas at mxc.ca
         Resolution|                            |FIXED

--- Comment #2 from Nick Lewycky <nicholas at mxc.ca> 2010-04-12 00:39:43 CDT ---
Believe it or not r99761 is so old that it predates access control in clang,
access control being the term for public/protected/private in C++. With a
current clang from svn head, you should get:

$ clang++ pr6818.cc
pr6818.cc:9:11: error: 'Elem' is a private member of 'List'
    List::Elem a;
          ^
pr6818.cc:3:16: note: declared private here
        struct Elem {
               ^
1 error generated.

Please reopen if that isn't the case with clang post-r100880.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 02:51:26 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 02:51:26 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6733] template class with default arg assigned from
 enum causes assert
In-Reply-To: <bug-6733-206@http.llvm.org/bugs/>
References: <bug-6733-206@http.llvm.org/bugs/>
Message-ID: <20100412075126.67C232A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6733

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-12 02:51:25 CDT ---
Fixed in r101029.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 05:10:55 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 05:10:55 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6820] New: 3 failing MC/Disassembler test cases
 with ./configure --enable-targets=x86
Message-ID: <bug-6820-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6820

           Summary: 3 failing MC/Disassembler test cases with ./configure
                    --enable-targets=x86
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: grosser at fim.uni-passau.de
                CC: llvmbugs at cs.uiuc.edu


Both CMake and autoconf have three failing test cases in MC/Dissassembler. It
looks like these tests need to be disabled if arm/neon/thumb support is not
compiled in.

By the way. Is there a MC component in the bugtracker?

Here the autoconf output:
==============================================================================
FAIL: /home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/arm-tests.txt
Failed with exit(1) at line 1
while running: llvm-mc --disassemble
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/arm-tests.txt
-triple=arm-apple-darwin9 | FileCheck
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/arm-tests.txt
llvm-mc: error: unable to get target for 'arm-apple-darwin9', see --version and
--triple.
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/arm-tests.txt:3:10:
error: expected string not found in input
# CHECK: b #0
         ^
<stdin>:1:1: note: scanning from here

^

FAIL: /home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/neon-tests.txt
Failed with exit(1) at line 1
while running: llvm-mc --disassemble
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/neon-tests.txt
-triple=arm-apple-darwin9 | FileCheck
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/neon-tests.txt
llvm-mc: error: unable to get target for 'arm-apple-darwin9', see --version and
--triple.
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/neon-tests.txt:3:10:
error: expected string not found in input
# CHECK: vbif q15, q7, q0
         ^
<stdin>:1:1: note: scanning from here

^

FAIL:
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/thumb-tests.txt
Failed with exit(1) at line 1
while running: llvm-mc --disassemble
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/thumb-tests.txt
-triple=thumb-apple-darwin9 | FileCheck
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/thumb-tests.txt
llvm-mc: error: unable to get target for 'thumb-apple-darwin9', see --version
and --triple.
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/thumb-tests.txt:3:10:
error: expected string not found in input
# CHECK: add r5, sp, #68
         ^
<stdin>:1:1: note: scanning from here

^
==============================================================================





And here one of the cmake ones:
==============================================================================
FAIL: LLVM :: MC/Disassembler/thumb-tests.txt (3537 of 4935)
******************** TEST 'LLVM :: MC/Disassembler/thumb-tests.txt' FAILED
********************
Script:
--
llvm-mc --disassemble
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/thumb-tests.txt
-triple=thumb-apple-darwin9 | FileCheck
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/thumb-tests.txt
--
Exit Code: 1
Command Output (stdout):
--
Command has output on stderr!

--
Command Output (stderr):
--
llvm-mc: error: unable to get target for 'thumb-apple-darwin9', see --version
and --triple.
/home/grosser/Projekte/llvm/git_trunk/test/MC/Disassembler/thumb-tests.txt:3:10:
error: expected string not found in input
# CHECK: add r5, sp, #68
         ^
<stdin>:1:1: note: scanning from here

^
==============================================================================

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 07:59:02 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 07:59:02 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6821] New: JIT DWARF EH data does not seem to emit
	LSDA
Message-ID: <bug-6821-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6821

           Summary: JIT DWARF EH data does not seem to emit LSDA
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: csdavec at swan.ac.uk
                CC: llvmbugs at cs.uiuc.edu


Calling _Unwind_GetLanguageSpecificData() from a personality function returns
NULL in JIT'd frames.  At least some of the EH data is being correctly emitted,
because the correct personality function is being called (this personality
function is only referenced in JIT'd code, so the fact that it's called at all
tells me something).  Statically compiling the same IR works fine.

I think this used to work, but unfortunately it's not a feature that I use very
often so I didn't notice exactly when my test started to fail.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 09:16:34 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 09:16:34 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6794] Assertion failed: ((Result ||
 D->isInvalidDecl()) && "declaration was not instantiated in this scope!")
In-Reply-To: <bug-6794-206@http.llvm.org/bugs/>
References: <bug-6794-206@http.llvm.org/bugs/>
Message-ID: <20100412141634.4F63A2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6794

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME

--- Comment #4 from Douglas Gregor <dgregor at apple.com> 2010-04-12 09:16:33 CDT ---
This is working fine for me. Please re-open the bug if it's still failing for
you,

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 10:03:58 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 10:03:58 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6794] Assertion failed: ((Result ||
 D->isInvalidDecl()) && "declaration was not instantiated in this scope!")
In-Reply-To: <bug-6794-206@http.llvm.org/bugs/>
References: <bug-6794-206@http.llvm.org/bugs/>
Message-ID: <20100412150358.B4D192A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6794

Albert Zeyer <ich at az2000.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|WORKSFORME                  |

--- Comment #5 from Albert Zeyer <ich at az2000.de> 2010-04-12 10:03:57 CDT ---
Still crashes:


macbook-7:openlierox az$ ../CppIDE/llvm-src/Debug/bin/clang++ test.cpp 
Assertion failed: ((Result || D->isInvalidDecl()) && "declaration was not
instantiated in this scope!"), function getInstantiationOf, file Sema.h, line
3570.
0  clang             0x00f6e050 main + 16152202
1  clang             0x00f6e496 main + 16153296
2  libSystem.B.dylib 0x94cbd2bb _sigtramp + 43
3  libSystem.B.dylib 0xffffffff _sigtramp + 1798581615
4  libSystem.B.dylib 0x94d3123a raise + 26
5  libSystem.B.dylib 0x94d3d679 abort + 73
6  libSystem.B.dylib 0x94d323db __assert_rtn + 101
7  clang             0x003a9c53 main + 3814029
8  clang             0x003a2fd7 main + 3786257
9  clang             0x0037429a main + 3594452
10 clang             0x00388d38 main + 3679090
11 clang             0x00375a9a main + 3600596
12 clang             0x0038937a main + 3680692
13 clang             0x00391e73 main + 3716269
14 clang             0x00389829 main + 3681891
15 clang             0x003918cd main + 3714823
16 clang             0x00391e01 main + 3716155
17 clang             0x0038a040 main + 3683962
18 clang             0x003925e0 main + 3718170
19 clang             0x00389896 main + 3682000
20 clang             0x00397645 main + 3738751
21 clang             0x00395bf2 main + 3732012
22 clang             0x003966af main + 3734761
23 clang             0x00396bc4 main + 3736062
24 clang             0x00395a3e main + 3731576
25 clang             0x00377b34 main + 3608942
26 clang             0x003a7ad3 main + 3805453
27 clang             0x003a7416 main + 3803728
28 clang             0x003a7c39 main + 3805811
29 clang             0x003a7416 main + 3803728
30 clang             0x00225909 main + 2223939
31 clang             0x005f4827 main + 6217313
32 clang             0x00223b5b main + 2216341
33 clang             0x0003e552 main + 228236
34 clang             0x0003e442 main + 227964
35 clang             0x0001fd11 main + 103243
36 clang             0x000031a7 _mh_execute_header + 8615
37 clang             0x00006aaf main + 233
38 clang             0x00002036 _mh_execute_header + 4150
Stack dump:
0.    Program arguments:
/Users/az/Programmierung/CppIDE/llvm-src/Debug/bin/clang -cc1 -triple
i386-apple-darwin9.0.0 -S -disable-free -main-file-name test.cpp -pic-level 1
-mdisable-fp-elim -target-cpu yonah -resource-dir
/Users/az/Programmierung/CppIDE/llvm-src/Debug/lib/clang/1.5 -ferror-limit 19
-fmessage-length 80 -fexceptions -fdiagnostics-show-option -fcolor-diagnostics
-o /var/folders/P9/P9wWY+Z8HDeOPixVmDZ0J++++TI/-Tmp-/cc-RamXbS.s -x c++
test.cpp 
1.    <eof> parser at end of file
2.    test.cpp:6:7: instantiating function definition
'__OlxVariable<int>::gusVar'
3.    test.cpp:8:24: instantiating function definition
'__OlxVariable<int>::gusVar()::GusVarWrapper::invoke'



Process:         clang [84594]
Path:            /Users/az/Programmierung/CppIDE/llvm-src/Debug/bin/clang
Identifier:      clang
Version:         ??? (???)
Code Type:       X86 (Native)
Parent Process:  clang++ [84593]

Interval Since Last Report:          152142 sec
Crashes Since Last Report:           2
Per-App Interval Since Last Report:  0 sec
Per-App Crashes Since Last Report:   1

Date/Time:       2010-04-12 17:03:41.308 +0200
OS Version:      Mac OS X 10.5.8 (9L30)
Report Version:  6
Anonymous UUID:  C99325D9-953E-44CB-A0FE-88CB8BF0B517

Exception Type:  EXC_CRASH (SIGABRT)
Exception Codes: 0x0000000000000000, 0x0000000000000000
Crashed Thread:  0

Application Specific Information:
Stack dump:
0.    Program arguments:
/Users/az/Programmierung/CppIDE/llvm-src/Debug/bin/clang -cc1 -triple
i386-apple-darwin9.0.0 -S -disable-free -main-file-name test.cpp -pic-level 1
-mdisable-fp-elim -target-cpu yonah -resource-dir
/Users/az/Programmierung/CppIDE/llvm-src/Debug/lib/clang/1.5 -ferror-limit 19
-fmessage-length 80 -fexceptions -fdiagnostics-show-option -fcolor-diagnostics
-o /var/folders/P9/P9wWY+Z8HDeOPixVmDZ0J++++TI/-Tmp-/cc-RamXbS.s -x c++
test.cpp 
1.    <eof> parser at end of file
2.    test.cpp:6:7: instantiating function definition
'__OlxVariable<int>::gusVar'
3.    test.cpp:8:24: instantiating function definition
'__OlxVariable<int>::gusVar()::GusVarWrapper::invoke'


Thread 0 Crashed:
0   libSystem.B.dylib                 0x94cbee42 __kill + 10
1   libSystem.B.dylib                 0x94d3123a raise + 26
2   libSystem.B.dylib                 0x94d3d622 __abort + 97
3   libSystem.B.dylib                 0x94d3d68a _cproc_fork_child + 0
4   libSystem.B.dylib                 0x94d323db __assert_rtn + 101
5   clang                             0x003a9c53
clang::Sema::LocalInstantiationScope::getInstantiationOf(clang::Decl const*) +
123
6   clang                             0x003a2fd7
clang::Sema::FindInstantiatedDecl(clang::SourceLocation, clang::NamedDecl*,
clang::MultiLevelTemplateArgumentList const&) + 147
7   clang                             0x0037429a (anonymous
namespace)::TemplateInstantiator::TransformDecl(clang::SourceLocation,
clang::Decl*) + 390
8   clang                             0x00388d38
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformDeclRefExpr(clang::DeclRefExpr*) +
272
9   clang                             0x00375a9a (anonymous
namespace)::TemplateInstantiator::TransformDeclRefExpr(clang::DeclRefExpr*) +
188
10  clang                             0x0038937a
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformExpr(clang::Expr*) + 724
11  clang                             0x00391e73
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformMemberExpr(clang::MemberExpr*) + 57
12  clang                             0x00389829
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformExpr(clang::Expr*) + 1923
13  clang                             0x003918cd
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformCallExpr(clang::CallExpr*) + 57
14  clang                             0x00391e01
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformCXXMemberCallExpr(clang::CXXMemberCallExpr*)
+ 47
15  clang                             0x0038a040
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformExpr(clang::Expr*) + 3994
16  clang                             0x003925e0
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformBinaryOperator(clang::BinaryOperator*)
+ 54
17  clang                             0x00389896
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformExpr(clang::Expr*) + 2032
18  clang                             0x00397645
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformIfStmt(clang::IfStmt*) + 287
19  clang                             0x00395bf2
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformStmt(clang::Stmt*) + 1162
20  clang                             0x003966af
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformCompoundStmt(clang::CompoundStmt*,
bool) + 121
21  clang                             0x00396bc4
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformCompoundStmt(clang::CompoundStmt*)
+ 56
22  clang                             0x00395a3e
clang::TreeTransform<(anonymous
namespace)::TemplateInstantiator>::TransformStmt(clang::Stmt*) + 726
23  clang                             0x00377b34
clang::Sema::SubstStmt(clang::Stmt*, clang::MultiLevelTemplateArgumentList
const&) + 178
24  clang                             0x003a7ad3
clang::Sema::InstantiateFunctionDefinition(clang::SourceLocation,
clang::FunctionDecl*, bool, bool) + 1259
25  clang                             0x003a7416
clang::Sema::PerformPendingImplicitInstantiations(bool) + 328
26  clang                             0x003a7c39
clang::Sema::InstantiateFunctionDefinition(clang::SourceLocation,
clang::FunctionDecl*, bool, bool) + 1617
27  clang                             0x003a7416
clang::Sema::PerformPendingImplicitInstantiations(bool) + 328
28  clang                             0x00225909
clang::Sema::ActOnEndOfTranslationUnit() + 37
29  clang                             0x005f4827
clang::Parser::ParseTopLevelDecl(clang::OpaquePtr<1>&) + 77 (Parser.cpp:345)
30  clang                             0x00223b5b
clang::ParseAST(clang::Preprocessor&, clang::ASTConsumer*, clang::ASTContext&,
bool, bool, clang::CodeCompleteConsumer*) + 435
31  clang                             0x0003e552
clang::ASTFrontendAction::ExecuteAction() + 266
32  clang                             0x0003e442
clang::FrontendAction::Execute() + 278
33  clang                             0x0001fd11
clang::CompilerInstance::ExecuteAction(clang::FrontendAction&) + 879
34  clang                             0x000031a7 cc1_main(char const**, char
const**, char const*, void*) + 1439
35  clang                             0x00006aaf main + 233 (driver.cpp:181)
36  clang                             0x00002036 start + 54

Thread 0 crashed with X86 Thread State (32-bit):
  eax: 0x00000000  ebx: 0x94d3d639  ecx: 0xbfffbacc  edx: 0x94cbee42
  edi: 0x000000c6  esi: 0x00faab08  ebp: 0xbfffbae8  esp: 0xbfffbacc
   ss: 0x0000001f  efl: 0x00000282  eip: 0x94cbee42   cs: 0x00000007
   ds: 0x0000001f   es: 0x0000001f   fs: 0x00000000   gs: 0x00000037
  cr2: 0x94d2d12c

Binary Images:
    0x1000 -  0x1154fd7 +clang ??? (???) <bd759702f41585582dc3731cbe2531cd>
/Users/az/Programmierung/CppIDE/llvm-src/Debug/bin/clang
0x8fe00000 - 0x8fe2db43  dyld 97.1 (???) <458eed38a009e5658a79579e7bc26603>
/usr/lib/dyld
0x92737000 - 0x9273efe9  libgcc_s.1.dylib ??? (???)
<f53c808e87d1184c0f9df63aef53ce0b> /usr/lib/libgcc_s.1.dylib
0x946e4000 - 0x94741ffb  libstdc++.6.dylib ??? (???)
<04b812dcec670daa8b7d2852ab14be60> /usr/lib/libstdc++.6.dylib
0x94c50000 - 0x94db7ff3  libSystem.B.dylib ??? (???)
<c8f52e158bf540cc000146ca8a705958> /usr/lib/libSystem.B.dylib
0x95237000 - 0x9523bfff  libmathCommon.A.dylib ??? (???)
/usr/lib/system/libmathCommon.A.dylib
0xffff0000 - 0xffff1780  libSystem.B.dylib ??? (???) /usr/lib/libSystem.B.dylib

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 11:00:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 11:00:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6770] clang::Sema::ActOnTag(...): Assertion
 `Previous.isSingleResult()' failed
In-Reply-To: <bug-6770-206@http.llvm.org/bugs/>
References: <bug-6770-206@http.llvm.org/bugs/>
Message-ID: <20100412160018.634712A6C12F@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6770

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-12 11:00:17 CDT ---
Fixed in r101040.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 12:11:34 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 12:11:34 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6787] Unable to step through a program compiled
 with clang in gdb
In-Reply-To: <bug-6787-206@http.llvm.org/bugs/>
References: <bug-6787-206@http.llvm.org/bugs/>
Message-ID: <20100412171134.E28232A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6787

devang.patel <devang.patel at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME

--- Comment #5 from devang.patel <devang.patel at gmail.com> 2010-04-12 12:11:34 CDT ---
This works for me.

$ ~/clean/D/Debug/bin/clang -g ~/Downloads/prog.c -o prog
rajkot:tmp yash$ gdb prog
...
...
(gdb) b main
Breakpoint 1 at 0x100000ed6: file prog.c, line 5.
(gdb) r
Starting program: /private/tmp/prog 
Reading symbols for shared libraries +. done

Breakpoint 1, main (argc=1, argv=0x7fff5fbffa88) at prog.c:5
5        for (int i = 0; i < 3; ++i)
(gdb) n
Current language:  auto; currently minimal
7            printf("i == %d\n",
(gdb) 
i == 0
5        for (int i = 0; i < 3; ++i)
(gdb) 
7            printf("i == %d\n",
(gdb) 
i == 1
5        for (int i = 0; i < 3; ++i)
(gdb) 
7            printf("i == %d\n",
(gdb) 
i == 2
5        for (int i = 0; i < 3; ++i)
(gdb) 
10        return 0;
(gdb) 
11    }
(gdb) 

$ ~/clean/D/Debug/bin/clang --version
clang version 1.5 (trunk 101039)
Target: x86_64-apple-darwin10
Thread model: posix

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 12:15:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 12:15:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6570] Implicit copy-assignment operator hidden by
 overloaded assignment operator
In-Reply-To: <bug-6570-206@http.llvm.org/bugs/>
References: <bug-6570-206@http.llvm.org/bugs/>
Message-ID: <20100412171556.618DB2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6570

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Douglas Gregor <dgregor at apple.com> 2010-04-12 12:15:56 CDT ---
Fixed in r101047.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 12:41:27 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 12:41:27 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6822] New: clang assertion when compiling
 reinterpret_cast<float3>(float4_var)
Message-ID: <bug-6822-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6822

           Summary: clang assertion when compiling
                    reinterpret_cast<float3>(float4_var)
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bixia.zheng at amd.com
                CC: llvmbugs at cs.uiuc.edu


float3 and float4 are defined by the test case below.
Clang seems to believe that BitWidth(float3) = 4 * BitWidth(float) while LLVM
Core VectorType::getBitWidth(float3) returns 3 * BitWidth(float). This
discrepancy causes the assertion error. Where are you doing to fix this, Clang
or LLVM Core?
Thanks for your attention to this bug!

command line:
================
clang -emit-llvm -ObjC++ -S vector03.c

error message:
================
Assertion failed: castIsValid(getOpcode(), S, Ty) && "Illegal BitCast", file
... llvm\lib\VMCore\Instructions.cpp, line 2732

test case:
================
typedef float float4 __attribute__((ext_vector_type(4)));
typedef float float3 __attribute__((ext_vector_type(3)));
//typedef float float2 __attribute__((ext_vector_type(2)));

typedef struct _mystruct {
    float3 field1;
    float field2;
    float field3;
    float field4;
    float field5;
} mystruct;

float3 foo() {
  float3 vec3;
  float4 vec4;
  vec4 = sizeof(mystruct);
  vec3 = reinterpret_cast<float3>(vec4);

  return vec3;
}

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 13:36:50 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 13:36:50 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6823] New: llvm-config --libs prints static
 libraries even if llvm is configured with --enable-shared
Message-ID: <bug-6823-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6823

           Summary: llvm-config --libs prints static libraries even if
                    llvm is configured with --enable-shared
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: doko at ubuntu.com
                CC: llvmbugs at cs.uiuc.edu


$ llvm-config --libs
-lLLVMXCoreCodeGen -lLLVMXCoreAsmPrinter -lLLVMXCoreInfo -lLLVMSystemZCodeGen
-lLLVMSystemZAsmPrinter -lLLVMSystemZInfo -lLLVMSparcCodeGen
-lLLVMSparcAsmPrinter -lLLVMSparcInfo -lLLVMPowerPCCodeGen
-lLLVMPowerPCAsmPrinter -lLLVMPowerPCInfo -lLLVMPIC16AsmPrinter
-lLLVMPIC16CodeGen -lLLVMPIC16Info -lLLVMMSP430CodeGen -lLLVMMSP430AsmPrinter
-lLLVMMSP430Info -lLLVMMSIL -lLLVMMSILInfo -lLLVMMipsAsmPrinter
-lLLVMMipsCodeGen -lLLVMMipsInfo -lLLVMLinker -lLLVMipo -lLLVMInterpreter
-lLLVMInstrumentation -lLLVMJIT -lLLVMExecutionEngine -lLLVMDebugger
-lLLVMCppBackend -lLLVMCppBackendInfo -lLLVMCellSPUCodeGen
-lLLVMCellSPUAsmPrinter -lLLVMCellSPUInfo -lLLVMCBackend -lLLVMCBackendInfo
-lLLVMBlackfinCodeGen -lLLVMBlackfinAsmPrinter -lLLVMBlackfinInfo
-lLLVMBitWriter -lLLVMX86AsmParser -lLLVMX86AsmPrinter -lLLVMX86CodeGen
-lLLVMX86Info -lLLVMAsmParser -lLLVMARMAsmPrinter -lLLVMARMCodeGen
-lLLVMARMInfo -lLLVMArchive -lLLVMBitReader -lLLVMAlphaCodeGen
-lLLVMSelectionDAG -lLLVMAlphaAsmPrinter -lLLVMAsmPrinter -lLLVMCodeGen
-lLLVMScalarOpts -lLLVMTransformUtils -lLLVMipa -lLLVMAnalysis -lLLVMTarget
-lLLVMMC -lLLVMCore -lLLVMAlphaInfo -lLLVMSupport -lLLVMSystem
-L/usr/lib/llvm/lib

I would expect that the shared library is the result for a configuration with
--enable-shared.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 14:27:05 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 14:27:05 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6824] New: crash in X86 DAG->DAG Instruction
	Selection
Message-ID: <bug-6824-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6824

           Summary: crash in X86 DAG->DAG Instruction Selection
           Product: new-bugs
           Version: trunk
          Platform: Other
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: kwm at FreeBSD.org
                CC: llvmbugs at cs.uiuc.edu


> cat fopen.c 
#include <stdio.h>
 int main ()
{
 FILE *f = fopen ("conftest.out", "w");
}

> clang -v
clang version 1.5 (trunk 101035)
Target: i386-unknown-freebsd8.0
Thread model: posix
> clang -O2 fopen.c
Stack dump:
0.    Program arguments: /usr/local/clang/bin/clang -cc1 -triple
i386-unknown-freebsd8.0 -S -disable-free -main-file-name fopen.c
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases -target-cpu
pentium4 -resource-dir /usr/local/clang/lib/clang/1.5 -O2 -ferror-limit 19
-fmessage-length 80 -fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics
-o /tmp/cc-rrmzgR.s -x c fopen.c 
1.    <eof> parser at end of file
2.    Code generation
3.    Running pass 'X86 DAG->DAG Instruction Selection' on function '@main'
clang: error: compiler command failed due to signal 11 (use -v to see
invocation)

-O1 also crashes, -O0 works.
This testcase works with clang revision 100677.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 15:54:39 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 15:54:39 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6717] Cannot handle template name-lookup ambiguities
In-Reply-To: <bug-6717-206@http.llvm.org/bugs/>
References: <bug-6717-206@http.llvm.org/bugs/>
Message-ID: <20100412205439.7CE312A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6717

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-12 15:54:39 CDT ---
Fixed in r101065.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 16:11:30 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 16:11:30 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6660] clang-2.7 regressions with respect to
	clang-2.6
In-Reply-To: <bug-6660-206@http.llvm.org/bugs/>
References: <bug-6660-206@http.llvm.org/bugs/>
Message-ID: <20100412211130.BCA782A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6660

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #9 from Chris Lattner <clattner at apple.com> 2010-04-12 16:11:30 CDT ---
Fixed here, please verify:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20100412/029259.html

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 16:12:16 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 16:12:16 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6168] clang emits missing fields as undef instead
	of 0
In-Reply-To: <bug-6168-206@http.llvm.org/bugs/>
References: <bug-6168-206@http.llvm.org/bugs/>
Message-ID: <20100412211216.AFD222A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6168

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #11 from Chris Lattner <clattner at apple.com> 2010-04-12 16:12:15 CDT ---
Minimal patch here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20100412/029259.html

If you'd like to still get undefs, that can be added back on top of this patch,
for now, correctness is most important.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 16:53:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 16:53:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6814] ExtWarn diagnostics shouldn't print
	[-pedantic]
In-Reply-To: <bug-6814-206@http.llvm.org/bugs/>
References: <bug-6814-206@http.llvm.org/bugs/>
Message-ID: <20100412215331.DA2822A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6814

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
            Summary|ExtWarn diagnostics should  |ExtWarn diagnostics
                   |have a controlling -W flag  |shouldn't print [-pedantic]

--- Comment #5 from Chris Lattner <clattner at apple.com> 2010-04-12 16:53:31 CDT ---
Fixed in r101071

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 17:09:55 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 17:09:55 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6787] Unable to step through a program compiled
 with clang in gdb
In-Reply-To: <bug-6787-206@http.llvm.org/bugs/>
References: <bug-6787-206@http.llvm.org/bugs/>
Message-ID: <20100412220955.5A5692A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6787

Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
                 CC|                            |asl at math.spbu.ru
         Resolution|WORKSFORME                  |

--- Comment #6 from Anton Korobeynikov <asl at math.spbu.ru> 2010-04-12 17:09:54 CDT ---
I don't think that test on darwin is a good way of verifying problems seen at
least on linux.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 17:48:34 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 17:48:34 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6825] New: Missing access control check for
 injected-class-names used as template-names
Message-ID: <bug-6825-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6825

           Summary: Missing access control check for injected-class-names
                    used as template-names
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


We aren't performing access control for injected-class-names that are used as
template names, e.g.,

terfin:clang dgregor$ cat t3.C
template <class T> struct Base { }; 

template<typename T> class Between : private Base<T> { };

template <class T> struct Derived : private Between<int>, private Between<char>
{
  Base<double> d;
};

Derived<double> dd;
terfin:clang dgregor$ clang++ -fsyntax-only t3.C
terfin:clang dgregor$ eccp -A t3.C
"t3.C", line 6: error: type "Base<T>::Base [with T=int]" (declared at line 1)
          is inaccessible
    Base<double> d;
    ^
          detected during instantiation of class "Derived<T> [with T=double]"
                    at line 9

1 error detected in the compilation of "t3.C".

The solution is not trivial, because we are performing some ambiguity
resolution (per C++ [temp.local]p4) when the injected-class-names are mapped to
class templates, so we will need to hold on to all of the paths to each of the
injected-class-names; some of those paths may not be accessible while others
are.

GCC doesn't diagnose this example, nor does EDG in non-strict mode, so it's not
critical.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 18:44:08 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 18:44:08 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6285] Second SCS of conversion function template:
 No exact match check
In-Reply-To: <bug-6285-206@http.llvm.org/bugs/>
References: <bug-6285-206@http.llvm.org/bugs/>
Message-ID: <20100412234408.017162A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6285

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #5 from Douglas Gregor <dgregor at apple.com> 2010-04-12 18:44:07 CDT ---
Implemented in r101089. I was pleasantly surprised to find that we'd already
implemented C++0x default function template arguments, so we're actually able
to test this.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 19:03:32 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 19:03:32 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6101] clang: preprocessor doesn't honor #line
	directives!
In-Reply-To: <bug-6101-206@http.llvm.org/bugs/>
References: <bug-6101-206@http.llvm.org/bugs/>
Message-ID: <20100413000332.396022A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6101

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-12 19:03:31 CDT ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20100412/029272.html

Thanks to Jordy Rose for the start of the patch.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 19:56:27 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 19:56:27 -0500 (CDT)
Subject: [LLVMbugs] [Bug 5866] Friend template specialization in class
 template loses specialization info
In-Reply-To: <bug-5866-206@http.llvm.org/bugs/>
References: <bug-5866-206@http.llvm.org/bugs/>
Message-ID: <20100413005627.0A5B92A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=5866

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #4 from John McCall <rjmccall at apple.com> 2010-04-12 19:56:26 CDT ---
And lo, it works and the hack is no more.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 19:57:37 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 19:57:37 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6018] Implement access control
In-Reply-To: <bug-6018-206@http.llvm.org/bugs/>
References: <bug-6018-206@http.llvm.org/bugs/>
Message-ID: <20100413005737.05C932A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6018

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #4 from John McCall <rjmccall at apple.com> 2010-04-12 19:57:36 CDT ---
Calling this done;  there are some significant holes, but nothing that's useful
to cover in an umbrella bug like this.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 20:03:05 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 20:03:05 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6569] clang fails assertion for namespace/friend
	construct.
In-Reply-To: <bug-6569-206@http.llvm.org/bugs/>
References: <bug-6569-206@http.llvm.org/bugs/>
Message-ID: <20100413010305.957072A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6569

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |rjmccall at apple.com
         Resolution|                            |DUPLICATE

--- Comment #1 from John McCall <rjmccall at apple.com> 2010-04-12 20:03:05 CDT ---


*** This bug has been marked as a duplicate of bug 6174 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 20:04:44 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 20:04:44 -0500 (CDT)
Subject: [LLVMbugs] [Bug 5848] clang crashes when instantiating a class
 template that has a nested friend class tempalte
In-Reply-To: <bug-5848-206@http.llvm.org/bugs/>
References: <bug-5848-206@http.llvm.org/bugs/>
Message-ID: <20100413010444.9B2B32A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=5848

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |rjmccall at apple.com
         Resolution|                            |FIXED

--- Comment #6 from John McCall <rjmccall at apple.com> 2010-04-12 20:04:44 CDT ---
This seems to be fixed properly.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 20:45:12 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 20:45:12 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6174] clang crashes on (non-template) friend
	declaration
In-Reply-To: <bug-6174-206@http.llvm.org/bugs/>
References: <bug-6174-206@http.llvm.org/bugs/>
Message-ID: <20100413014512.5D32D2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6174

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from John McCall <rjmccall at apple.com> 2010-04-12 20:45:11 CDT ---
Fixed in r101112.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 22:03:19 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 22:03:19 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6826] New: GraphWriter delete the generated file
 before "dotty" load it.
Message-ID: <bug-6826-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6826

           Summary: GraphWriter delete the generated file before "dotty"
                    load it.
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P
         Component: Support Libraries
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: etherzhhb at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4671)
 --> (http://llvm.org/bugs/attachment.cgi?id=4671)
the patch make GraphWriter work on windows.

In fact, Dotty spawns another app on windows(not cygwin or __MINGW32__), too.

so 
ifdef __MINGW32__
should change to 
#if defined (__MINGW32__) || defined (_WINDOWS)

to make GraphWriter work.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 12 23:35:47 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 12 Apr 2010 23:35:47 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6826] GraphWriter delete the generated file before
 "dotty" load it.
In-Reply-To: <bug-6826-206@http.llvm.org/bugs/>
References: <bug-6826-206@http.llvm.org/bugs/>
Message-ID: <20100413043547.9D9AB2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6826

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-12 23:35:47 CDT ---
Applied in r101116, thanks!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 13 02:46:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 13 Apr 2010 02:46:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6207] clang misunderstands special member functions
 when they are declared friend.
In-Reply-To: <bug-6207-206@http.llvm.org/bugs/>
References: <bug-6207-206@http.llvm.org/bugs/>
Message-ID: <20100413074603.8C5222A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6207

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |rjmccall at apple.com
         Resolution|                            |FIXED

--- Comment #1 from John McCall <rjmccall at apple.com> 2010-04-13 02:46:03 CDT ---
Fixed in r101122.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 13 03:38:08 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 13 Apr 2010 03:38:08 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6827] New: friended template functions don't get to
 bypass protections
Message-ID: <bug-6827-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6827

           Summary: friended template functions don't get to bypass
                    protections
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: oliver at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


clang++ fails to correctly respect friend specifier in template class pointing
to a template function

struct Test { int i; };
template <typename T> class Foo;
template <typename T> Foo<T> foo(T* t){ return Foo<T>(t, true); }
template <typename T> class Foo {
public:
    Foo(T*);
    friend Foo<T> foo<T>(T*);
private:
    Foo(T*, bool);
};


Test globalT;
Foo<Test> f = foo(&globalT);

Compiles in g++ but fails in clang :-(

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 13 04:34:00 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 13 Apr 2010 04:34:00 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6828] New: Assertion failed : isStackEmpty() &&
 Stack not empty at end of basic block?
Message-ID: <bug-6828-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6828

           Summary: Assertion failed : isStackEmpty() && Stack not empty
                    at end of basic block?
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: yann at droneaud.fr
                CC: llvmbugs at cs.uiuc.edu


While building wine with llvm/clang built with Debug+Check, I've found another
ICE different from bug #3387 : according to the log, this is in pass 'X86 FP
Stackifier':

clang: X86FloatingPoint.cpp:300:
bool<unnamed>::FPS::processBasicBlock(llvm::MachineFunction&,
llvm::MachineBasicBlock&): Assertion `isStackEmpty() && "Stack not empty at end
of basic block?"' failed.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 13 10:08:32 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 13 Apr 2010 10:08:32 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6066] Ambiguity with conversion-to-base
In-Reply-To: <bug-6066-206@http.llvm.org/bugs/>
References: <bug-6066-206@http.llvm.org/bugs/>
Message-ID: <20100413150832.5D41F2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6066

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-13 10:08:32 CDT ---
Fixed in r101132.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 13 10:51:01 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 13 Apr 2010 10:51:01 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6078] Ambiguous conversion sequence in surrogate
 function call => assertion failure
In-Reply-To: <bug-6078-206@http.llvm.org/bugs/>
References: <bug-6078-206@http.llvm.org/bugs/>
Message-ID: <20100413155101.7E2BD2A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6078

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-13 10:51:01 CDT ---
Fixed in r101133.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 13 11:27:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 13 Apr 2010 11:27:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6807] llvm::SimplifyFortifiedLibCalls::fold crash
In-Reply-To: <bug-6807-206@http.llvm.org/bugs/>
References: <bug-6807-206@http.llvm.org/bugs/>
Message-ID: <20100413162731.EF9A92A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6807

Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |

--- Comment #9 from Duncan Sands <baldrick at free.fr> 2010-04-13 11:27:31 CDT ---
The original testcase is still failing.  I will attach unreduced IR in a
moment.
In the meantime, here is a new reduced testcase that crashes llc in the same
way:

; ModuleID = 'bugpoint-reduced-simplified.bc'
target datalayout =
"e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64"
target triple = "x86_64-unknown-linux-gnu"

module asm "\09.ident\09\22GCC: (GNU) 4.6.0 20100413 (experimental) LLVM:
101132M\22"

define i32 @stpncpy(i8* %__dest, i8* %__src, i64 %__n) nounwind {
entry:
  %0 = tail call i64 @llvm.objectsize.i64(i8* %__dest, i1 false) ; <i64>
[#uses=1]
  br i1 undef, label %return, label %"5"

"5":                                              ; preds = %entry
  %1 = tail call i32 (...)* @__stpncpy_chk(i8* %__dest, i8* %__src, i64 %__n,
i64 %0) nounwind ; <i32> [#uses=1]
  ret i32 %1

return:                                           ; preds = %entry
  ret i32 undef
}

declare i64 @llvm.objectsize.i64(i8*, i1) nounwind readonly

declare i32 @__stpncpy_chk(...)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 13 11:43:25 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 13 Apr 2010 11:43:25 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6807] llvm::SimplifyFortifiedLibCalls::fold crash
In-Reply-To: <bug-6807-206@http.llvm.org/bugs/>
References: <bug-6807-206@http.llvm.org/bugs/>
Message-ID: <20100413164325.79CE42A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6807

Eric Christopher <echristo at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #13 from Eric Christopher <echristo at gmail.com> 2010-04-13 11:43:24 CDT ---
Have no idea how I missed that.

Fixed in r101139.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 13 12:51:29 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 13 Apr 2010 12:51:29 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6829] New: MS extension for nameless struct members
	missing
Message-ID: <bug-6829-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6829

           Summary: MS extension for nameless struct members missing
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: amine48rz at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Here's a testcase : 

typedef struct { int x; } STRUCT1;
typedef struct { STRUCT1; int y; } STRUCT2;

clang -fms-extensions issues a warning : declaration does not declare anything

typedef struct { STRUCT1; int y; } STRUCT2;
                 ^~~~~~~

While gcc -fms-extensions doesn't.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 13 15:43:30 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 13 Apr 2010 15:43:30 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6830] New: Clang complains the qualified class name
 in a constructor call doesn't refer to a value
Message-ID: <bug-6830-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6830

           Summary: Clang complains the qualified class name in a
                    constructor call doesn't refer to a value
           Product: clang
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: zhanyong.wan at gmail.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


$ cat t.cc
namespace foo {

class X {
 public:
  X() {}
};

}  // namespace foo

class Z {
 public:
  explicit Z(const foo::X& x) {}

  void Work() {}
};

void Test() {
  Z(foo::X()).Work();
}

$ clang -c t.cc
t.cc:18:10: error: 'X' does not refer to a value
  Z(foo::X()).Work();
    ~~~~~^
t.cc:3:7: note: declared at
class X {
      ^
2 diagnostics generated.

The error goes away if we move X out of namespace foo.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 13 17:18:52 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 13 Apr 2010 17:18:52 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6827] friended template functions don't get to
 bypass protections
In-Reply-To: <bug-6827-206@http.llvm.org/bugs/>
References: <bug-6827-206@http.llvm.org/bugs/>
Message-ID: <20100413221852.B772A2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6827

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from John McCall <rjmccall at apple.com> 2010-04-13 17:18:52 CDT ---
Fixed in r101184.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 13 18:47:58 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 13 Apr 2010 18:47:58 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6468] 97592: -fblocks codegen with template
 reference arg crash
In-Reply-To: <bug-6468-206@http.llvm.org/bugs/>
References: <bug-6468-206@http.llvm.org/bugs/>
Message-ID: <20100413234758.6DC082A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6468

Fariborz Jahanian <fjahanian at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |fjahanian at apple.com
         Resolution|                            |FIXED

--- Comment #1 from Fariborz Jahanian <fjahanian at apple.com> 2010-04-13 18:47:57 CDT ---
Fixed in TOT: http://llvm.org/viewvc/llvm-project?view=rev&revision=101196

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 13 19:24:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 13 Apr 2010 19:24:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6514] valid template code rejected (friend decl.
 misinterpreted)
In-Reply-To: <bug-6514-206@http.llvm.org/bugs/>
References: <bug-6514-206@http.llvm.org/bugs/>
Message-ID: <20100414002459.BF5222A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6514

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |rjmccall at apple.com
         Resolution|                            |FIXED

--- Comment #2 from John McCall <rjmccall at apple.com> 2010-04-13 19:24:59 CDT ---
Fixed in r101198.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 14 04:37:33 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 14 Apr 2010 04:37:33 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6831] New: Spurious ambiguity error related to
 template parameter usage.
Message-ID: <bug-6831-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6831

           Summary: Spurious ambiguity error related to template parameter
                    usage.
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: zaffanella at cs.unipr.it
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


For the following test program:

namespace NA { struct S; }
namespace NB { struct S; }

using namespace NA;
using namespace NB;

template <typename S> void foo();

clang++ reports the following error:

$ clang++ -fsyntax-only bug.cc
error: reference to 'S' is ambiguous
bug.cc:1:23: note: candidate found by name lookup is 'NA::S'
namespace NA { struct S; }
                      ^
bug.cc:2:23: note: candidate found by name lookup is 'NB::S'
namespace NB { struct S; }
                      ^
1 error generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 14 08:28:14 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 14 Apr 2010 08:28:14 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6832] New: ScalarReplAggregates generate wrong
 alignement on store
Message-ID: <bug-6832-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6832

           Summary: ScalarReplAggregates generate wrong alignement on
                    store
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: quickslyver at free.fr
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4676)
 --> (http://llvm.org/bugs/attachment.cgi?id=4676)
before scalarrepl

Command used: opt bugr.bc -scalarrepl -o - | llvm-dis

The generated alignment of most stores are wrong.


$ llvm-dis -o - bugr.bc 
; ModuleID = 'bugr.bc'
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-n32"
target triple = "arm-u-u"

%0 = type { %struct.anon, %struct.anon }
%struct.anon = type { [4 x i8], [4 x i8] }

@c = external global %0                           ; <%0*> [#uses=1]

define arm_aapcscc void @good() nounwind {
entry:
  %x0 = alloca %struct.anon, align 4              ; <%struct.anon*> [#uses=2]
  %tmp = bitcast %struct.anon* %x0 to i8*         ; <i8*> [#uses=1]
  call void @llvm.memset.p0i8.i32(i8* %tmp, i8 0, i32 8, i32 4, i1 false)
  %tmp1 = bitcast %struct.anon* %x0 to i8*        ; <i8*> [#uses=1]
  call void @llvm.memcpy.p0i8.p0i8.i32(i8* getelementptr inbounds (%0* @c, i32
0, i32 0, i32 0, i32 0), i8* %tmp1, i32 8, i32 4, i1 false)
  ret void
}

declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind

declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32,
i1) nounwind

--------------------------------------------------------------------------------
$ opt bugr.bc -scalarrepl -o - | llvm-dis
; ModuleID = '<stdin>'
target datalayout =
"e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-n32"
target triple = "arm-u-u"

%0 = type { %struct.anon, %struct.anon }
%struct.anon = type { [4 x i8], [4 x i8] }

@c = external global %0                           ; <%0*> [#uses=1]

define arm_aapcscc void @good() nounwind {
entry:
  %0 = bitcast i8* getelementptr inbounds (%0* @c, i32 0, i32 0, i32 0, i32 0)
to %struct.anon* ; <%struct.anon*> [#uses=1]
  %1 = bitcast i8* getelementptr inbounds (%0* @c, i32 0, i32 0, i32 0, i32 0)
to [4 x i8]* ; <[4 x i8]*> [#uses=4]
  %.010 = getelementptr inbounds [4 x i8]* %1, i32 0, i32 0 ; <i8*> [#uses=1]
  store i8 0, i8* %.010, align 4
  %.112 = getelementptr inbounds [4 x i8]* %1, i32 0, i32 1 ; <i8*> [#uses=1]
  store i8 0, i8* %.112, align 4
  %.2 = getelementptr inbounds [4 x i8]* %1, i32 0, i32 2 ; <i8*> [#uses=1]
  store i8 0, i8* %.2, align 4
  %.3 = getelementptr inbounds [4 x i8]* %1, i32 0, i32 3 ; <i8*> [#uses=1]
  store i8 0, i8* %.3, align 4
  %.1 = getelementptr inbounds %struct.anon* %0, i32 0, i32 1 ; <[4 x i8]*>
[#uses=4]
  %.1.0 = getelementptr inbounds [4 x i8]* %.1, i32 0, i32 0 ; <i8*> [#uses=1]
  store i8 0, i8* %.1.0, align 4
  %.1.1 = getelementptr inbounds [4 x i8]* %.1, i32 0, i32 1 ; <i8*> [#uses=1]
  store i8 0, i8* %.1.1, align 4
  %.1.2 = getelementptr inbounds [4 x i8]* %.1, i32 0, i32 2 ; <i8*> [#uses=1]
  store i8 0, i8* %.1.2, align 4
  %.1.3 = getelementptr inbounds [4 x i8]* %.1, i32 0, i32 3 ; <i8*> [#uses=1]
  store i8 0, i8* %.1.3, align 4
  ret void
}

declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind

declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32,
i1) nounwind

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 14 11:23:20 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 14 Apr 2010 11:23:20 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6833] New: Combination of GVN, ScalarRepl,
 and other passes misoptimizes code
Message-ID: <bug-6833-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6833

           Summary: Combination of GVN, ScalarRepl, and other passes
                    misoptimizes code
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: Interprocedural Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: criswell at uiuc.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4677)
 --> (http://llvm.org/bugs/attachment.cgi?id=4677)
Input assembly which triggers the bug

This bug was found in the release_27 branch of LLVM using Mac OS X (Snow
Leopard).

The following command reproduces the problem:

opt -preverify -domtree -verify -scalarrepl -indvars -internalize -inline -gvn 
bugpoint-reduced-simplified.bc -f -o /tmp/f.bc

The resulting bitcode file creates a constant expression with the wrong index
(the index is extremely large).

Attached are LLVM assembly files of the input and output of the above command;
notice that the GEP constant expression for plus8dir is incorrect.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 14 11:44:17 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 14 Apr 2010 11:44:17 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6834] New: clang installs empty directories like
 usr/include/CMakeFiles
Message-ID: <bug-6834-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6834

           Summary: clang installs empty directories like
                    usr/include/CMakeFiles
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: markus at unixforces.net
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4679)
 --> (http://llvm.org/bugs/attachment.cgi?id=4679)
list of empty directories

clang from trunk installs many empty directories like "usr/include/CMakeFiles".
See attached file for a list.

The attached list is from revision 101244, but I saw this earlier this week,
too.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 14 12:04:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 14 Apr 2010 12:04:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6835] New: Makefiles.rules wrong handling of
	CXXFLAGS
Message-ID: <bug-6835-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6835

           Summary: Makefiles.rules wrong handling of CXXFLAGS
           Product: Build scripts
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Makefiles
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: julien.etienne at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Hello,

There is a mistake in the current Makefile.rules file leading to the
propagation of -fno-rtti gcc option even when it should not be set.

To reproduce the issue:
1- Get llvm trunk (I am at revision 101257)
2- Set the CXXFLAGS variable in your environment, for example:
export CXXFLAGS="-O2 -pipe -march=native"
3- Launch configure: ./configure
4- Launch make: make

You should get an error like this:
[...]
make[2]: Entering directory `/home/julien/src/llvm/utils/TableGen'
llvm[2]: Compiling ARMDecoderEmitter.cpp for Debug build
In file included from ARMDecoderEmitter.cpp:21:
Record.h: In constructor ?llvm::VarBitInit::VarBitInit(llvm::TypedInit*,
unsigned int)?:
Record.h:980: error: ?dynamic_cast? not permitted with -fno-rtti
Record.h: In constructor
?llvm::VarListElementInit::VarListElementInit(llvm::TypedInit*, unsigned int)?:
Record.h:1003: error: ?dynamic_cast? not permitted with -fno-rtti
Record.h:1005: error: ?dynamic_cast? not permitted with -fno-rtti
ARMDecoderEmitter.cpp: In function ?bit_value_t bitFromBits(llvm::BitsInit&,
unsigned int)?:
ARMDecoderEmitter.cpp:188: error: ?dynamic_cast? not permitted with -fno-rtti
make[2]: *** [/home/julien/src/llvm/utils/TableGen/Debug/ARMDecoderEmitter.o]
Error 1
make[2]: Leaving directory `/home/julien/src/llvm/utils/TableGen'
make[1]: *** [TableGen/.makeall] Error 2
make[1]: Leaving directory `/home/julien/src/llvm/utils'
make: *** [all] Error 1

This is caused by the option -fno-rtti that should not be set as
TableGen/Makefile contains "REQUIRES_RTTI := 1".

This is due to a mistake in Makefile.rules that can be fixed with the following
patch:
--- Makefile.rules      2010-04-14 18:44:09.882395333 +0200
+++ Makefile.rules.fixed        2010-04-14 18:43:56.371458018 +0200
@@ -388,7 +388,6 @@
 # If REQUIRES_RTTI=1 is specified then don't disable run-time type id.
 ifneq ($(REQUIRES_RTTI), 1)
   CXX.Flags += -fno-rtti
-  CXXFLAGS += -fno-rtti
 endif

 ifdef ENABLE_COVERAGE

The CXXFLAGS should not be set here, CXX.Flags is enough.
With this patch the REQUIRES_RTTI behavior is fine.

Best Regards,
Julien

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 14 13:25:25 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 14 Apr 2010 13:25:25 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6836] New: Fix broken tests in Makefiles for
 platforms without dynamic libs
Message-ID: <bug-6836-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6836

           Summary: Fix broken tests in Makefiles for platforms without
                    dynamic libs
           Product: new-bugs
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: reeuwijk at few.vu.nl
                CC: llvmbugs at cs.uiuc.edu


The attached patch fixes two related bugs in the current trunk.

1) The current trunk version of Makefile.config has a list of platforms that do
not support dynamic libraries, and a variable TARGET_HAS_DYNAMIC_LIBS that is
supposedly only set if the platform does support dynamic libraries.

At least, that is the intention. Unfortunately, the current code does the exact
opposite, and only sets TARGET_HAS_DYNAMIC_LIBS if the platform /is/ in the
list. The attached patch corrects this by reversing the test (ifeq -> ifneq).

Also, I wonder if the test shouldn't look at TARGET_OS, instead of HOST_OS, as
is done now. I haven't changed that.


2) The two places where TARGET_HAS_DYNAMIC_LIBS are used also have their test
in reverse, meaning that parts of the code that require dynamic libraries are
only compiled if the variable is NOT set. The attached patch also fixes this.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 14 13:28:33 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 14 Apr 2010 13:28:33 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6837] New: Null pointer reference
Message-ID: <bug-6837-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6837

           Summary: Null pointer reference
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: John.Thompson.JTSoftware at gmail.com
                CC: llvmbugs at cs.uiuc.edu


In the InitListExpr::InitListExpr in Expr.cpp I'm seeing a call to this
function from InitListChecker::getStructuredSubobjectInit where the initExprs
parameter is 0, causing an assertion failure in the collections stuff starting
from this line (approx. Expr.cpp:741):

    InitExprs.insert(C, InitExprs.end(), initExprs, initExprs+numInits);

The caller is explicitly passing 0 in the third argument:

  InitListExpr *Result
    = new (SemaRef.Context) InitListExpr(SemaRef.Context,
                                         InitRange.getBegin(), 0, 0,
                                         InitRange.getEnd());

so it seems there is an arguments mismatch or logic error.  To work-around, I
just check for a null pointer and avoid the first call above, but because I'm
compiling a huge file with lots of errors, I don't know if it's related to
improper error recovery or not, though it seems this problem happens before the
first error is emitted.

If the above is not sufficient, I can take the time to try to narrow down a
test case.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 14 15:31:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 14 Apr 2010 15:31:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6833] Combination of GVN, ScalarRepl,
 and other passes misoptimizes code
In-Reply-To: <bug-6833-206@http.llvm.org/bugs/>
References: <bug-6833-206@http.llvm.org/bugs/>
Message-ID: <20100414203131.7FA382A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6833

John T. Criswell <criswell at uiuc.edu> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #3 from John T. Criswell <criswell at uiuc.edu> 2010-04-14 15:31:31 CDT ---
I believe you are correct in that the two GEPs exhibit equivalent behavior on a
64-bit target (the result pointer is 8 bytes less than the source pointer). 
Furthermore, each GEP indexes the pointer out of the bounds of a valid memory
object, so both are equally incorrect.

I'm marking the bug resolved and invalid.

Thanks.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 14 18:59:53 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 14 Apr 2010 18:59:53 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6840] New: Friend class cannot access protected
	field of parent
Message-ID: <bug-6840-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6840

           Summary: Friend class cannot access protected field of parent
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ori at avtalion.name
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


The following code compiles in g++, but not in clang++

class A {
protected:
    static int a;
};

class B : public A {
    friend class C;
};

class C {
    void foo() { 
        A::a = 0;
    }
};


$ clang++ -c a.cpp
a.cpp:12:6: error: 'a' is a protected member of 'A'
                A::a = 0;
                   ^
a.cpp:3:13: note: declared protected here
        static int a;
                   ^
1 error generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 14 23:48:32 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 14 Apr 2010 23:48:32 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6841] New: Compiling a BLAS library causes ICE if
 optimization is used.
Message-ID: <bug-6841-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6841

           Summary: Compiling a BLAS library causes ICE if optimization is
                    used.
           Product: tools
           Version: 2.6
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: gfortran
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: source at sharpsteen.net
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4682)
 --> (http://llvm.org/bugs/attachment.cgi?id=4682)
Fortran source code for R BLAS library.  Causes ICE when compiled by
llvm-gfortran using any optimization level other than 0.

Hello,

This bug report regards llvm-gfortran version 4.2.1

I am running OS X 10.6.3 and compiled llvm version 2.6 using the sources
provided by:

  http://www.llvm.org/releases/2.6/llvm-2.6.tar.gz

I then compiled llvm-gcc from the following source:

  http://www.llvm.org/releases/2.6/llvm-gcc-4.2-2.6.source.tar.gz

The output of llvm-gfortran -v is as follows:

Using built-in specs.
Target: x86_64-apple-darwin10
Configured with: ../configure
--prefix=/Users/Sharpie/opt/brew/Cellar/llvm-gcc/4.2 --program-prefix=llvm-
--enable-llvm=/Users/Sharpie/opt/brew --enable-languages=c,c++,,fortran
--with-gmp=/Users/Sharpie/opt/brew/ --with-mpfr=/Users/Sharpie/opt/brew/
--with-gxx-include-dir=/usr/include/c++/4.2.1
--with-libiconv-prefix=/Users/Sharpie/opt/brew --build=x86_64-apple-darwin10
--host=x86_64-apple-darwin10 --target=x86_64-apple-darwin10 --with-arch=nocona
--with-tune=core2
Thread model: posix
gcc version 4.2.1 (Based on Apple Inc. build 5649) (LLVM build)

I compiled a few test benchmarks and llvm-gfortran worked with optimization
levels O1, O2, O3 and O4.  I then tried to test out the new compiler by using
it to build GNU R version 2.10. 1 at which point it fails with an ICE while
compiling the BLAS library contains in the R source.

The compilation of the file blas.f fails in isolation:

llvm-gfortran -O1 -c blas.f

Assertion failed: (Operands[i]->isLoopInvariant(l) && "Operands of AddRec must
be loop-invariant!"), function SCEVAddRecExpr, file
/Users/Sharpie/source/llvm-2.6/include/llvm/Analysis/ScalarEvolutionExpressions.h,
line 371.
blas.f:7125: internal compiler error: Abort trap
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://llvm.org/bugs/> for instructions.


Optimization levels 2 through 4 fail with identical error messages.  Disabling
optimization results in a successful compilation:

llvm-gfortran -O0 -c blas.f

gfortran 4.2.1 also succeeds:

gfortran -O3 -c blas.f


The file blas.f is attached.  Let me know if I can provide any other details!

-Charlie

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 01:33:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 01:33:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6841] Compiling a BLAS library causes ICE if
 optimization is used.
In-Reply-To: <bug-6841-206@http.llvm.org/bugs/>
References: <bug-6841-206@http.llvm.org/bugs/>
Message-ID: <20100415063351.C1FDF2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6841

Charlie Sharpsteen <source at sharpsteen.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Charlie Sharpsteen <source at sharpsteen.net> 2010-04-15 01:33:51 CDT ---
(In reply to comment #1)
> Can you try an llvm 2.7 pre-release or SVN head?  LLVM 2.6 is really really
> old.

Recompiled with llvm svn revision 101343
llvm-gcc svn revision 101333

Problem is not present.  Good luck with the 2.7 release!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 02:25:20 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 02:25:20 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6842] New: inliner inlines recursive call,
 defeats loop-deletion
Message-ID: <bug-6842-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6842

           Summary: inliner inlines recursive call, defeats loop-deletion
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Interprocedural Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


In opt -O2 on the attached testcase, there is a loop in the output which is
dead but not deleted.

The loop comes from a simple tail-recursive function.

The inliner is inlining the recursive function into itself (one level).

In this testcase, the function has a single exit, but inlining it into itself
causes it to have two exits. When tailcallelim visits it, the function turns
into a loop and the two exits turn into two loop exits. Loop deletion doesn't
know how do delete loops with multiple exits.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 04:14:00 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 04:14:00 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6843] New: analyzer not considering unsignedness of
 temporary unsigned ints
Message-ID: <bug-6843-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6843

           Summary: analyzer not considering unsignedness of temporary
                    unsigned ints
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: jultay108 at googlemail.com
                CC: llvmbugs at cs.uiuc.edu


when comparing a temporary unsigned integer in a if clause the analyzer does
not consider the unsignedness and checks impossible branches.

testcase:
called with:
scan-build gcc test.c
(svn 101237 debug mode)

#include <stdlib.h>
#include <stdio.h>

unsigned int f(void)
{
  return 2u;
}

int main(void)
{
  int * a = NULL;
  unsigned int count = f();
  // always true even if it overflows
  if ((count + 1u) >= 0u) {
    a = malloc(sizeof(int));
    if (a == NULL)
      return 0;
    *a = 1;
  }
  // analyzer gives NULL dereference warning
  printf("%d\n", *a);
  return 0;
}

it only occurs if the value in the if is a temporary
e.g.
  count = f() + 1u;
  if (count >= 0u) {
correctly handles the the always true if

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 07:28:17 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 07:28:17 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6844] New: analyzer crash when using C stdarg
Message-ID: <bug-6844-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6844

           Summary: analyzer crash when using C stdarg
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: jultay108 at googlemail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4685)
 --> (http://llvm.org/bugs/attachment.cgi?id=4685)
failure logs from the scan directory

using svn: 101364, Debug

clangs static analyzer crashes when it checks a C function using variable
number of arguments (from stdarg.h)

testcase:
#include <stdarg.h>
void
map(int srcID, ...)
{
  va_list             ap;
  int                 i;

  va_start(ap, srcID);
  for (i = 0; i < srcID; i++) {
    int v = va_arg(ap, int);
  }
  va_end(ap);
}

called with:
scan-build gcc -std=c99 -c file.c

logs attached

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 09:53:12 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 09:53:12 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6837] Null pointer reference
In-Reply-To: <bug-6837-206@http.llvm.org/bugs/>
References: <bug-6837-206@http.llvm.org/bugs/>
Message-ID: <20100415145312.C9BC92A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6837

John Thompson <John.Thompson.JTSoftware at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from John Thompson <John.Thompson.JTSoftware at gmail.com> 2010-04-15 09:53:12 CDT ---
It's fixed.  Thanks!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 10:55:04 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 10:55:04 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6846] New: Integer literal to vector cast reports
	error
Message-ID: <bug-6846-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6846

           Summary: Integer literal to vector cast reports error
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: John.Thompson.JTSoftware at gmail.com
                CC: llvmbugs at cs.uiuc.edu


gcc seems to allow this:

vector unsigned int smask = (vector unsigned int)(vector int)(-1);

But Clang reports:

error: too few elements in vector initialization (expected 4 elements, have 1)

(Using -faltivec.)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 11:52:00 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 11:52:00 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6847] New: llvm-gcc build broken for ARM due to
	debug info
Message-ID: <bug-6847-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6847

           Summary: llvm-gcc build broken for ARM due to debug info
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bob.wilson at apple.com
                CC: llvmbugs at cs.uiuc.edu


My nightly build of llvm-gcc for ARM failed last night.  I've attached a
reduced testcase, and I've verified that it still fails on trunk.  It appears
to be related to debug info.  To reproduce:

$ llc -mtriple=arm-apple-darwin < testcase.ll

The failure is:
Assertion failed: (SubUsed && "Using an undefined register!"), function
forward, file
/Users/bwilson/local/llvm/llvm/lib/CodeGen/RegisterScavenging.cpp, line 193.
0  llc               0x00000001008a4364 PrintStackTrace(void*) + 38
1  llc               0x00000001008a4844 SignalHandler(int) + 312
2  libSystem.B.dylib 0x00007fff8422beaa _sigtramp + 26
3  libSystem.B.dylib 0x00007fff841d0bea tiny_malloc_from_free_list + 1196
4  libSystem.B.dylib 0x00007fff842a7e74 __pthread_markcancel + 0
5  llc               0x00000001005cd564 llvm::RegScavenger::forward() + 1148
6  llc               0x00000001001fea01
llvm::RegScavenger::forward(llvm::ilist_iterator<llvm::MachineInstr>) + 109
7  llc               0x0000000100599f46
llvm::PEI::scavengeFrameVirtualRegs(llvm::MachineFunction&) + 1444
8  llc               0x000000010059b9c4
llvm::PEI::runOnMachineFunction(llvm::MachineFunction&) + 458
9  llc               0x000000010055a407
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) + 85
10 llc               0x00000001008224f8
llvm::FPPassManager::runOnFunction(llvm::Function&) + 348
11 llc               0x0000000100823a67
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 79
12 llc               0x0000000100823c2e
llvm::FunctionPassManager::run(llvm::Function&) + 136
13 llc               0x000000010002eb85 main + 2914
14 llc               0x000000010002d890 start + 52
15 llc               0x0000000000000002 start + 4294780838
Stack dump:
0.    Program arguments: /Users/bwilson/local/llvm/install/bin/llc
-mtriple=arm-apple-darwin 
1.    Running pass 'Prolog/Epilog Insertion & Frame Finalization' on function
'@__addvsi3'
Abort trap

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 12:25:58 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 12:25:58 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6848] New: suspicious symbols lib
	libEnhancedDisassembler.so
Message-ID: <bug-6848-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6848

           Summary: suspicious symbols lib libEnhancedDisassembler.so
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


The following symbols are among those exported from libEnhancedDisassembler.so
that are suspicious for not using one of the other namespace prefixes used by
other symbols in the library:

X86Disassembler.o:
00000000 T x86DisassemblerDebug

X86DisassemblerDecoder.o:
00002a8c T decodeInstruction
00000000 D x86DisassemblerContexts
00000000 D x86DisassemblerInstrSpecifiers
000206e0 D x86DisassemblerOneByteOpcodes
0002c6e0 D x86DisassemblerThreeByte38Opcodes
000326e0 D x86DisassemblerThreeByte3AOpcodes
000266e0 D x86DisassemblerTwoByteOpcodes

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 12:33:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 12:33:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6844] analyzer crash when using C stdarg
In-Reply-To: <bug-6844-206@http.llvm.org/bugs/>
References: <bug-6844-206@http.llvm.org/bugs/>
Message-ID: <20100415173359.2ECC12A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6844

Ted Kremenek <kremenek at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |kremenek at apple.com
         Resolution|                            |FIXED

--- Comment #2 from Ted Kremenek <kremenek at apple.com> 2010-04-15 12:33:58 CDT ---
Fixed in r101378.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 15:29:05 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 15:29:05 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6847] llvm-gcc build broken for ARM due to debug
	info
In-Reply-To: <bug-6847-206@http.llvm.org/bugs/>
References: <bug-6847-206@http.llvm.org/bugs/>
Message-ID: <20100415202905.6A5A02A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6847

Jakob Stoklund Olesen <stoklund at 2pi.dk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |stoklund at 2pi.dk
         Resolution|                            |FIXED

--- Comment #2 from Jakob Stoklund Olesen <stoklund at 2pi.dk> 2010-04-15 15:29:05 CDT ---
Fixed in r101392

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 16:22:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 16:22:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6611] (PR5718 is closed) Assertion failed:
 (ExprType->isFunctionType() && "Unknown scalar value"),
 function EmitLoadOfLValue, file CGExpr.cpp, line 579.
In-Reply-To: <bug-6611-206@http.llvm.org/bugs/>
References: <bug-6611-206@http.llvm.org/bugs/>
Message-ID: <20100415212259.A32FE2A6C12F@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6611

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #3 from Douglas Gregor <dgregor at apple.com> 2010-04-15 16:22:59 CDT ---
Closing due to the lack of a test case. Please re-open with a test case if the
problem persists.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 18:48:22 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 18:48:22 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6831] Clang is confused by template parameter name
 shadowing other declarations
In-Reply-To: <bug-6831-206@http.llvm.org/bugs/>
References: <bug-6831-206@http.llvm.org/bugs/>
Message-ID: <20100415234822.473FD2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6831

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-15 18:48:22 CDT ---
Fixed in r101419.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 19:51:12 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 19:51:12 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6848] suspicious symbols lib
	libEnhancedDisassembler.so
In-Reply-To: <bug-6848-206@http.llvm.org/bugs/>
References: <bug-6848-206@http.llvm.org/bugs/>
Message-ID: <20100416005112.B60002A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6848

Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Dan Gohman <gohman at apple.com> 2010-04-15 19:51:12 CDT ---
I did some Makefilery and implement that, so this is now fixed.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 20:06:28 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 20:06:28 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6832] ScalarReplAggregates generate wrong
	alignement on store
In-Reply-To: <bug-6832-206@http.llvm.org/bugs/>
References: <bug-6832-206@http.llvm.org/bugs/>
Message-ID: <20100416010629.1483C2A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6832

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-15 20:06:28 CDT ---
Fixed here, thanks!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20100412/099896.html

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 20:53:02 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 20:53:02 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6850] New: Unhelpful diagnostic when template calls
 later-declared function not found by ADL
Message-ID: <bug-6850-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6850

           Summary: Unhelpful diagnostic when template calls
                    later-declared function not found by ADL
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


When a template calls another function, the target of the call has to be
declared before the implementation, unless it's found by ADL, in which case it
can be declared later. Clang's error message doesn't provide either of those
hints, and it should. This has confused at least one clang early adopter inside
of google.


$ cat test.cc
#include <utility>
using std::pair;

void my_template(int) {}

#ifdef WORK
template<typename T, typename U>
void my_template(const pair<T, U>& p);
#endif

template<typename T>
void call_template(const T& t) {
    my_template(t);
}

#ifdef WORK2
namespace std {
#endif
template<typename T, typename U>
void my_template(const pair<T, U>& p) {
}
#ifdef WORK2
}
#endif

void foo() {
    pair<int, int> p(3, 4);
    call_template(p);
}
$ clang++  -c test.cc
test.cc:13:5: error: no matching function for call to 'my_template'
    my_template(t);
    ^~~~~~~~~~~
test.cc:28:5: note: in instantiation of function template specialization
'call_template<std::pair<int, int> >' requested here
    call_template(p);
    ^
test.cc:4:6: note: candidate function not viable: no known conversion from
'std::pair<int, int> const' to 'int' for 1st argument
void my_template(int) {}
     ^
1 error generated.
$ clang++  -c test.cc -DWORK
$ clang++  -c test.cc -DWORK2
$

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 15 23:57:06 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 15 Apr 2010 23:57:06 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6501] QOI: arguments passed/required mismatch
	diagnostic
In-Reply-To: <bug-6501-206@http.llvm.org/bugs/>
References: <bug-6501-206@http.llvm.org/bugs/>
Message-ID: <20100416045706.6ADB12A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6501

Eric Christopher <echristo at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #4 from Eric Christopher <echristo at gmail.com> 2010-04-15 23:57:06 CDT ---
Finished off thusly:

[issola:llvm/tools/clang] echristo% svn ci
Sending        include/clang/Basic/DiagnosticSemaKinds.td
Sending        lib/Sema/SemaChecking.cpp
Sending        lib/Sema/SemaExpr.cpp
Sending        lib/Sema/SemaExprObjC.cpp
Transmitting file data ....
Committed revision 101442.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 16 04:52:46 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 16 Apr 2010 04:52:46 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6851] New: Spurious mismatch of out-of-line method
 definition and method declaration.
Message-ID: <bug-6851-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6851

           Summary: Spurious mismatch of out-of-line method definition and
                    method declaration.
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: zaffanella at cs.unipr.it
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


The following program:

template <bool v>
struct S;

struct N {
  template <bool w>
  S< S<w>::cond && 1 > foo();
};

struct Alien;
bool operator&&(const Alien&, const Alien&);

template <bool w>
S< S<w>::cond && 1 > N::foo() { }


generates a spurious error in clang++ (r101446):

$ llvm/Debug/bin/clang++ -fsyntax-only bug.cc
bug.cc:13:25: error: out-of-line definition of 'foo' does not match any
      declaration in 'N'
S< S<w>::cond && 1 > N::foo() { }
                     ~~~^
1 error generated.

If the declaration of operator&& is commented out, the error disappears.
The same happens if the declaration of operator&& is moved before the
declaration of N or after the definition of N::foo().

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 16 10:51:02 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 16 Apr 2010 10:51:02 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6854] New: analyzer: Assertion
 `castTy->isIntegerType() fails when doing reinterpret casting
Message-ID: <bug-6854-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6854

           Summary: analyzer: Assertion `castTy->isIntegerType() fails
                    when doing reinterpret casting
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: jultay108 at googlemail.com
                CC: llvmbugs at cs.uiuc.edu


svn 101464 (Debug)

I get the assertion
clang: SimpleSValuator.cpp:116: virtual
clang::SVal<unnamed>::SimpleSValuator::EvalCastL(clang::Loc, clang::QualType):
Assertion `castTy->isIntegerType()' failed

when analyzing following code with scan-build gcc -c file.c:

void
crash(void * arg)
{
  void * a = arg;
  *(void**)a = arg;
  float f = *(float*) a;
}

it does not crash when one skips the assignment on the dereferenced left side
pointer or the reinterpret float cast.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 16 11:01:46 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 16 Apr 2010 11:01:46 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6842] inliner inlines recursive call,
	defeats loop-deletion
In-Reply-To: <bug-6842-206@http.llvm.org/bugs/>
References: <bug-6842-206@http.llvm.org/bugs/>
Message-ID: <20100416160146.63B562A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6842

Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Dan Gohman <gohman at apple.com> 2010-04-16 11:01:46 CDT ---
Fixed in r101471.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 16 11:53:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 16 Apr 2010 11:53:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6634] Consumes basically all of the virtual memory
In-Reply-To: <bug-6634-206@http.llvm.org/bugs/>
References: <bug-6634-206@http.llvm.org/bugs/>
Message-ID: <20100416165303.951FD2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6634

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME

--- Comment #2 from Douglas Gregor <dgregor at apple.com> 2010-04-16 11:53:03 CDT ---
This is no longer crashing for me. Please re-open if it's still causing
problems for you.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 16 12:01:46 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 16 Apr 2010 12:01:46 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6461] wrong error: functions that differ only in
 their return type cannot be overloaded
In-Reply-To: <bug-6461-206@http.llvm.org/bugs/>
References: <bug-6461-206@http.llvm.org/bugs/>
Message-ID: <20100416170146.140AD2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6461

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #9 from Douglas Gregor <dgregor at apple.com> 2010-04-16 12:01:45 CDT ---
This is core issue 1043.

Informal discussion seemed to favor Clang's behavior here (rejecting the code),
so I'm going to close this bug as INVALID. In the (IMO, unlikely) event that
core issue 1043 is resolved the other way, we'll change Clang at this point.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 16 12:55:06 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 16 Apr 2010 12:55:06 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6854] crash when casting symbolic memory address to
	a float
In-Reply-To: <bug-6854-206@http.llvm.org/bugs/>
References: <bug-6854-206@http.llvm.org/bugs/>
Message-ID: <20100416175506.6449A2A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6854

Ted Kremenek <kremenek at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Ted Kremenek <kremenek at apple.com> 2010-04-16 12:55:06 CDT ---
Fixed in r101499.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 16 14:01:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 16 Apr 2010 14:01:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6855] New: friend does not have access to private
 enum in declaration
Message-ID: <bug-6855-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6855

           Summary: friend does not have access to private enum in
                    declaration
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: evan at chromium.org
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4688)
 --> (http://llvm.org/bugs/attachment.cgi?id=4688)
code (also pasted in bug)

[class.access] p5 in C++98 has an example and says this should be well-formed.

class C {
  enum X {
    X_VALUE
  };

  friend void foo(X);
};

// BUG: C::X is not available here
void foo(C::X x) {
  // C::X is correctly available here
  C::X x2 = C::X_VALUE;
}


$ clang++ test.cc
test.cc:12:13: error: 'X' is a private member of 'C'
void foo(C::X x) {
            ^
test.cc:4:8: note: declared private here
  enum X {
       ^
1 error generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 16 18:19:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 16 Apr 2010 18:19:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6766] clang crash with a complex C99 structure
	initialisation
In-Reply-To: <bug-6766-206@http.llvm.org/bugs/>
References: <bug-6766-206@http.llvm.org/bugs/>
Message-ID: <20100416231956.0D56D2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6766

Nuno Lopes <nunoplopes at sapo.pt> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |nunoplopes at sapo.pt
         Resolution|                            |FIXED

--- Comment #1 from Nuno Lopes <nunoplopes at sapo.pt> 2010-04-16 18:19:55 CDT ---
fixed in r101568.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 00:05:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 00:05:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6665] asm variables
In-Reply-To: <bug-6665-206@http.llvm.org/bugs/>
References: <bug-6665-206@http.llvm.org/bugs/>
Message-ID: <20100417050559.3325B2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6665

Nick Lewycky <nicholas at mxc.ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #1 from Nick Lewycky <nicholas at mxc.ca> 2010-04-17 00:05:58 CDT ---


*** This bug has been marked as a duplicate of bug 3933 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 01:54:26 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 01:54:26 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6766] clang crash with a complex C99 structure
	initialisation
In-Reply-To: <bug-6766-206@http.llvm.org/bugs/>
References: <bug-6766-206@http.llvm.org/bugs/>
Message-ID: <20100417065426.3D5562A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6766

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-17 01:54:25 CDT ---
The patch for this caused several miscompilations, and had to be reverted in
r101623

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 04:43:02 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 04:43:02 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6856] New: X86ELFWriterInfo.ctor: unneeded local
	var is64Bit
Message-ID: <bug-6856-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6856

           Summary: X86ELFWriterInfo.ctor: unneeded local var is64Bit
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: le_afo at gmx.net
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4689)
 --> (http://llvm.org/bugs/attachment.cgi?id=4689)
uni diff for X86ELFWriterInfo.cpp

X86ELFWriterInfo::X86ELFWriterInfo(TargetMachine &TM) :
bool is64Bit = TM.getTargetData()->getPointerSizeInBits() == 64;

local var is64Bit is not needed as its a member variable of parent class
TargetELFWriterInfo and will be initialized in TargetELFWriterInfo.ctor
the same way

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 05:41:06 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 05:41:06 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6857] New: Expression statement in local scope
 causes inaccurate diagnostics.
Message-ID: <bug-6857-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6857

           Summary: Expression statement in local scope causes inaccurate
                    diagnostics.
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: zaffanella at cs.unipr.it
                CC: llvmbugs at cs.uiuc.edu


The following program, using non-standard extensions:

void foo(void) {
  struct s {
    int i[({typedef int Int; 3;})];
  };
}

causes the generation of the following inaccurate diagnostic ("file scope"):

llvm/Debug/bin/clang -fsyntax-only bug.c
bug.c:3:11: error: statement expression not allowed at file scope
    int i[({typedef int Int; 3;})];
          ^
1 error generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 10:55:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 10:55:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6858] New: Use of freed memory in
 InlineCostAnalyzer::getInlineCost
Message-ID: <bug-6858-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6858

           Summary: Use of freed memory in
                    InlineCostAnalyzer::getInlineCost
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4691)
 --> (http://llvm.org/bugs/attachment.cgi?id=4691)
testcase .ll

This is the reason changing std::map into ValueMap broke the dragonegg build.
It looks like the memory was freed when growing the ValueMap:

$ valgrind opt -inline -disable-output use_after_free.ll
==31149== Memcheck, a memory error detector
==31149== Copyright (C) 2002-2009, and GNU GPL'd, by Julian Seward et al.
==31149== Using Valgrind-3.6.0.SVN and LibVEX; rerun with -h for copyright info
==31149== Command: opt -inline -disable-output use_after_free.ll
==31149== 
==31149== Invalid read of size 4
==31149==    at 0x9D1081:
llvm::InlineCostAnalyzer::getInlineCost(llvm::CallSite,
llvm::SmallPtrSet<llvm::Function const*, 16u>&) (InlineCost.cpp:366)
==31149==    by 0x8039EF: (anonymous
namespace)::SimpleInliner::getInlineCost(llvm::CallSite) (InlineSimple.cpp:40)
==31149==    by 0x8046D4: llvm::Inliner::shouldInline(llvm::CallSite)
(Inliner.cpp:204)
==31149==    by 0x8053D4: llvm::Inliner::runOnSCC(llvm::CallGraphSCC&)
(Inliner.cpp:380)
==31149==    by 0x9985A2: (anonymous
namespace)::CGPassManager::RunPassOnSCC(llvm::Pass*, llvm::CallGraphSCC&,
llvm::CallGraph&, bool&) (CallGraphSCCPass.cpp:109)
==31149==    by 0x9994E8: (anonymous
namespace)::CGPassManager::runOnModule(llvm::Module&)
(CallGraphSCCPass.cpp:355)
==31149==    by 0xB7CF12: llvm::MPPassManager::runOnModule(llvm::Module&)
(PassManager.cpp:1492)
==31149==    by 0xB7D452: llvm::PassManagerImpl::run(llvm::Module&)
(PassManager.cpp:1573)
==31149==    by 0xB7D98A: llvm::PassManager::run(llvm::Module&)
(PassManager.cpp:1610)
==31149==    by 0x7B46A5: main (opt.cpp:544)
==31149==  Address 0x62d73c0 is 7,280 bytes inside a block of size 7,680 free'd
==31149==    at 0x4C26E54: operator delete(void*) (vg_replace_malloc.c:387)
==31149==    by 0x802856:
llvm::DenseMap<llvm::ValueMapCallbackVH<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> >,
llvm::InlineCostAnalyzer::FunctionInfo,
llvm::DenseMapInfo<llvm::ValueMapCallbackVH<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> > >,
llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> >::grow(unsigned
int) (DenseMap.h:402)
==31149==    by 0x8022EC:
llvm::DenseMap<llvm::ValueMapCallbackVH<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> >,
llvm::InlineCostAnalyzer::FunctionInfo,
llvm::DenseMapInfo<llvm::ValueMapCallbackVH<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> > >,
llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo>
>::InsertIntoBucket(llvm::ValueMapCallbackVH<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> > const&,
llvm::InlineCostAnalyzer::FunctionInfo const&,
std::pair<llvm::ValueMapCallbackVH<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> >,
llvm::InlineCostAnalyzer::FunctionInfo>*) (DenseMap.h:282)
==31149==    by 0x801B30:
llvm::DenseMap<llvm::ValueMapCallbackVH<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> >,
llvm::InlineCostAnalyzer::FunctionInfo,
llvm::DenseMapInfo<llvm::ValueMapCallbackVH<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> > >,
llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo>
>::FindAndConstruct(llvm::ValueMapCallbackVH<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> > const&)
(DenseMap.h:207)
==31149==    by 0x80132E:
llvm::DenseMap<llvm::ValueMapCallbackVH<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> >,
llvm::InlineCostAnalyzer::FunctionInfo,
llvm::DenseMapInfo<llvm::ValueMapCallbackVH<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> > >,
llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo>
>::operator[](llvm::ValueMapCallbackVH<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo> > const&)
(DenseMap.h:211)
==31149==    by 0x800E1F: llvm::ValueMap<llvm::Function const*,
llvm::InlineCostAnalyzer::FunctionInfo, llvm::ValueMapConfig<llvm::Function
const*>, llvm::DenseMapInfo<llvm::InlineCostAnalyzer::FunctionInfo>
>::operator[](llvm::Function const* const&) (ValueMap.h:161)
==31149==    by 0x9D0F40:
llvm::InlineCostAnalyzer::getInlineCost(llvm::CallSite,
llvm::SmallPtrSet<llvm::Function const*, 16u>&) (InlineCost.cpp:319)
==31149==    by 0x8039EF: (anonymous
namespace)::SimpleInliner::getInlineCost(llvm::CallSite) (InlineSimple.cpp:40)
==31149==    by 0x8046D4: llvm::Inliner::shouldInline(llvm::CallSite)
(Inliner.cpp:204)
==31149==    by 0x8053D4: llvm::Inliner::runOnSCC(llvm::CallGraphSCC&)
(Inliner.cpp:380)
==31149==    by 0x9985A2: (anonymous
namespace)::CGPassManager::RunPassOnSCC(llvm::Pass*, llvm::CallGraphSCC&,
llvm::CallGraph&, bool&) (CallGraphSCCPass.cpp:109)
==31149==    by 0x9994E8: (anonymous
namespace)::CGPassManager::runOnModule(llvm::Module&)
(CallGraphSCCPass.cpp:355)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 12:28:44 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 12:28:44 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6856] X86ELFWriterInfo.ctor: unneeded local var
	is64Bit
In-Reply-To: <bug-6856-206@http.llvm.org/bugs/>
References: <bug-6856-206@http.llvm.org/bugs/>
Message-ID: <20100417172844.054812A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6856

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-17 12:28:43 CDT ---
Applied in r101648, thanks.  Please send patches like this directly to the
llvm-commits list with "[PATCH]" in the subject line, no need to open a PR
unless you think it is controversial.  Thanks!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 12:58:16 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 12:58:16 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6858] Use of freed memory in
	InlineCostAnalyzer::getInlineCost
In-Reply-To: <bug-6858-206@http.llvm.org/bugs/>
References: <bug-6858-206@http.llvm.org/bugs/>
Message-ID: <20100417175816.BF3672A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6858

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Chris Lattner <clattner at apple.com> 2010-04-17 12:58:16 CDT ---
Fixed in a few patches leading up to r101658.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 14:02:53 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 14:02:53 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6332] Assertion failed: (Ty->isSized() && "Cannot
 getTypeInfo() on a type that is unsized!")
In-Reply-To: <bug-6332-206@http.llvm.org/bugs/>
References: <bug-6332-206@http.llvm.org/bugs/>
Message-ID: <20100417190253.322C02A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6332

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #4 from Chris Lattner <clattner at apple.com> 2010-04-17 14:02:52 CDT ---
Fixed in r101662, thanks for the great testcase!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 15:46:05 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 15:46:05 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6491] asmparser should check types on constant
	forward refs
In-Reply-To: <bug-6491-206@http.llvm.org/bugs/>
References: <bug-6491-206@http.llvm.org/bugs/>
Message-ID: <20100417204605.974AD2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6491

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-17 15:46:05 CDT ---
Fixed in r101670

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 21:48:08 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 21:48:08 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6860] New: ice: Live range not found
Message-ID: <bug-6860-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6860

           Summary: ice: Live range not found
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


As below.  Seen on x64 Linux.  

[regehr at gamow tmp403]$ clang -O1 small.c

clang: SimpleRegisterCoalescing.cpp:646: bool
llvm::SimpleRegisterCoalescing::ReMaterializeTrivialDef(llvm::LiveInterval&,
unsigned int, unsigned int, llvm::MachineInstr*): Assertion `SrcLR !=
SrcInt.end() && "Live range not found!"' failed.
0  clang           0x0000000001246a0f
1  clang           0x0000000001248912
2  libpthread.so.0 0x00007fbb0fbc3190
3  libc.so.6       0x00007fbb0eec94b5 gsignal + 53
4  libc.so.6       0x00007fbb0eeccf50 abort + 384
5  libc.so.6       0x00007fbb0eec2481 __assert_fail + 241
6  clang           0x0000000000ef2f8f
7  clang           0x0000000000ef35cb
8  clang           0x0000000000ef80af
9  clang           0x0000000000ef94a7
10 clang           0x0000000000ef9c46
11 clang           0x0000000000efaaa8
12 clang           0x00000000011ca3fd
13 clang           0x00000000011ca5db
14 clang           0x00000000011ca89f
15 clang           0x0000000000415357
16 clang           0x0000000000415bd5
17 clang           0x000000000060eb09
18 clang           0x0000000000419721
19 clang           0x00000000004099bb
20 clang           0x000000000040cd2b main + 2491
21 libc.so.6       0x00007fbb0eeb4abd __libc_start_main + 253
22 clang           0x0000000000407199
Stack dump:
0.    Program arguments:
/uusoc/exports/scratch/regehr/z/compiler-install/llvm-gcc-r101534-install/bin/clang
-cc1 -triple x86_64-unknown-linux-gnu -S -disable-free -main-file-name small.c
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -resource-dir
/uusoc/exports/scratch/regehr/z/compiler-install/llvm-gcc-r101534-install/lib/clang/1.5
-O1 -ferror-limit 19 -fmessage-length 80 -fgnu-runtime
-fdiagnostics-show-option -fcolor-diagnostics -o /tmp/cc-ELwZYl.s -x c small.c 
1.    <eof> parser at end of file
2.    Code generation
3.    Running pass 'Simple Register Coalescing' on function '@func_1'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

[regehr at gamow tmp403]$ cat small.c

typedef short int int16_t;
typedef int int32_t;
typedef unsigned short int uint16_t;
typedef unsigned int uint32_t;
struct S0
{
  int32_t f0;
  volatile uint16_t f1;
  const uint32_t f3;
  volatile int16_t f4;
  volatile uint32_t f6;
  uint32_t f7;
};
struct S2
{
  volatile struct S0 f2;
  volatile struct S0 f3;
  volatile struct S0 f5;
};
struct S2 g_557 = {
};

struct S2
func_1 (void)
{
  return g_557;
}

[regehr at gamow tmp403]$ clang -v

clang version 1.5 (trunk 101534)
Target: x86_64-unknown-linux-gnu
Thread model: posix

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 21:54:50 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 21:54:50 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6861] New: possibly bogus asm generation on x64
Message-ID: <bug-6861-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6861

           Summary: possibly bogus asm generation on x64
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu, chenyang at cs.utah.edu


[regehr at gamow tmp404]$ clang -O2 small.c -c -w

/tmp/cc-ETcUQX.s: Assembler messages:
/tmp/cc-ETcUQX.s:40: Error: suffix or operands invalid for `movsbq'
/tmp/cc-ETcUQX.s:41: Error: suffix or operands invalid for `movq'
/tmp/cc-ETcUQX.s:44: Error: suffix or operands invalid for `xor'
/tmp/cc-ETcUQX.s:49: Error: Incorrect register `%esi' used with `q' suffix
/tmp/cc-ETcUQX.s:51: Error: Incorrect register `%esi' used with `q' suffix
clang: error: assembler command failed with exit code 1 (use -v to see
invocation)

[regehr at gamow tmp404]$ cat small.c

typedef signed char int8_t;
typedef int int32_t;
typedef long int int64_t;
typedef unsigned short int uint16_t;
static int8_t
safe_sub_func_int64_t_s_s (int64_t si1, int8_t si2)
{
  return (si1 ^ si2) & (si1 ^ (si1 ^ si2) << 1) - si2 ^ si2 ? : si1 - si2;
}

int8_t g_7;
int8_t *g_19 = &g_7;
int8_t **g_30 = &g_19;
int8_t **const *g_29 = &g_30;
int8_t g_34[2];
int32func (int32_t * **p_50, int8_t p_51)
{
  int8_t l_253;
  int8_t l_255;
  uint16_t l_282[3][4];
  int i, j, k, l, m;
  for (i = 0; i < 1; i++)
    for (i = 0; i < 1; i++)
      l_282[i][j];
  if (0)
    {
    }
  else
    l_253 = ***g_29;
  if (l_253)
    goto lbl_264;
  for (0; p_51; p_51 = safe_sub_func_int64_t_s_s (p_51, 1))
    {
      for (l; l_255; l_255 = safe_sub_func_int64_t_s_s)
      lbl_264:g_34[0] &= 1;
    }
}

[regehr at gamow tmp404]$ clang -v

clang version 1.5 (trunk 101534)
Target: x86_64-unknown-linux-gnu
Thread model: posix

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 22:16:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 22:16:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6824] crash in X86 DAG->DAG Instruction Selection
In-Reply-To: <bug-6824-206@http.llvm.org/bugs/>
References: <bug-6824-206@http.llvm.org/bugs/>
Message-ID: <20100418031618.C00652A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6824

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #7 from Chris Lattner <clattner at apple.com> 2010-04-17 22:16:18 CDT ---
Sounds like your system compiler is miscompiling llvm.  This works for me in
debug, release, and release-asserts builds.

If clang successfully bootstraps for you, you could build it at -O0 then use
that to build clang at -O3.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 22:38:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 22:38:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6862] New: ice: Type is not SCEVable
Message-ID: <bug-6862-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6862

           Summary: ice: Type is not SCEVable
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu, chenyang at cs.utah.edu


Created an attachment (id=4692)
 --> (http://llvm.org/bugs/attachment.cgi?id=4692)
failure inducing C code

This is an odd case-- slightly different versions of the same test input were
producing several different error messages and also segfaults.  So perhaps it's
a memory safety problem leading to data structure corruption.

regehr at john-home:~/volatile/bugs/tmp313$ clang -O1 small.c
small.c:70:2: warning: implicit declaration of function 'func_17' is invalid in
      C99 [-Wimplicit-function-declaration]
        func_17 ((((*g_16) >=
        ^
small.c:69:15: warning: incompatible integer to pointer conversion assigning to
      'int32_t *' from 'int'
      (*l_44) =
              ^
clang: ScalarEvolution.cpp:2372: const llvm::Type*
llvm::ScalarEvolution::getEffectiveSCEVType(const llvm::Type*) const: Assertion
`isSCEVable(Ty) && "Type is not SCEVable!"' failed.
0  clang 0x08fc5438
Stack dump:
0.    Program arguments:
/home/regehr/z/compiler-install/llvm-gcc-r101617-install/bin/clang -cc1 -triple
i386-pc-linux-gnu -S -disable-free -main-file-name small.c -mrelocation-model
static -mdisable-fp-elim -mconstructor-aliases -target-cpu pentium4
-resource-dir
/home/regehr/z/compiler-install/llvm-gcc-r101617-install/lib/clang/1.5 -O1
-ferror-limit 19 -fmessage-length 80 -fgnu-runtime -fdiagnostics-show-option
-fcolor-diagnostics -o /tmp/cc-wAmn5u.s -x c small.c 
1.    <eof> parser at end of file
2.    Per-module optimization passes
3.    Running pass 'CallGraph Pass Manager' on module 'small.c'.
4.    Running pass 'Loop Pass Manager' on function '@func_10'
5.    Running pass 'Canonicalize Induction Variables' on basic block
'%for.body17'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

regehr at john-home:~/volatile/bugs/tmp313$ clang -v

clang version 1.5 (trunk 101617)
Target: i386-pc-linux-gnu
Thread model: posix

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 23:01:34 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 23:01:34 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6863] New: crash on invalid
Message-ID: <bug-6863-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6863

           Summary: crash on invalid
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


This crashes clang++:


bool test(int FD,  foo_t &FileActions) {

  if (int Err = foo(&FileActions))
    ;
  return false;
}

$ clang++ t.ii
t.ii:2:20: error: unknown type name 'foo_t'
bool test(int FD,  foo_t &FileActions) {
                   ^
t.ii:5:5: warning: if statement has empty body [-Wempty-body]
    ;
    ^
Assertion failed: (RefCount >= 1 && "Referencing already-destroyed
statement!"), function getStmtClass, file
/Users/sabre/llvm/tools/clang/lib/Sema/../../include/clang/AST/Stmt.h, line
202.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 17 23:05:30 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 17 Apr 2010 23:05:30 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6864] New: poor recovery after typo
Message-ID: <bug-6864-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6864

           Summary: poor recovery after typo
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


In this example:

typedef int mytype_t;

int foo() {
  mytype_tt x;
  return x+x;
} 

I get two errors:

t.cpp:5:3: error: use of undeclared identifier 'mytype_tt'; did you mean
      'mytype_t'?
  mytype_tt x;
  ^
t.cpp:6:10: error: use of undeclared identifier 'x'
  return x+x;
         ^

The first one is great (but missing a fixit), the second is completely bogus. 
Clang should recover and continue parsing the code as if I wrote "mytype_t".

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 18 00:00:21 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 18 Apr 2010 00:00:21 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6863] crash on invalid
In-Reply-To: <bug-6863-206@http.llvm.org/bugs/>
References: <bug-6863-206@http.llvm.org/bugs/>
Message-ID: <20100418050021.244C52A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6863

Nick Lewycky <nicholas at mxc.ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |nicholas at mxc.ca
         Resolution|                            |DUPLICATE

--- Comment #1 from Nick Lewycky <nicholas at mxc.ca> 2010-04-18 00:00:20 CDT ---


*** This bug has been marked as a duplicate of bug 6720 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 18 03:34:42 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 18 Apr 2010 03:34:42 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6855] friend does not have access to private enum
 in declaration
In-Reply-To: <bug-6855-206@http.llvm.org/bugs/>
References: <bug-6855-206@http.llvm.org/bugs/>
Message-ID: <20100418083442.EF27B2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6855

Chandler Carruth <chandlerc at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |chandlerc at gmail.com
         Resolution|                            |FIXED

--- Comment #1 from Chandler Carruth <chandlerc at gmail.com> 2010-04-18 03:34:42 CDT ---
Fixed in r101706.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 18 04:22:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 18 Apr 2010 04:22:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6177] Reference binding failure affects viability
In-Reply-To: <bug-6177-206@http.llvm.org/bugs/>
References: <bug-6177-206@http.llvm.org/bugs/>
Message-ID: <20100418092251.C23412A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6177

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Douglas Gregor <dgregor at apple.com> 2010-04-18 04:22:51 CDT ---
Fixed in r101712.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 18 06:04:14 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 18 Apr 2010 06:04:14 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6865] New: Stop give vodka to static analyzer :)
Message-ID: <bug-6865-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6865

           Summary: Stop give vodka to static analyzer :)
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: aytharn at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Definitively a totally nonsense error, or I am wrong?
See the attachment to see in the class Block.m what clang tell me as error!!!!
(both with SDK 3.2 and 4.0beta)

1) First of all, the value is incremented in each line (and to increment
something you need either to read the prev value or to use black magic, my
guess is clang does the first way...)

2) Second after this block is created I MUST use this value FOREACH block with
no exceptions to add it in the 2D engine Cocos2d... Look at the [self
addChild...] I use to pass this value generated before, how this crazy clang
could think that the value is never read???

for (int i = 0; i < [playerOne.currentBlock.bricks count]; i++) {
        Brick *brick = [playerOne.currentBlock.bricks objectAtIndex:i];
        brick.bColumn += 4;
        brick.bRow += 18;
        brick.position = ccp(pad + brick.bColumn * (side + 3), pad + brick.bRow
* side); 
        [blockOne.bricks addObject:brick];
        [self addChild:[playerOne.currentBlock.bricks objectAtIndex:i] z:1
tag:brick.myTag];
    }


Thank you!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 18 06:16:42 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 18 Apr 2010 06:16:42 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6866] New: MSVC9 hangs on compilation of
	X86Disassembler.cpp
Message-ID: <bug-6866-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6866

           Summary: MSVC9 hangs on compilation of X86Disassembler.cpp
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sothman at web.de
                CC: llvmbugs at cs.uiuc.edu


When building in release mode with MSVC9, compilation hangs on
lib\Target\X86\Disassembler\X86Disassembler.cpp

The issue doesn't occur on a debug build, and can in fact be circumvented by
compiling without optimizations, so this appears actually to be a problem with
the MSVC9 optimizer rather than a bug in the LLVM, but it's probably more
practical and timely possible to work around than wait for a MSVC bugfix.

It looks like the problem is related to r101179, more specifically to
lib\Target\X86\X86GenEDInfo.inc and its instInfoX86 array, which MSVC's
optimizer seems to have trouble with.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 18 07:03:12 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 18 Apr 2010 07:03:12 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6867] New: Unexpected linkage decl
Message-ID: <bug-6867-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6867

           Summary: Unexpected linkage decl
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: sidorenko at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Hi,

I'm trying to build Etoile trunk (rev 6032)
(svn://svn.gna.org/svn/etoile/trunk/Etoile) using latest clang/llvm (rev
101709) on Fedora 12. Here is what happens:

clang version 1.5 (trunk 101709)
Target: i386-pc-linux-gnu
Thread model: posix
 "/usr/local/bin/clang" -cc1 -triple i386-pc-linux-gnu -S -disable-free
-main-file-name LLVMCodeGen.mm -pic-level 2 -mdisable-fp-elim
-mconstructor-aliases -target-cpu pentium4 -v -g -resource-dir
/usr/local/lib/clang/1.5 -dependency-file
obj/LanguageKitCodeGen.obj/LLVMCodeGen.mm.d -MT
obj/LanguageKitCodeGen.obj/LLVMCodeGen.mm.o -MP -DGNU_RUNTIME_VERSION=1
-DGNU_RUNTIME_VERSION=1 -D_GNU_SOURCE -DGNU_RUNTIME_VERSION=1 -D_GNU_SOURCE
-DGNU_RUNTIME_VERSION=1 -DGNUSTEP -DGNUSTEP_BASE_LIBRARY=1 -DGNU_GUI_LIBRARY=1
-DGNU_RUNTIME=1 -DGNUSTEP_BASE_LIBRARY=1 -D_REENTRANT -DGSWARN -DGSDIAGNOSE
-D_DEBUG -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS
-I/usr/lib/gcc/i686-redhat-linux/4.4.3/include -I/usr/local/include
-I/home/andrey/projects/Etoile/Build
-I/home/andrey/projects/Etoile/Languages/LanguageKit
-I/home/andrey/projects/Etoile/Languages/LanguageKit/Headers
-I/home/andrey/projects/Etoile/Build
-I/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen
-I/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen/Headers
-I/home/andrey/projects/Etoile/Build
-I/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen
-I/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen/Headers
-I/home/andrey/projects/Etoile/Build
-I/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen
-I/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen/Headers
-I./derived_src -I. -I/home/andrey/GNUstep/Library/Headers
-I/usr/GNUstep/Local/Library/Headers -I/usr/GNUstep/System/Library/Headers -O2
-Wall -Wno-import -Woverloaded-virtual -w -fconstant-string-class
NSConstantString -ferror-limit 19 -fmessage-length 0 -fno-rtti -fgnu-runtime
-fdiagnostics-show-option -o /tmp/cc-JTlMvF.s -x objective-c++ LLVMCodeGen.mm
clang -cc1 version 1.5 based upon llvm 2.8svn hosted on i386-pc-linux-gnu
ignoring nonexistent directory
"/home/andrey/projects/Etoile/Languages/LanguageKit/Headers"
ignoring nonexistent directory
"/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen/Headers"
ignoring nonexistent directory
"/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen/Headers"
ignoring nonexistent directory
"/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen/Headers"
ignoring nonexistent directory "/home/andrey/GNUstep/Library/Headers"
ignoring nonexistent directory "/usr/include/c++/4.4.3/x86_64-pc-linux-gnu/32"
ignoring nonexistent directory "/usr/include/c++/4.4.3/i686-pc-linux-gnu/"
ignoring nonexistent directory "/usr/include/c++/4.4"
ignoring nonexistent directory "/usr/include/c++/4.4/x86_64-linux-gnu/32"
ignoring nonexistent directory "/usr/include/c++/4.4/backward"
ignoring nonexistent directory "/usr/include/c++/4.4"
ignoring nonexistent directory "/usr/include/c++/4.4/i486-linux-gnu/64"
ignoring nonexistent directory "/usr/include/c++/4.4/backward"
ignoring nonexistent directory "/usr/include/c++/4.1.3"
ignoring nonexistent directory "/usr/include/c++/4.1.3/i486-linux-gnu/"
ignoring nonexistent directory "/usr/include/c++/4.1.3/backward"
ignoring nonexistent directory "/usr/include/c++/4.3.3"
ignoring nonexistent directory "/usr/include/c++/4.3.3/x86_64-linux-gnu/32"
ignoring nonexistent directory "/usr/include/c++/4.3.3/backward"
ignoring nonexistent directory "/usr/include/c++/4.4.1"
ignoring nonexistent directory "/usr/include/c++/4.4.1/x86_64-linux-gnu/32"
ignoring nonexistent directory "/usr/include/c++/4.4.1/backward"
ignoring nonexistent directory "/usr/include/c++/4.1.2"
ignoring nonexistent directory "/usr/include/c++/4.1.2/i386-redhat-linux/"
ignoring nonexistent directory "/usr/include/c++/4.1.2/backward"
ignoring nonexistent directory "/usr/include/c++/4.3.0"
ignoring nonexistent directory "/usr/include/c++/4.3.0/i386-redhat-linux/"
ignoring nonexistent directory "/usr/include/c++/4.3.0/backward"
ignoring nonexistent directory "/usr/include/c++/4.3.2"
ignoring nonexistent directory "/usr/include/c++/4.3.2/i386-redhat-linux/"
ignoring nonexistent directory "/usr/include/c++/4.3.2/backward"
ignoring nonexistent directory "/usr/include/c++/4.3.2"
ignoring nonexistent directory "/usr/include/c++/4.3.2/x86_64-redhat-linux/32"
ignoring nonexistent directory "/usr/include/c++/4.3.2/backward"
ignoring nonexistent directory "/usr/include/c++/4.4.1"
ignoring nonexistent directory "/usr/include/c++/4.4.1/i586-redhat-linux/"
ignoring nonexistent directory "/usr/include/c++/4.4.1/backward"
ignoring nonexistent directory "/usr/include/c++/4.4.2"
ignoring nonexistent directory "/usr/include/c++/4.4.2/i686-redhat-linux/"
ignoring nonexistent directory "/usr/include/c++/4.4.2/backward"
ignoring nonexistent directory "/usr/include/c++/4.3"
ignoring nonexistent directory "/usr/include/c++/4.3/i586-suse-linux/"
ignoring nonexistent directory "/usr/include/c++/4.3/backward"
ignoring nonexistent directory "/usr/include/c++/4.3"
ignoring nonexistent directory "/usr/include/c++/4.3/x86_64-suse-linux/32"
ignoring nonexistent directory "/usr/include/c++/4.3/backward"
ignoring nonexistent directory "/usr/include/c++/4.4"
ignoring nonexistent directory "/usr/include/c++/4.4/i586-suse-linux/"
ignoring nonexistent directory "/usr/include/c++/4.4/backward"
ignoring nonexistent directory "/usr/include/c++/4.4"
ignoring nonexistent directory "/usr/include/c++/4.4/x86_64-suse-linux/"
ignoring nonexistent directory "/usr/include/c++/4.4/backward"
ignoring nonexistent directory "/usr/include/c++/4.3.1"
ignoring nonexistent directory "/usr/include/c++/4.3.1/i686-pc-linux-gnu/"
ignoring nonexistent directory "/usr/include/c++/4.3.1/backward"
ignoring nonexistent directory "/usr/include/c++/4.3.1"
ignoring nonexistent directory
"/usr/include/c++/4.3.1/x86_64-unknown-linux-gnu/"
ignoring nonexistent directory "/usr/include/c++/4.3.1/backward"
ignoring nonexistent directory
"/usr/lib/gcc/i686-pc-linux-gnu/4.3.4/include/g++-v4"
ignoring nonexistent directory
"/usr/lib/gcc/i686-pc-linux-gnu/4.3.4/include/g++-v4/i686-pc-linux-gnu/"
ignoring nonexistent directory
"/usr/lib/gcc/i686-pc-linux-gnu/4.3.4/include/g++-v4/backward"
ignoring nonexistent directory
"/usr/lib/gcc/i686-pc-linux-gnu/4.3.2/include/g++-v4"
ignoring nonexistent directory
"/usr/lib/gcc/i686-pc-linux-gnu/4.3.2/include/g++-v4/i686-pc-linux-gnu/"
ignoring nonexistent directory
"/usr/lib/gcc/i686-pc-linux-gnu/4.3.2/include/g++-v4/backward"
ignoring nonexistent directory
"/usr/lib/gcc/i686-pc-linux-gnu/4.1.2/include/g++-v4"
ignoring nonexistent directory
"/usr/lib/gcc/i686-pc-linux-gnu/4.1.2/include/g++-v4/i686-pc-linux-gnu/"
ignoring nonexistent directory
"/usr/lib/gcc/i686-pc-linux-gnu/4.1.2/include/g++-v4/backward"
ignoring nonexistent directory "/usr/include/c++/4.3"
ignoring nonexistent directory "/usr/include/c++/4.3/i486-pc-linux-gnu/"
ignoring nonexistent directory "/usr/include/c++/4.3/backward"
ignoring nonexistent directory "/usr/include/c++/4.3"
ignoring nonexistent directory "/usr/include/c++/4.3/i486-linux-gnu/"
ignoring nonexistent directory "/usr/include/c++/4.3/backward"
ignoring nonexistent directory
"/usr/lib/gcc/x86_64-pc-linux-gnu/4.1.2/include/g++-v4"
ignoring nonexistent directory
"/usr/lib/gcc/x86_64-pc-linux-gnu/4.1.2/include/g++-v4/i686-pc-linux-gnu/"
ignoring nonexistent directory
"/usr/lib/gcc/x86_64-pc-linux-gnu/4.1.2/include/g++-v4/backward"
ignoring nonexistent directory
"/usr/lib/gcc/x86_64-pc-linux-gnu/4.3.2/include/g++-v4"
ignoring nonexistent directory
"/usr/lib/gcc/x86_64-pc-linux-gnu/4.3.2/include/g++-v4/x86_64-pc-linux-gnu/"
ignoring nonexistent directory
"/usr/lib/gcc/x86_64-pc-linux-gnu/4.3.2/include/g++-v4/backward"
ignoring nonexistent directory
"/usr/lib/gcc/x86_64-pc-linux-gnu/4.4.3/include/g++-v4"
ignoring nonexistent directory
"/usr/lib/gcc/x86_64-pc-linux-gnu/4.4.3/include/g++-v4/x86_64-pc-linux-gnu/32"
ignoring nonexistent directory
"/usr/lib/gcc/x86_64-pc-linux-gnu/4.4.3/include/g++-v4/backward"
ignoring duplicate directory "/home/andrey/projects/Etoile/Build"
ignoring duplicate directory "/home/andrey/projects/Etoile/Build"
ignoring duplicate directory
"/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen"
ignoring duplicate directory "/home/andrey/projects/Etoile/Build"
ignoring duplicate directory
"/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen"
ignoring duplicate directory
"/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen"
ignoring duplicate directory "/usr/include/c++/4.4.3"
ignoring duplicate directory "/usr/include/c++/4.4.3/backward"
ignoring duplicate directory "/usr/include/c++/4.4.3"
ignoring duplicate directory "/usr/include/c++/4.4.3/backward"
ignoring duplicate directory "/usr/local/include"
  as it is a non-system directory that duplicates a system directory
#include "..." search starts here:
#include <...> search starts here:
 /usr/lib/gcc/i686-redhat-linux/4.4.3/include
 /home/andrey/projects/Etoile/Build
 /home/andrey/projects/Etoile/Languages/LanguageKit
 /home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen
 ./derived_src
 /usr/GNUstep/Local/Library/Headers
 /usr/GNUstep/System/Library/Headers
 /usr/local/lib/clang/1.5/include
 /usr/include/c++/4.4.3
 /usr/include/c++/4.4.3/backward
 /usr/include/c++/4.4.3/i686-redhat-linux
 /usr/local/include
 /usr/include
End of search list.
clang: Mangle.cpp:310: const clang::DeclContext* IgnoreLinkageSpecDecls(const
clang::DeclContext*): Assertion `cast<LinkageSpecDecl>(DC)->getLanguage() ==
LinkageSpecDecl::lang_cxx && "Unexpected linkage decl!"' failed.
0  clang     0x08fb04f2
1  clang     0x08fb03c3
2            0x00af1400 __kernel_sigreturn + 0
3  libc.so.6 0x0013c34a abort + 378
4  libc.so.6 0x00133bd8 __assert_fail + 248
5  clang     0x081b9d43
6  clang     0x081b9d99
7  clang     0x081bea14
8  clang     0x081be43e
9  clang     0x081baee3
10 clang     0x081bac28
11 clang     0x081ba032
12 clang     0x081b9b86
13 clang     0x081b9b18
14 clang     0x081bee7e
15 clang     0x0819e7c9
16 clang     0x0819e6f5
17 clang     0x081a021d
18 clang     0x081a3bad
19 clang     0x0819d3b6
20 clang     0x0805eaf1
21 clang     0x084474a6
22 clang     0x08448ed3
23 clang     0x082b600b
24 clang     0x086498ec
25 clang     0x082b45e2
26 clang     0x0807975c
27 clang     0x080793cd
28 clang     0x08064afd
29 clang     0x0804df69
30 clang     0x08052ded main + 232
31 libc.so.6 0x00126bb6 __libc_start_main + 230
32 clang     0x0804cb31
Stack dump:
0.    Program arguments: /usr/local/bin/clang -cc1 -triple i386-pc-linux-gnu -S
-disable-free -main-file-name LLVMCodeGen.mm -pic-level 2 -mdisable-fp-elim
-mconstructor-aliases -target-cpu pentium4 -v -g -resource-dir
/usr/local/lib/clang/1.5 -dependency-file
obj/LanguageKitCodeGen.obj/LLVMCodeGen.mm.d -MT
obj/LanguageKitCodeGen.obj/LLVMCodeGen.mm.o -MP -DGNU_RUNTIME_VERSION=1
-DGNU_RUNTIME_VERSION=1 -D_GNU_SOURCE -DGNU_RUNTIME_VERSION=1 -D_GNU_SOURCE
-DGNU_RUNTIME_VERSION=1 -DGNUSTEP -DGNUSTEP_BASE_LIBRARY=1 -DGNU_GUI_LIBRARY=1
-DGNU_RUNTIME=1 -DGNUSTEP_BASE_LIBRARY=1 -D_REENTRANT -DGSWARN -DGSDIAGNOSE
-D_DEBUG -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS
-I/usr/lib/gcc/i686-redhat-linux/4.4.3/include -I/usr/local/include
-I/home/andrey/projects/Etoile/Build
-I/home/andrey/projects/Etoile/Languages/LanguageKit
-I/home/andrey/projects/Etoile/Languages/LanguageKit/Headers
-I/home/andrey/projects/Etoile/Build
-I/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen
-I/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen/Headers
-I/home/andrey/projects/Etoile/Build
-I/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen
-I/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen/Headers
-I/home/andrey/projects/Etoile/Build
-I/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen
-I/home/andrey/projects/Etoile/Languages/LanguageKit/CodeGen/Headers
-I./derived_src -I. -I/home/andrey/GNUstep/Library/Headers
-I/usr/GNUstep/Local/Library/Headers -I/usr/GNUstep/System/Library/Headers -O2
-Wall -Wno-import -Woverloaded-virtual -w -fconstant-string-class
NSConstantString -ferror-limit 19 -fmessage-length 0 -fno-rtti -fgnu-runtime
-fdiagnostics-show-option -o /tmp/cc-JTlMvF.s -x objective-c++ LLVMCodeGen.mm 
1.    <eof> parser at end of file
2.    /usr/include/c++/4.4.3/bits/basic_string.h:1624:7: instantiating function
definition 'std::basic_string<char, std::char_traits<char>,
std::allocator<char> >::data'
3.    /usr/include/c++/4.4.3/bits/basic_string.h:1624:7: LLVM IR generation of
declaration 'std::basic_string<char, std::char_traits<char>,
std::allocator<char> >::data'
4.    /usr/include/c++/4.4.3/bits/basic_string.h:1624:7: Mangling declaration
'std::basic_string<char, std::char_traits<char>, std::allocator<char> >::data'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)
make[5]: *** [obj/LanguageKitCodeGen.obj/LLVMCodeGen.mm.o] Error 250
make[4]: *** [internal-framework-run-compile-submake] Error 2
make[3]: *** [LanguageKitCodeGen.all.framework.variables] Error 2
make[2]: *** [internal-all] Error 2
make[1]: *** [internal-all] Error 2
make: *** [internal-all] Error 2

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 18 09:19:43 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 18 Apr 2010 09:19:43 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6868] New: [cbe] cfrac test fails
Message-ID: <bug-6868-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6868

           Summary: [cbe] cfrac test fails
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: pawel.worach at gmail.com
                CC: llvmbugs at cs.uiuc.edu


$ lli --version
Low Level Virtual Machine (http://llvm.org/):
  llvm version 2.7
  Optimized build with assertions.
  Built Apr 17 2010 (00:16:44).
  Host: x86_64-unknown-freebsd8.0
  Host CPU: penryn

  Registered Targets:
    x86    - 32-bit X86: Pentium-Pro and above
    x86-64 - 64-bit X86: EM64T and AMD64

$ lli cfrac.llvm.bc 41757646344123832613190542166099121
41757646344123832613190542166099121 = 49384934934843479 * 845554345656569399
$ llc -march=c -o cfrac.c cfrac.llvm.bc
$ gcc cfrac.c -o cfrac -lm
<ton of warnings>
$ ./cfrac 41757646344123832613190542166099121
Illegal instruction (core dumped)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 18 13:01:14 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 18 Apr 2010 13:01:14 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6868] [cbe] cfrac test fails
In-Reply-To: <bug-6868-206@http.llvm.org/bugs/>
References: <bug-6868-206@http.llvm.org/bugs/>
Message-ID: <20100418180114.979382A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6868

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-18 13:01:14 CDT ---
The CBE is pretty hopelessly broken.  This is a likely dupe of one of 2399,
605, 1658, etc.

*** This bug has been marked as a duplicate of bug 605 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 18 14:07:35 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 18 Apr 2010 14:07:35 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6766] clang crash with a complex C99 structure
	initialisation
In-Reply-To: <bug-6766-206@http.llvm.org/bugs/>
References: <bug-6766-206@http.llvm.org/bugs/>
Message-ID: <20100418190735.277002A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6766

Nuno Lopes <nunoplopes at sapo.pt> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Nuno Lopes <nunoplopes at sapo.pt> 2010-04-18 14:07:34 CDT ---
Sorry for the regression. I've addressed the problem and recommited the patch
in r101726.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 18 16:02:53 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 18 Apr 2010 16:02:53 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6769] IRgen crashes emitting code for static member
 function of local class within a static member function
In-Reply-To: <bug-6769-206@http.llvm.org/bugs/>
References: <bug-6769-206@http.llvm.org/bugs/>
Message-ID: <20100418210253.27ECA2A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6769

Fariborz Jahanian <fjahanian at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Fariborz Jahanian <fjahanian at apple.com> 2010-04-18 16:02:52 CDT ---
Fixed in http://llvm.org/viewvc/llvm-project?view=rev&revision=101756

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 18 22:15:58 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 18 Apr 2010 22:15:58 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6871] New: Move lib/CodeGen/SelectionDAGBuilder.h
 to llvm/CodeGen/ directory
Message-ID: <bug-6871-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6871

           Summary: Move lib/CodeGen/SelectionDAGBuilder.h to
                    llvm/CodeGen/ directory
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: etherzhhb at gmail.com
                CC: llvmbugs at cs.uiuc.edu


It would be great to move lib/CodeGen/SelectionDAGBuilder.h to
include/llvm/CodeGen/ directory so others llvm tools/projects can use the
SelectionDAGBuilder to build custom DAG.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 05:33:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 05:33:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6872] New: BASH 4.1 fails to compile using self
 hosted clang (r101748)
Message-ID: <bug-6872-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6872

           Summary: BASH 4.1 fails to compile using self hosted clang
                    (r101748)
           Product: clang
           Version: trunk
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: diegoiast at gmail.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


As a test for clang I downloaded bash 4.1, run configure and tried to compile.
The configure phase worked, but the compilation did not. Here is the command
that failed. Host's gcc seems to work.

The operating system is Fedora 12, 32bit. clang is a two phase self hosted
compiler based on subversion r101748.


[elcuco at pinky ~/rpmbuild/BUILD/bash-4.1/build/lib/intl]
/home/elcuco/src/llvm/bootstrap-stage-2/bin/clang -v -c
-DLOCALEDIR=\"/usr/local/share/locale\"
-DLOCALE_ALIAS_PATH=\"/usr/local/share/locale\" -DLIBDIR=\"/usr/local/libdata\"
-DIN_LIBINTL -DENABLE_RELOCATABLE=1 -DIN_LIBRARY
-DINSTALLDIR=\"/usr/local/lib\" -DNO_XMALLOC
-Dset_relocation_prefix=libintl_set_relocation_prefix
-Drelocate=libintl_relocate -DDEPENDS_ON_LIBICONV=1 -DHAVE_CONFIG_H -DSHELL -I.
-I../../../lib/intl -I/home/elcuco/rpmbuild/BUILD/bash-4.1/build -I../../..  -g
-O2  ../../../lib/intl/bindtextdom.c 
clang version 1.5 (trunk 101748)
Target: i386-pc-linux-gnu
Thread model: posix
 "/home/elcuco/src/llvm/bootstrap-stage-2/bin/clang" -cc1 -triple
i386-pc-linux-gnu -S -disable-free -main-file-name bindtextdom.c
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases -target-cpu
pentium4 -v -g -resource-dir
/home/elcuco/src/llvm/bootstrap-stage-2/lib/clang/1.5
-DLOCALEDIR="/usr/local/share/locale"
-DLOCALE_ALIAS_PATH="/usr/local/share/locale" -DLIBDIR="/usr/local/libdata"
-DIN_LIBINTL -DENABLE_RELOCATABLE=1 -DIN_LIBRARY -DINSTALLDIR="/usr/local/lib"
-DNO_XMALLOC -Dset_relocation_prefix=libintl_set_relocation_prefix
-Drelocate=libintl_relocate -DDEPENDS_ON_LIBICONV=1 -DHAVE_CONFIG_H -DSHELL -I.
-I../../../lib/intl -I/home/elcuco/rpmbuild/BUILD/bash-4.1/build -I../../.. -O2
-ferror-limit 19 -fmessage-length 143 -fgnu-runtime -fdiagnostics-show-option
-fcolor-diagnostics -o /tmp/cc-ktaKVv.s -x c ../../../lib/intl/bindtextdom.c
clang -cc1 version 1.5 based upon llvm 2.7svn hosted on i386-pc-linux-gnu
#include "..." search starts here:
#include <...> search starts here:
 .
 ../../../lib/intl
 /home/elcuco/rpmbuild/BUILD/bash-4.1/build
 ../../..
 /home/elcuco/src/llvm/bootstrap-stage-2/lib/clang/1.5/include
 /usr/local/include
 /usr/include
End of search list.
Stack dump:
0.      Program arguments: /home/elcuco/src/llvm/bootstrap-stage-2/bin/clang
-cc1 -triple i386-pc-linux-gnu -S -disable-free -main-file-name bindtextdom.c
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases -target-cpu
pentium4 -v -g -resource-dir
/home/elcuco/src/llvm/bootstrap-stage-2/lib/clang/1.5
-DLOCALEDIR="/usr/local/share/locale"
-DLOCALE_ALIAS_PATH="/usr/local/share/locale" -DLIBDIR="/usr/local/libdata"
-DIN_LIBINTL -DENABLE_RELOCATABLE=1 -DIN_LIBRARY -DINSTALLDIR="/usr/local/lib"
-DNO_XMALLOC -Dset_relocation_prefix=libintl_set_relocation_prefix
-Drelocate=libintl_relocate -DDEPENDS_ON_LIBICONV=1 -DHAVE_CONFIG_H -DSHELL -I.
-I../../../lib/intl -I/home/elcuco/rpmbuild/BUILD/bash-4.1/build -I../../.. -O2
-ferror-limit 19 -fmessage-length 143 -fgnu-runtime -fdiagnostics-show-option
-fcolor-diagnostics -o /tmp/cc-ktaKVv.s -x c ../../../lib/intl/bindtextdom.c 
1.      <eof> parser at end of file
2.      Per-module optimization passes
3.      Running pass 'CallGraph Pass Manager' on module
'../../../lib/intl/bindtextdom.c'.
4.      Running pass 'Loop Pass Manager' on function '@set_binding_values'
5.      Running pass 'Loop Invariant Code Motion' on basic block '%while.cond'
clang: error: compiler command failed due to signal 11 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 10:31:06 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 10:31:06 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6873] New: LLVM Tutorial program
Message-ID: <bug-6873-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6873

           Summary: LLVM Tutorial program
           Product: new-bugs
           Version: 2.6
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: release blocker
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: andrew.james.bate at googlemail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4698)
 --> (http://llvm.org/bugs/attachment.cgi?id=4698)
copy of shell session showing failure and fix

The tutorial program built with the compile flags shown crashes when I try to
call putchard from Kalidiscope.

After some digging around I found it could be fixed by compiling with 
'-rdynamic'

This puts all symbols where dl_sym() and hence LLVM can find them, probably
should use LLVM APIs to get just putchard searchable.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 11:29:01 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 11:29:01 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6874] New: clang cannot compile this decl yet
 (namespace af = ...)
Message-ID: <bug-6874-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6874

           Summary: clang cannot compile this decl yet (namespace af =
                    ...)
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ich at az2000.de
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


/Users/az/Programmierung/openlierox/src/gusanos/part_type.cpp:320:2: error: 
      cannot compile this decl yet
        namespace af = OmfgScript::ActionParamFlags;
        ^
1 error generated.


Haven't found any other report reporting this so I'm doing the report hereby.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 11:42:27 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 11:42:27 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6875] New: llc -print-after-all crashes
Message-ID: <bug-6875-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6875

           Summary: llc -print-after-all crashes
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


pes ~$ cat fabs.ll 
; ModuleID = '/home/rdivacky/fabsf.c'
target datalayout =
"e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64"
target triple = "x86_64-unknown-freebsd8.0"

define float @foo(float %a) nounwind {
entry:
  %retval = alloca float, align 4                 ; <float*> [#uses=2]
  %a.addr = alloca float, align 4                 ; <float*> [#uses=2]
  %b = alloca float, align 4                      ; <float*> [#uses=2]
  store float %a, float* %a.addr
  %tmp = load float* %a.addr                      ; <float> [#uses=1]
  %call = call float @fabsf(float %tmp)           ; <float> [#uses=1]
  store float %call, float* %b
  %tmp1 = load float* %b                          ; <float> [#uses=1]
  store float %tmp1, float* %retval
  %0 = load float* %retval                        ; <float> [#uses=1]
  ret float %0
}

declare float @fabsf(float) nounwind readnone

pes ~$ llc -O2 fabs.ll && echo success
success

pes ~$ llc -O2 fabs.ll -print-after-all && echo success
Assertion failed: (!PMS.empty() && "Unable to find appropriate Pass Manager"),
function assignPassManager, file PassManager.cpp, line 1695.
Stack dump:
0.    Program arguments: llc -O2 fabs.ll -print-after-all 
Abort (core dumped)

the attached patch fixes that to some extent but it crashes on something else

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 11:49:42 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 11:49:42 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6747] missed optimization for inline virtual methods
In-Reply-To: <bug-6747-206@http.llvm.org/bugs/>
References: <bug-6747-206@http.llvm.org/bugs/>
Message-ID: <20100419164942.0592F2A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6747

Rafael ?vila de Esp?ndola <rafael.espindola at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #5 from Rafael ?vila de Esp?ndola <rafael.espindola at gmail.com> 2010-04-19 11:49:41 CDT ---
Fixed in 101757.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 13:29:29 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 13:29:29 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6876] New: Unbounded recursion in MachineCSE
Message-ID: <bug-6876-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6876

           Summary: Unbounded recursion in MachineCSE
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicolas.geoffray at gmail.com
                CC: llvmbugs at cs.uiuc.edu


On very large functions, MachineCSE stack overflows because there are too many
recursive calls to processBlock (line 362 of lib/CodeGen/MachineCSE.cpp).

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 15:39:57 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 15:39:57 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6877] New: Valgrind errors with Clang C++
Message-ID: <bug-6877-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6877

           Summary: Valgrind errors with Clang C++
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com,
                    fjahanian at apple.com


Created an attachment (id=4700)
 --> (http://llvm.org/bugs/attachment.cgi?id=4700)
Failing test

When running Clang under valgrind for the attached file, we're getting a lot of
"uninitialized value" errors out of valgrind. The example itself comes from the
test in PR6794, which asserts consistently for the reporter but not for me. The
valgrind errors are at least consistent and seem related.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 15:54:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 15:54:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6821] JIT DWARF EH data does not seem to emit LSDA
In-Reply-To: <bug-6821-206@http.llvm.org/bugs/>
References: <bug-6821-206@http.llvm.org/bugs/>
Message-ID: <20100419205440.9AE1F2A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6821

Bill Wendling <wendling at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #4 from Bill Wendling <wendling at apple.com> 2010-04-19 15:54:40 CDT ---
Since there is no testcase and I've recently fixed something in this area, I'm
going to declare this closed. Please refile if you have a bug still with
testcase.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 16:31:43 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 16:31:43 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6495] extern const file-level variable with
 initializer raises warning
In-Reply-To: <bug-6495-206@http.llvm.org/bugs/>
References: <bug-6495-206@http.llvm.org/bugs/>
Message-ID: <20100419213143.983972A6C12F@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6495

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Douglas Gregor <dgregor at apple.com> 2010-04-19 16:31:43 CDT ---
Fixed in r101818.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 16:50:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 16:50:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 5958] common symbol in "extern C" not generated
In-Reply-To: <bug-5958-206@http.llvm.org/bugs/>
References: <bug-5958-206@http.llvm.org/bugs/>
Message-ID: <20100419215051.3B1852A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=5958

Fariborz Jahanian <fjahanian at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Fariborz Jahanian <fjahanian at apple.com> 2010-04-19 16:50:50 CDT ---
It is not reproducible any more:
extern "C" {
int EXTERN_C_A;
int EXTERN_C_B = 4;
};

int foo(void) {
    return EXTERN_C_A;
}

int bar(void) {
    return EXTERN_C_B;
}

% clang -###
Apple clang version 1.5 (trunk 101769)
Target: x86_64-apple-darwin10
Thread model: posix
% clang -c test.cpp
[% nm -nm test.o | grep EXTERN
0000000000000034 (__DATA,__data) external _EXTERN_C_B
00000000000000b0 (__DATA,__common) external _EXTERN_C_A

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 17:11:33 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 17:11:33 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6862] ice: Type is not SCEVable
In-Reply-To: <bug-6862-206@http.llvm.org/bugs/>
References: <bug-6862-206@http.llvm.org/bugs/>
Message-ID: <20100419221133.55E492A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6862

Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Dan Gohman <gohman at apple.com> 2010-04-19 17:11:32 CDT ---
Fixed in r101819.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 19:33:54 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 19:33:54 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6878] New: -cc1 should default to -fexceptions
Message-ID: <bug-6878-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6878

           Summary: -cc1 should default to -fexceptions
           Product: clang
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Driver
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rjmccall at apple.com
                CC: llvmbugs at cs.uiuc.edu


Currently clang -cc1 defaults to -fno-exceptions.  -cc1 should really default
to standards compliance.

I expect the main overhead in fixing this will be changing the run lines on a
bunch of codegen tests.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 20:12:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 20:12:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6866] MSVC9 hangs on compilation of
	X86Disassembler.cpp
In-Reply-To: <bug-6866-206@http.llvm.org/bugs/>
References: <bug-6866-206@http.llvm.org/bugs/>
Message-ID: <20100420011203.363652A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6866

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Chris Lattner <clattner at apple.com> 2010-04-19 20:12:02 CDT ---
Works for me, committed in r101854, thanks!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 19 22:03:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 19 Apr 2010 22:03:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6879] New: intermittent ice: Why are we unswitching
 on a constant
Message-ID: <bug-6879-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6879

           Summary: intermittent ice: Why are we unswitching on a constant
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu, chenyang at cs.utah.edu


Created an attachment (id=4701)
 --> (http://llvm.org/bugs/attachment.cgi?id=4701)
failure inducing C

This one is no fun!  The bug only happens about 50% of the time, something
nondeterministic has crept into the execution.  Valgrind seems to think it is a
use-after-free.

regehr at john-home:~/volatile/bugs/tmp314$ clang -O0 -S -emit-llvm small.c -o
small.bc -w
regehr at john-home:~/volatile/bugs/tmp314$ opt -O3 small.bcWARNING: You're
attempting to print out a bitcode file.
This is inadvisable as it may cause display problems. If
you REALLY want to taste LLVM bitcode first-hand, you
can force output with the `-f' option.

opt: LoopUnswitch.cpp:850:
void<unnamed>::LoopUnswitch::RewriteLoopBodyWithConditionConstant(llvm::Loop*,
llvm::Value*, llvm::Constant*, bool): Assertion `!isa<Constant>(LIC) && "Why
are we unswitching on a constant?"' failed.
0  opt 0x084d9af8
Stack dump:
0.    Program arguments: opt -O3 small.bc 
1.    Running pass 'CallGraph Pass Manager' on module 'small.bc'.
2.    Running pass 'Loop Pass Manager' on function '@func_39'
3.    Running pass 'Unswitch loops' on basic block '%for.cond.i'
Aborted
regehr at john-home:~/volatile/bugs/tmp314$ clang -v
clang version 1.5 (trunk 101819)
Target: i386-pc-linux-gnu
Thread model: posix

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 00:09:39 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 00:09:39 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6879] intermittent ice: Why are we unswitching on a
	constant
In-Reply-To: <bug-6879-206@http.llvm.org/bugs/>
References: <bug-6879-206@http.llvm.org/bugs/>
Message-ID: <20100420050939.09D552A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6879

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #4 from Chris Lattner <clattner at apple.com> 2010-04-20 00:09:38 CDT ---
Fixed here, thanks!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20100419/100135.html

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 00:44:58 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 00:44:58 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6525] noalias is incorrect on sret arguments
In-Reply-To: <bug-6525-206@http.llvm.org/bugs/>
References: <bug-6525-206@http.llvm.org/bugs/>
Message-ID: <20100420054458.C2CD82A6C12E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6525

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #6 from Chris Lattner <clattner at apple.com> 2010-04-20 00:44:57 CDT ---
Clang fixed in r101877.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 01:00:47 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 01:00:47 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6469] llvm-as asserts on invalid IR involving
 trailing comma with a misleading message
In-Reply-To: <bug-6469-206@http.llvm.org/bugs/>
References: <bug-6469-206@http.llvm.org/bugs/>
Message-ID: <20100420060047.A47C82A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6469

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-20 01:00:47 CDT ---
I fixed this awhile ago, now both examples produce "error: expected metadata
after comma".  Thanks.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 01:21:02 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 01:21:02 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6340] Bug with empty C99 structure initialization
In-Reply-To: <bug-6340-206@http.llvm.org/bugs/>
References: <bug-6340-206@http.llvm.org/bugs/>
Message-ID: <20100420062102.1CBA02A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6340

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-20 01:21:01 CDT ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20100419/100139.html

This was a nasty regression, thanks for pointing it out!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 01:37:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 01:37:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6873] LLVM Tutorial program
In-Reply-To: <bug-6873-206@http.llvm.org/bugs/>
References: <bug-6873-206@http.llvm.org/bugs/>
Message-ID: <20100420063740.9FBD02A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6873

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-20 01:37:40 CDT ---
The tutorial already says:

"If you are compiling this on Linux, make sure to add the "-rdynamic" option as
well. This makes sure that the external functions are resolved properly at
runtime."

http://llvm.org/docs/tutorial/LangImpl4.html

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 04:34:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 04:34:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6880] New: Segfault while Releasing pass 'Basic
 CallGraph Construction' (regression)
Message-ID: <bug-6880-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6880

           Summary: Segfault while Releasing pass 'Basic CallGraph
                    Construction' (regression)
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: quickslyver at free.fr
                CC: llvmbugs at cs.uiuc.edu


$cat test.c
int main();
void _exit(int);

void foo()
{
}

void _crt_init() 
{
  _exit(main());
}

$clang test.c -o test.bc -c -emit-llvm && llvm-ld test.bc
-internalize-public-api-list=foo,_crt_init
0  llvm-ld   0x08690c5d
1  llvm-ld   0x086911c4
2            0xb7fdf400 __kernel_sigreturn + 0
3  llvm-ld   0x084bb0ec
4  llvm-ld   0x084b9f88
5  llvm-ld   0x0860708f llvm::PMDataManager::freePass(llvm::Pass*,
llvm::StringRef, llvm::PassDebuggingString) + 127
6  llvm-ld   0x086072e8 llvm::PMDataManager::removeDeadPasses(llvm::Pass*,
llvm::StringRef, llvm::PassDebuggingString) + 316
7  llvm-ld   0x08607ccb llvm::FPPassManager::runOnFunction(llvm::Function&) +
591
8  llvm-ld   0x08607d73 llvm::FPPassManager::runOnModule(llvm::Module&) + 109
9  llvm-ld   0x08607892 llvm::MPPassManager::runOnModule(llvm::Module&) + 398
10 llvm-ld   0x08609088 llvm::PassManagerImpl::run(llvm::Module&) + 124
11 llvm-ld   0x086090ed llvm::PassManager::run(llvm::Module&) + 39
12 llvm-ld   0x08323f66 llvm::Optimize(llvm::Module*) + 735
13 llvm-ld   0x0832b061 main + 1881
14 libc.so.6 0xb7d33455 __libc_start_main + 229
15 llvm-ld   0x08323361
Stack dump:
0.    Program arguments: llvm-ld test.bc
-internalize-public-api-list=foo,_crt_init 
1.    Running pass 'Function Pass Manager' on module 'a.out'.
2.    Releasing pass 'Basic CallGraph Construction'
Segmentation fault

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 07:17:35 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 07:17:35 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6880] Segfault while Releasing pass 'Basic
 CallGraph Construction' (regression)
In-Reply-To: <bug-6880-206@http.llvm.org/bugs/>
References: <bug-6880-206@http.llvm.org/bugs/>
Message-ID: <20100420121735.AEBC12A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6880

Benjamin Kramer <benny.kra at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |benny.kra at gmail.com
         Resolution|                            |FIXED

--- Comment #1 from Benjamin Kramer <benny.kra at gmail.com> 2010-04-20 07:17:35 CDT ---
Fixed in r101897.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 10:36:37 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 10:36:37 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6881] New: variable incorrectly optimized out
Message-ID: <bug-6881-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6881

           Summary: variable incorrectly optimized out
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


pes ~/clangbsd$ cat weak.c
extern int _DYNAMIC;
#pragma weak _DYNAMIC

void foo() {
   if (&_DYNAMIC != 0)
      printf("foo bar\n");
}
pes ~/clangbsd$ clang -c weak.c && nm weak.o
0000000000000000 T foo
                 U printf

pes ~/clangbsd$ gcc -c weak.c && nm weak.o
                 w _DYNAMIC
0000000000000000 T foo
                 U puts

this happens because lib/AST/ExprConstant.cpp:1180 logic is flawed. It does not
work for weak variable which are 0. thus the if() check is optimized away. We
need to check if the variable is weakref-ed before deciding it cannot be 0.

this proof of concept (disabling the opt) makes the test work:

Index: lib/AST/ExprConstant.cpp
===================================================================
--- lib/AST/ExprConstant.cpp    (revision 101893)
+++ lib/AST/ExprConstant.cpp    (working copy)
@@ -1178,7 +1178,7 @@
       // Reject any bases from the normal codepath; we special-case
comparisons
       // to null.
       if (LHSValue.getLValueBase()) {
-        if (!E->isEqualityOp())
+//        if (!E->isEqualityOp())
           return false;
         if (RHSValue.getLValueBase() || !RHSValue.getLValueOffset().isZero())
           return false;
@@ -1187,7 +1187,7 @@
           return false;
         return Success(bres ^ (E->getOpcode() == BinaryOperator::EQ), E);
       } else if (RHSValue.getLValueBase()) {
-        if (!E->isEqualityOp())
+//        if (!E->isEqualityOp())
           return false;
         if (LHSValue.getLValueBase() || !LHSValue.getLValueOffset().isZero())
           return false;

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 12:38:48 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 12:38:48 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6882] New: clang crashes on spirit.cpp
Message-ID: <bug-6882-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6882

           Summary: clang crashes on spirit.cpp
           Product: clang
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


clang crashes on SingleSource/Benchmarks/Misc-C++-EH/spirit.cpp.

Here's a backtrace from gdb:

Program received signal SIGSEGV, Segmentation fault.
0x000000000040a546 in clang::IdentifierInfo::getLength (this=0x0) at
/llvm/tools/clang/lib/Frontend/../../include/clang/Basic/IdentifierTable.h:98
98      if (Entry) return Entry->getKeyLength();
(gdb) where
#0  0x000000000040a546 in clang::IdentifierInfo::getLength (this=0x0) at
/llvm/tools/clang/lib/Frontend/../../include/clang/Basic/IdentifierTable.h:98
#1  0x000000000044df6b in clang::IdentifierInfo::getName (this=0x0) at
/llvm/tools/clang/lib/Frontend/../../include/clang/Basic/IdentifierTable.h:110
#2  0x000000000057cfbe in mangleSourceName (this=0x7fffffff9d90, II=0x0) at
Mangle.cpp:638
#3  0x000000000057efa9 in mangleType (this=0x7fffffff9d90, T=0x32d8b90) at
Mangle.cpp:1171
#4  0x000000000057e38f in mangleType (this=0x7fffffff9d90, T=...) at
/llvm/tools/clang/lib/CodeGen/../../include/clang/AST/TypeNodes.def:95
#5  0x000000000057ec91 in mangleType (this=0x7fffffff9d90, T=0x32d8e70) at
Mangle.cpp:1118
#6  0x000000000057e05b in mangleType (this=0x7fffffff9d90, T=...) at
/llvm/tools/clang/lib/CodeGen/../../include/clang/AST/TypeNodes.def:67
#7  0x000000000057e7c7 in mangleBareFunctionType (this=0x7fffffff9d90,
T=0x32d8ed0, MangleReturnType=true) at Mangle.cpp:1026
#8  0x000000000057c0c1 in mangleFunctionEncoding (this=0x7fffffff9d90,
FD=0x32d8f60) at Mangle.cpp:298
#9  0x000000000057bf5c in mangle (this=0x7fffffff9d90, D=0x4cb5d50, Prefix=...)
at Mangle.cpp:252
#10 0x00000000005813f4 in clang::CodeGen::MangleContext::mangleName
(this=0x2069480, D=0x4cb5d50, Res=...) at Mangle.cpp:1897
#11 0x000000000056096d in clang::CodeGen::CodeGenModule::getMangledName
(this=0x2069210, Buffer=..., ND=0x4cb5d50) at CodeGenModule.cpp:193
#12 0x000000000056089a in clang::CodeGen::CodeGenModule::getMangledName
(this=0x2069210, Buffer=..., GD=...) at CodeGenModule.cpp:176
#13 0x00000000005626ba in clang::CodeGen::CodeGenModule::EmitGlobal
(this=0x2069210, GD=...) at CodeGenModule.cpp:711
#14 0x000000000056648d in clang::CodeGen::CodeGenModule::EmitTopLevelDecl
(this=0x2069210, D=0x4cb5d50) at CodeGenModule.cpp:1767

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 13:27:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 13:27:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6883] New: diamond in anonymous namespace fails to
	link
Message-ID: <bug-6883-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6883

           Summary: diamond in anonymous namespace fails to link
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


This program fails to link:

namespace {
class B1 { };
class VD1 : virtual public B1 { };
class VD2 : virtual public B1 { };
class VD3 : public VD1, public VD2 { };
}

int main(void) {
  VD3 vd3;
}

$ g++ a.cc
$ llvm/Debug/bin/clang++ a.cc
/usr/bin/ld: /tmp/cc-5CadNh.o: in function (anonymous
namespace)::VD3::VD3():a.cc(.text+0x3b): error: undefined reference to
'_ZTTN12_GLOBAL__N_13VD3E1'
/usr/bin/ld: /tmp/cc-5CadNh.o: in function (anonymous
namespace)::VD3::VD3():a.cc(.text+0x73): error: undefined reference to
'_ZTTN12_GLOBAL__N_13VD3E2'
collect2: ld returned 1 exit status
clang: error: linker command failed with exit code 1 (use -v to see invocation)

Excluding the usual stuff, here's what 'nm' on a.out from g++ shows:

00000000004005b8 t _ZN33_GLOBAL__N_a.cc_00000000_6EBD49AD2B1C2Ev
00000000004005c2 t _ZN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD1C2Ev
00000000004005e2 t _ZN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD2C2Ev
0000000000400602 t _ZN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD3C1Ev
00000000004007d0 r _ZTCN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD3E0_NS_3VD1E
00000000004007f0 r _ZTCN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD3E8_NS_3VD2E
0000000000400950 V _ZTIN33_GLOBAL__N_a.cc_00000000_6EBD49AD2B1E
0000000000400860 V _ZTIN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD1E
00000000004008a0 V _ZTIN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD2E
0000000000400820 V _ZTIN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD3E
00000000004009a0 V _ZTSN33_GLOBAL__N_a.cc_00000000_6EBD49AD2B1E
0000000000400920 V _ZTSN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD1E
0000000000400960 V _ZTSN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD2E
00000000004008e0 V _ZTSN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD3E
0000000000400780 r _ZTTN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD3E
0000000000401ea0 B _ZTVN10__cxxabiv117__class_type_infoE
0000000000401e40 B _ZTVN10__cxxabiv121__vmi_class_type_infoE
00000000004007a0 r _ZTVN33_GLOBAL__N_a.cc_00000000_6EBD49AD3VD3E

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 15:43:52 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 15:43:52 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6881] variable incorrectly optimized out
In-Reply-To: <bug-6881-206@http.llvm.org/bugs/>
References: <bug-6881-206@http.llvm.org/bugs/>
Message-ID: <20100420204352.AB1702A6C12F@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6881

Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #2 from Eli Friedman <sharparrow1 at yahoo.com> 2010-04-20 15:43:52 CDT ---


*** This bug has been marked as a duplicate of bug 4386 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 17:03:42 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 17:03:42 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6644] [C++ CodeGen] Crash with polymorphic class
 inside extern "C" block
In-Reply-To: <bug-6644-206@http.llvm.org/bugs/>
References: <bug-6644-206@http.llvm.org/bugs/>
Message-ID: <20100420220342.2EF352A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6644

Fariborz Jahanian <fjahanian at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Fariborz Jahanian <fjahanian at apple.com> 2010-04-20 17:03:41 CDT ---
Fixed in TOT: http://llvm.org/viewvc/llvm-project?view=rev&revision=101948

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 17:26:42 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 17:26:42 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6136] Abort trap in Kaleidoscope/toy example,
	"at the end"
In-Reply-To: <bug-6136-206@http.llvm.org/bugs/>
References: <bug-6136-206@http.llvm.org/bugs/>
Message-ID: <20100420222642.DC9672A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6136

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-20 17:26:42 CDT ---
The toy example doesn't work with the interpreter.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 18:19:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 18:19:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 1135] Change "+" to "|" may cause problem in loop
 dependence analysis
In-Reply-To: <bug-1135-206@http.llvm.org/bugs/>
References: <bug-1135-206@http.llvm.org/bugs/>
Message-ID: <20100420231903.8183B2A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=1135

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #8 from Chris Lattner <clattner at apple.com> 2010-04-20 18:19:02 CDT ---
Fixed in r101958

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 18:21:21 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 18:21:21 -0500 (CDT)
Subject: [LLVMbugs] [Bug 4648] JIT depends on AsmPrinter due to dwarfwriter
In-Reply-To: <bug-4648-206@http.llvm.org/bugs/>
References: <bug-4648-206@http.llvm.org/bugs/>
Message-ID: <20100420232121.7F5502A6C12F@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=4648

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX

--- Comment #12 from Chris Lattner <clattner at apple.com> 2010-04-20 18:21:20 CDT ---
We plan to make the jit use asmwriter, so this is wontfix.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 18:22:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 18:22:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 4823] test/Frontend/darwin-version.c fails
In-Reply-To: <bug-4823-206@http.llvm.org/bugs/>
References: <bug-4823-206@http.llvm.org/bugs/>
Message-ID: <20100420232240.44BA72A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=4823

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER

--- Comment #18 from Chris Lattner <clattner at apple.com> 2010-04-20 18:22:39 CDT ---
No idea what the status of this is, but if it is really a broken compiler, it's
done.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 18:24:52 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 18:24:52 -0500 (CDT)
Subject: [LLVMbugs] [Bug 5009] inliner missing opportunity with CPS style
	code
In-Reply-To: <bug-5009-206@http.llvm.org/bugs/>
References: <bug-5009-206@http.llvm.org/bugs/>
Message-ID: <20100420232452.3A5B82A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=5009

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-20 18:24:52 CDT ---


*** This bug has been marked as a duplicate of bug 6724 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 19:21:30 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 19:21:30 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6876] Unbounded recursion in MachineCSE
In-Reply-To: <bug-6876-206@http.llvm.org/bugs/>
References: <bug-6876-206@http.llvm.org/bugs/>
Message-ID: <20100421002130.A4B552A6C124@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6876

Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #4 from Evan Cheng <evan.cheng at apple.com> 2010-04-20 19:21:30 CDT ---
Fixed. r101964

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 19:43:52 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 19:43:52 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6842] inliner inlines recursive call,
	defeats loop-deletion
In-Reply-To: <bug-6842-206@http.llvm.org/bugs/>
References: <bug-6842-206@http.llvm.org/bugs/>
Message-ID: <20100421004352.1261F2A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6842

Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |

--- Comment #3 from Dan Gohman <gohman at apple.com> 2010-04-20 19:43:51 CDT ---
r101471 is reverted, because it caused slowdowns in testcases which depended on
recursive-function inlining.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 20:23:29 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 20:23:29 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6871] Move lib/CodeGen/SelectionDAGBuilder.h to
 llvm/CodeGen/ directory
In-Reply-To: <bug-6871-206@http.llvm.org/bugs/>
References: <bug-6871-206@http.llvm.org/bugs/>
Message-ID: <20100421012330.14A842A6C12C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6871

Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #6 from Dan Gohman <gohman at apple.com> 2010-04-20 20:23:29 CDT ---
This looks good.  Instead of making those functions be members of
TargetLowering though, I'd prefer to leave them as standalone functions, since
they don't require access to TargetLowering internals.  I've applied this
change in r101975.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 21:05:55 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 21:05:55 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6884] New: Clang fails to account for no-return
 destructors in warnings
Message-ID: <bug-6884-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6884

           Summary: Clang fails to account for no-return destructors in
                    warnings
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chandlerc at gmail.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Below, the noretutrn 'abort' function correctly prevents any warnings about
returning from a non-void function, but a noreturn destructor on a temporary
object fails to do this. These structures can be used to stream out a fatal
message prior to calling abort, among other cases.

% cat t.cc
void abort() __attribute__((noreturn));

struct abort_struct {
  abort_struct() {}
  ~abort_struct() __attribute__((noreturn)) { abort(); }
};

int f(int x) {
  switch (x) { default: abort(); }
}

int g(int x) {
  switch (x) { default: abort_struct(); }
}

% ./bin/clang -fsyntax-only -Wall t.cc
t.cc:14:1: warning: control reaches end of non-void function [-Wreturn-type]
}
^
1 warning generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 20 21:51:55 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 20 Apr 2010 21:51:55 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6885] New: Diagnostic for friend declaration inside
 of a namespace with same name
Message-ID: <bug-6885-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6885

           Summary: Diagnostic for friend declaration inside of a
                    namespace with same name
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: jason.haslam at gmail.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


I don't know if this is valid but gcc accepts it:

hack:~ jason$ cat test.cpp 
namespace A {
    class B {
        friend class A;
    };
}

hack:~ jason$ clang++ test.cpp 
test.cpp:3:16: error: redefinition of 'A' as different kind of symbol
                friend class A;
                             ^
test.cpp:1:11: note: previous definition is here
namespace A {
          ^
1 error generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 21 00:43:42 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 21 Apr 2010 00:43:42 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6781] Template diagnostics should suppress
 instantiation contexts when there are lots
In-Reply-To: <bug-6781-206@http.llvm.org/bugs/>
References: <bug-6781-206@http.llvm.org/bugs/>
Message-ID: <20100421054342.2B2172A6C12D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6781

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-21 00:43:41 CDT ---
This is now fixed in trunk.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 21 13:26:30 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 21 Apr 2010 13:26:30 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6886] New: Clang++ fails to resolve overloads when
 requesting the address of a static function
Message-ID: <bug-6886-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6886

           Summary: Clang++ fails to resolve overloads when requesting the
                    address of a static function
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: oliver at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4707)
 --> (http://llvm.org/bugs/attachment.cgi?id=4707)
Testcase

Clang doesn't correctly handle function overload resolution when taking the
address of a static function.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 21 16:14:43 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 21 Apr 2010 16:14:43 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6887] New: assertion failed: Default argument is
 not yet instantiated!
Message-ID: <bug-6887-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6887

           Summary: assertion failed: Default argument is not yet
                    instantiated!
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4708)
 --> (http://llvm.org/bugs/attachment.cgi?id=4708)
testcase

This is a crash on valid during irgen. The attached example fails:

$ g++ -c 000.cc
$ clang++ -fsyntax-only 000.cc
$ clang++ -c -flto 000.cc
clang: Decl.cpp:821: clang::Expr* clang::ParmVarDecl::getDefaultArg():
Assertion `!hasUninstantiatedDefaultArg() && "Default argument is not yet
instantiated!"' failed.
0  clang           0x0000000001d0c092
1  clang           0x0000000001d0c57e
2  libpthread.so.0 0x00007f36497177d0
3  libc.so.6       0x00007f3648a3b095 gsignal + 53
4  libc.so.6       0x00007f3648a3caf0 abort + 272
5  libc.so.6       0x00007f3648a342df __assert_fail + 239
6  clang           0x00000000012fdfab clang::ParmVarDecl::getDefaultArg() + 137
7  clang           0x0000000000f792d1 clang::ParmVarDecl::getDefaultArg() const
+ 21
8  clang           0x0000000001015db6
9  clang           0x0000000001018b0b
clang::CodeGen::CodeGenFunction::EmitClassMemberwiseCopy(llvm::Value*,
llvm::Value*, clang::CXXRecordDecl const*, clang::CXXRecordDecl const*,
clang::QualType) + 269
10 clang           0x0000000001018fc4
clang::CodeGen::CodeGenFunction::SynthesizeCXXCopyConstructor(llvm::SmallVector<std::pair<clang::VarDecl
const*, clang::QualType>, 16u> const&) + 1206
11 clang           0x0000000001019439
clang::CodeGen::CodeGenFunction::EmitConstructorBody(llvm::SmallVector<std::pair<clang::VarDecl
const*, clang::QualType>, 16u>&) + 483
12 clang           0x0000000001001293
clang::CodeGen::CodeGenFunction::GenerateCode(clang::CodeGen::GlobalDecl,
llvm::Function*) + 1311
13 clang           0x0000000000f47ab0
clang::CodeGen::CodeGenModule::EmitCXXConstructor(clang::CXXConstructorDecl
const*, clang::CXXCtorType) + 270
14 clang           0x0000000000f1c4aa
clang::CodeGen::CodeGenModule::EmitGlobalDefinition(clang::CodeGen::GlobalDecl)
+ 180
15 clang           0x0000000000f1ce56
clang::CodeGen::CodeGenModule::EmitDeferred() + 372
16 clang           0x0000000000f1d07b clang::CodeGen::CodeGenModule::Release()
+ 21
17 clang           0x0000000000f16927
18 clang           0x0000000000debb13
19 clang           0x0000000001020379 clang::ParseAST(clang::Preprocessor&,
clang::ASTConsumer*, clang::ASTContext&, bool, bool,
clang::CodeCompleteConsumer*) + 735
20 clang           0x0000000000e01c28 clang::ASTFrontendAction::ExecuteAction()
+ 256
21 clang           0x0000000000e01d25 clang::FrontendAction::Execute() + 239
22 clang           0x0000000000def724
clang::CompilerInstance::ExecuteAction(clang::FrontendAction&) + 706
23 clang           0x0000000000dd92cd cc1_main(char const**, char const**, char
const*, void*) + 1765
24 clang           0x0000000000dddf9f main + 251
25 libc.so.6       0x00007f3648a271c4 __libc_start_main + 244
26 clang           0x0000000000dd8179
Stack dump:
0.      Program arguments: /usr/local/google/home/nlewycky/llvm/Debug/bin/clang
-cc1 -triple x86_64-unknown-linux-gnu -emit-llvm-bc -disable-free
-main-file-name 000.cc -mrelocation-model static -mdisable-fp-elim
-mconstructor-aliases -munwind-tables -target-cpu x86-64 -resource-dir
/usr/local/google/home/nlewycky/llvm/Debug/lib/clang/1.5 -ferror-limit 19
-fmessage-length 80 -fexceptions -fgnu-runtime -fdiagnostics-show-option
-fcolor-diagnostics -o 000.o -x c++ 000.cc 
1.      <eof> parser at end of file
2.      Per-file LLVM IR generation
3.      000.cc:6:8: Generating code for declaration 'B::B'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 21 17:31:02 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 21 Apr 2010 17:31:02 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6888] New: Missing error string
Message-ID: <bug-6888-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6888

           Summary: Missing error string
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM assembly language parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


The following is obviously invalid syntax:

define void @p(<3 x i32>* %p, <3 x i32> %x) {
  store <3 x i32> %x, i32 1, i32 1>, <3 x i32>* %p
  ret void
}

llvm-as diagnoses the error, but doesn't provide an error string, so llvm-as
fails like this:

$ llvm-as t.ll
llvm-as:
$

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 21 21:35:34 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 21 Apr 2010 21:35:34 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6889] New: memory unsafety crash bug
Message-ID: <bug-6889-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6889

           Summary: memory unsafety crash bug
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu, chenyang at cs.utah.edu


Seen on x86 Linux.  Valgrind reports use of freed RAM.  I cannot seem to repro
using opt.

regehr at john-home:~/volatile/bugs/tmp315$ clang -c -w -O1 small.c
clang:
/home/regehr/z/compiler-build/llvm-r102031/include/llvm/Support/Casting.h:202:
typename llvm::cast_retty<To, From>::ret_type llvm::cast(const Y&) [with X =
llvm::CallInst, Y = llvm::Instruction*]: Assertion `isa<X>(Val) && "cast<Ty>()
argument of incompatible type!"' failed.
0  clang 0x08fcee58
Stack dump:
0.    Program arguments:
/home/regehr/z/compiler-install/llvm-gcc-r102031-install/bin/clang -cc1 -triple
i386-pc-linux-gnu -S -disable-free -main-file-name small.c -mrelocation-model
static -mdisable-fp-elim -mconstructor-aliases -target-cpu pentium4
-resource-dir
/home/regehr/z/compiler-install/llvm-gcc-r102031-install/lib/clang/1.5 -O1 -w
-ferror-limit 19 -ftemplate-backtrace-limit 10 -fmessage-length 89
-fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics -o /tmp/cc-bODr6C.s
-x c small.c 
1.    <eof> parser at end of file
2.    Per-module optimization passes
3.    Running pass 'CallGraph Pass Manager' on module 'small.c'.
4.    Running pass 'Loop Pass Manager' on function '@int326'
5.    Running pass 'Loop Invariant Code Motion' on basic block
'%for.cond5.preheader'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

regehr at john-home:~/volatile/bugs/tmp315$ clang -v

clang version 1.5 (trunk 102031)
Target: i386-pc-linux-gnu
Thread model: posix

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 21 21:52:12 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 21 Apr 2010 21:52:12 -0500 (CDT)
Subject: [LLVMbugs] [Bug 5567] clang crashes when bitfields in union have
 type of integer but the integer is aligned at char boundaries.
In-Reply-To: <bug-5567-206@http.llvm.org/bugs/>
References: <bug-5567-206@http.llvm.org/bugs/>
Message-ID: <20100422025212.62C84312836B@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=5567

Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #5 from Daniel Dunbar <daniel at zuster.org> 2010-04-21 21:52:11 CDT ---
Fixed here: http://llvm.org/viewvc/llvm-project?view=rev&revision=102045

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 21 21:52:41 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 21 Apr 2010 21:52:41 -0500 (CDT)
Subject: [LLVMbugs] [Bug 5591] Incorrect packed bitfields assignment
In-Reply-To: <bug-5591-206@http.llvm.org/bugs/>
References: <bug-5591-206@http.llvm.org/bugs/>
Message-ID: <20100422025241.33AFC312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=5591

Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #7 from Daniel Dunbar <daniel at zuster.org> 2010-04-21 21:52:40 CDT ---
Fixed here: http://llvm.org/viewvc/llvm-project?view=rev&revision=102045

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 03:22:52 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 03:22:52 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6891] New: clang 2.6 -O1 faster than clang 2.7
Message-ID: <bug-6891-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6891

           Summary: clang 2.6 -O1 faster than clang 2.7
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Timing results on the crc32 code in zlib, all code compiled as 64-bit, on an
Intel Core 2 Quad Q9550.
64/32 refers to the element size in the crc32 table.

     (64) O1    O2    O3    Os  (32) O1    O2    O3   Os
gcc-4.4   87.2  92.8  93.3  98.9     90.0  92.8  92.8 100.0
gcc-4.5   94.4  92.8  92.8  98.9     92.2  92.8  92.8  99.4
clang-2.6 82.2  88.9  88.9  88.9     85.6  87.2  86.7  86.7
clang-2.7 91.7  90.6  92.8  90.6     86.1  85.6  87.2  83.9
dragonegg 89.4  87.2  86.7  86.7     90.0  86.7  86.7  87.8

clang 2.6 -O1 / 64 is better than everything else, while LLVM usually prefers
32-bit elements, and -O2/-Os. (gcc tends to prefer -O1 and 64).

It would be interesting to investigate why clang 2.6 -O1 is faster, and fix
clang 2.7 (well trunk/2.8) to be faster too.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 03:34:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 03:34:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6892] New: X86 test failures on Cygwin 1.7
Message-ID: <bug-6892-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6892

           Summary: X86 test failures on Cygwin 1.7
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jay.foad at antixlabs.com
                CC: llvmbugs at cs.uiuc.edu


I'm running llvm 2.7 prerelease 2 on Cygwin 1.7.5(0.225/5/3). I get the
following failures from "make check":

FAIL: /home/foad/llvm/llvm-2.7/test/CodeGen/X86/2009-03-13-PHIElimBug.ll for
PR3
784
Failed with exit(1) at line 1
while running:  llc <
/home/foad/llvm/llvm-2.7/test/CodeGen/X86/2009-03-13-PHIEl
imBug.ll -march=x86 | FileCheck
/home/foad/llvm/llvm-2.7/test/CodeGen/X86/2009-0
3-13-PHIElimBug.ll
/home/foad/llvm/llvm-2.7/test/CodeGen/X86/2009-03-13-PHIElimBug.ll:31:15:
error:
 expected string not found in input
; CHECK-NEXT: Llabel1:
              ^
<stdin>:10:2: note: scanning from here
 call _g
 ^
<stdin>:10:2: note: possible intended match here
 call _g
 ^

FAIL: /home/foad/llvm/llvm-2.7/test/CodeGen/X86/2009-03-16-PHIElimInLPad.ll
Failed with exit(1) at line 1
while running:  llc <
/home/foad/llvm/llvm-2.7/test/CodeGen/X86/2009-03-16-PHIEl
imInLPad.ll -march=x86 -asm-verbose | FileCheck
/home/foad/llvm/llvm-2.7/test/Co
deGen/X86/2009-03-16-PHIElimInLPad.ll
/home/foad/llvm/llvm-2.7/test/CodeGen/X86/2009-03-16-PHIElimInLPad.ll:23:10:
err
or: expected string not found in input
; CHECK: lpad
         ^
<stdin>:1:2: note: scanning from here
 .def _phi; .scl 2; .type 32; .endef
 ^
<stdin>:1:7: note: possible intended match here
 .def _phi; .scl 2; .type 32; .endef
      ^
FAIL: /home/foad/llvm/llvm-2.7/test/CodeGen/X86/dyn-stackalloc.ll
Failed with exit(1) at line 2
while running:  llc <
/home/foad/llvm/llvm-2.7/test/CodeGen/X86/dyn-stackalloc.l
l -march=x86 | egrep {\$4294967280|-16}
child process exited abnormally
FAIL: /home/foad/llvm/llvm-2.7/test/CodeGen/X86/loop-strength-reduce4.ll
Failed with exit(1) at line 2
while running:  llc <
/home/foad/llvm/llvm-2.7/test/CodeGen/X86/loop-strength-re
duce4.ll -march=x86 -relocation-model=pic | FileCheck
/home/foad/llvm/llvm-2.7/t
est/CodeGen/X86/loop-strength-reduce4.ll -check-prefix=PIC
/home/foad/llvm/llvm-2.7/test/CodeGen/X86/loop-strength-reduce4.ll:16:8: error:
expected string not found in input
; PIC: cmpl $64
       ^
<stdin>:1:2: note: scanning from here
 .def _foo; .scl 2; .type 32; .endef
 ^
<stdin>:8:3: note: possible intended match here
 movl $-64, %ecx
  ^

FAIL: /home/foad/llvm/llvm-2.7/test/CodeGen/X86/twoaddr-coalesce.ll
Failed with exit(1) at line 1
while running:  llc <
/home/foad/llvm/llvm-2.7/test/CodeGen/X86/twoaddr-coalesce
.ll -march=x86 | /usr/bin/grep mov | count 4
Expected 4 lines, got 5.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 07:40:00 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 07:40:00 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6893] New: cbe creates illegal c from 'insertvalue'
Message-ID: <bug-6893-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6893

           Summary: cbe creates illegal c from 'insertvalue'
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: kalle.raiskila at nokia.com
                CC: llvmbugs at cs.uiuc.edu


Compiling the following code with the cbe produces bad C

  define void @tester(){
    %vitem = insertvalue [2 x float] zeroinitializer, float 3.0, 0 ; 
    ret void
  }


~/tmp$ gcc -c tmp.c
tmp.c: In function 'tester':
tmp.c:147: error: 'struct l_unnamed0' has no member named 'field0'


'struct l_unnamed0' is defined as:
  /* Structure contents */
  struct l_unnamed0 { float array[2]; };

gcc error comes from:
  struct l_unnamed0 llvm_cbe_vitem;
  llvm_cbe_vitem.field0 = 0x1.8p+1;



llvm version tested: r102057

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 09:06:41 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 09:06:41 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6495] extern const file-level variable with
 initializer raises warning
In-Reply-To: <bug-6495-206@http.llvm.org/bugs/>
References: <bug-6495-206@http.llvm.org/bugs/>
Message-ID: <20100422140641.6FE9D312836E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6495

Daniel Lowe <dlowe at bitmuse.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |

--- Comment #4 from Daniel Lowe <dlowe at bitmuse.com> 2010-04-22 09:06:40 CDT ---
Almost fixed.  The test case I gave was fixed, but this doesn't work:

$ echo 'extern const int my_ints[] = {42,43,44};' >example_a.cc
$ echo 'int main(int argc, char *argv[]) {extern int my_ints[]; return
my_ints[0];}' >example_b.cc
$ clang++ example_a.cc example_b.cc
example_a.cc:1:18: warning: 'extern' variable has an initializer
extern const int my_ints[] = {42,43,44};
                 ^
1 diagnostic generated.

Thanks for the great response time, though!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 09:36:34 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 09:36:34 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6495] extern const file-level variable with
 initializer raises warning
In-Reply-To: <bug-6495-206@http.llvm.org/bugs/>
References: <bug-6495-206@http.llvm.org/bugs/>
Message-ID: <20100422143634.5F7E8312836F@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6495

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #5 from Douglas Gregor <dgregor at apple.com> 2010-04-22 09:36:33 CDT ---
Well, that's embarrassing. Fixed for real in r102066.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 09:45:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 09:45:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6120] Call to noreturn pointer not marked noreturn
In-Reply-To: <bug-6120-206@http.llvm.org/bugs/>
References: <bug-6120-206@http.llvm.org/bugs/>
Message-ID: <20100422144540.5BBAC312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6120

Nuno Lopes <nunoplopes at sapo.pt> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |nunoplopes at sapo.pt
         Resolution|                            |FIXED

--- Comment #1 from Nuno Lopes <nunoplopes at sapo.pt> 2010-04-22 09:45:40 CDT ---
It works as of r102065.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 09:45:49 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 09:45:49 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6894] New: Debugging information with STL sort
 causes assembler error
Message-ID: <bug-6894-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6894

           Summary: Debugging information with STL sort causes assembler
                    error
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dlowe at bitmuse.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4718)
 --> (http://llvm.org/bugs/attachment.cgi?id=4718)
Assembly output of clang++

Here's the test case:

$ cat >example_a.cc <<EOF
#include <vector>
#include <algorithm>
void example() { std::vector<int> v; std::sort(v.begin(), v.end()); }
EOF
$ clang++ -c example_a.cc
# No output
$ clang++ -c -g example_a.cc
/tmp/cc-QxPGI1.s: Assembler messages:
/tmp/cc-QxPGI1.s:12855: Error: invalid sections for operation on `.Ltmp265' and
`.Ltext_begin' setting `.Lset571'
/tmp/cc-QxPGI1.s:12855: Error: invalid sections for operation on `.Ltmp272' and
`.Ltext_begin' setting `.Lset572'
/tmp/cc-QxPGI1.s:12855: Error: invalid sections for operation on `.Ltmp274' and
`.Ltext_begin' setting `.Lset573'
/tmp/cc-QxPGI1.s:12855: Error: invalid sections for operation on `.Ltmp276' and
`.Ltext_begin' setting `.Lset574'
clang: error: assembler command failed with exit code 1 (use -v to see
invocation)

I've attached the assembly output.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 11:27:08 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 11:27:08 -0500 (CDT)
Subject: [LLVMbugs] [Bug 4961] ARM GCC __sync_synchronize does not generate
	code
In-Reply-To: <bug-4961-206@http.llvm.org/bugs/>
References: <bug-4961-206@http.llvm.org/bugs/>
Message-ID: <20100422162708.C7576312836B@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=4961

Jim Grosbach <grosbach at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |grosbach at apple.com
         Resolution|                            |FIXED

--- Comment #2 from Jim Grosbach <grosbach at apple.com> 2010-04-22 11:27:08 CDT ---
__sync_synchronize() currently generates a 'dmb' instruction for ARMv7 and 'mov
   r0, #0 ; mcr p15, 0, r0, c7, c10, 5' for ARMv6.

In the future, we may want to adjust the details depending on environmental
specifics (single vs. multi core, differing OS requirements, etc). It's
probably worth opening separate PR's for the details of those as they become
clear.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 11:30:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 11:30:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 5151] Error while generating ARM code with llvm-gcc
In-Reply-To: <bug-5151-206@http.llvm.org/bugs/>
References: <bug-5151-206@http.llvm.org/bugs/>
Message-ID: <20100422163051.9A4D7312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=5151

Jim Grosbach <grosbach at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |grosbach at apple.com
         Resolution|                            |FIXED

--- Comment #2 from Jim Grosbach <grosbach at apple.com> 2010-04-22 11:30:51 CDT ---
This looks to be fixed. The attached test case compiles without error.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 12:01:27 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 12:01:27 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6162] thumb2 mode doesn't distinguish the v7-M
	variant
In-Reply-To: <bug-6162-206@http.llvm.org/bugs/>
References: <bug-6162-206@http.llvm.org/bugs/>
Message-ID: <20100422170127.78EA7312836E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6162

Jim Grosbach <grosbach at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #4 from Jim Grosbach <grosbach at apple.com> 2010-04-22 12:01:27 CDT ---
This is the same root issue as 5717.

*** This bug has been marked as a duplicate of bug 5717 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 13:44:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 13:44:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6886] Clang++ fails to resolve overloads when
 requesting the address of a static function
In-Reply-To: <bug-6886-206@http.llvm.org/bugs/>
References: <bug-6886-206@http.llvm.org/bugs/>
Message-ID: <20100422184431.53900312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6886

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from John McCall <rjmccall at apple.com> 2010-04-22 13:44:31 CDT ---
Fixed in r102107.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 14:53:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 14:53:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6846] Integer literal to vector cast reports error
In-Reply-To: <bug-6846-206@http.llvm.org/bugs/>
References: <bug-6846-206@http.llvm.org/bugs/>
Message-ID: <20100422195356.69006312836B@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6846

John Thompson <John.Thompson.JTSoftware at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from John Thompson <John.Thompson.JTSoftware at gmail.com> 2010-04-22 14:53:56 CDT ---
Fixed in r101863.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 14:56:29 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 14:56:29 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6895] New: Vector literal/casting confusion.
Message-ID: <bug-6895-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6895

           Summary: Vector literal/casting confusion.
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: John.Thompson.JTSoftware at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I've run into a problem with vector literal/casting confusion:

vector float v = (vector float)((vector unsigned int)(0x49800000));

vecinit1.cpp:1:33: error: cannot initialize a vector element of type 'float'
with an lvalue of type '__vector unsigned int'
vector float v = (vector float)((vector unsigned int)(0x49800000));

It appears to be treating the outer paren expression as a vector literal
instead of a cast.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 15:52:13 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 15:52:13 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6894] Debugging information with STL sort causes
 assembler error
In-Reply-To: <bug-6894-206@http.llvm.org/bugs/>
References: <bug-6894-206@http.llvm.org/bugs/>
Message-ID: <20100422205213.C6198312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6894

devang.patel <devang.patel at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |devang.patel at gmail.com
         Resolution|                            |FIXED

--- Comment #2 from devang.patel <devang.patel at gmail.com> 2010-04-22 15:52:13 CDT ---
Fixed. r102127.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 16:02:43 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 16:02:43 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6872] BASH 4.1 fails to compile using self hosted
 clang (r101748)
In-Reply-To: <bug-6872-206@http.llvm.org/bugs/>
References: <bug-6872-206@http.llvm.org/bugs/>
Message-ID: <20100422210243.5D7F3312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6872

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #2 from Douglas Gregor <dgregor at apple.com> 2010-04-22 16:02:42 CDT ---
We can't do anything without preprocessed source. Please attach preprocessor
source and re-open the bug.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 16:04:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 16:04:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6756] exception object not destructed
In-Reply-To: <bug-6756-206@http.llvm.org/bugs/>
References: <bug-6756-206@http.llvm.org/bugs/>
Message-ID: <20100422210431.CC9C5312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6756

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from John McCall <rjmccall at apple.com> 2010-04-22 16:04:31 CDT ---
Fixed in r102039.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 17:29:21 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 17:29:21 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6896] New: Unhelpful diagnostic on attempt to
 partially specialize member function
Message-ID: <bug-6896-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6896

           Summary: Unhelpful diagnostic on attempt to partially
                    specialize member function
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: quality-of-implementation
          Severity: enhancement
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Yes, this is invalid, but the compiler should notice and point out the user's
mistake.


jyasskin-macbookpro15:~/tmp$ cat test.cc
class F {
 public:
  template <typename T>
  F(const T& value);
};

// Intending to partially-specialize F::F...
template <unsigned N>
inline F::F(const char (&value)[N]) {
}
jyasskin-macbookpro15:~/tmp$ ~/src/llvm/clang/obj/Debug/bin/clang++  -c test.cc
test.cc:9:11: error: out-of-line definition of 'F' does not match any
declaration in 'F'
inline F::F(const char (&value)[N]) {
       ~~~^
1 error generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 18:24:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 18:24:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 5998] thumb2 UBFX not generated
In-Reply-To: <bug-5998-206@http.llvm.org/bugs/>
References: <bug-5998-206@http.llvm.org/bugs/>
Message-ID: <20100422232431.3DC11312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=5998

Jim Grosbach <grosbach at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Jim Grosbach <grosbach at apple.com> 2010-04-22 18:24:30 CDT ---
Fixed in r102144

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 18:41:19 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 18:41:19 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6897] New: clang has a hardcoded version 1.0 (which
 should be 1.1 instead)
Message-ID: <bug-6897-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6897

           Summary: clang has a hardcoded version 1.0 (which should be 1.1
                    instead)
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Driver
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: doko at ubuntu.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4719)
 --> (http://llvm.org/bugs/attachment.cgi?id=4719)
suggested fix for 2.7

didn't check this in a vanilla build, just in a package build. patch attached,
tested on the 2.7 branch only. maybe a better fix should be applied on the
trunk.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 19:51:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 19:51:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6888] Missing error string
In-Reply-To: <bug-6888-206@http.llvm.org/bugs/>
References: <bug-6888-206@http.llvm.org/bugs/>
Message-ID: <20100423005104.058AE312836B@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6888

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-22 19:51:03 CDT ---
Fixed in r102154, thanks.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 20:43:19 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 20:43:19 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6898] New: Want better diagnostic for missing
 std::hash<T> specialization
Message-ID: <bug-6898-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6898

           Summary: Want better diagnostic for missing std::hash<T>
                    specialization
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: quality-of-implementation
          Severity: enhancement
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


$ cat test.cc
#include <ext/hash_set>
#include <string>

void foo() {
    __gnu_cxx::hash_set<std::string> hs;
    hs.insert("Hello world");
}
$ clang++ -c test.cc
In file included from test.cc:1:
In file included from /usr/include/c++/4.0.0/ext/hash_set:64:
/usr/include/c++/4.0.0/ext/hashtable.h:596:16: error: type 'hasher const' (aka
'__gnu_cxx::hash<std::basic_string<char> > const') does not provide a call
operator
      { return _M_hash(__key) % __n; }
               ^~~~~~~
/usr/include/c++/4.0.0/ext/hashtable.h:600:16: note: in instantiation of member
function '__gnu_cxx::hashtable<std::basic_string<char>,
std::basic_string<char>, __gnu_cxx::hash<std::basic_string<char> >,
      std::_Identity<std::basic_string<char> >,
std::equal_to<std::basic_string<char> >, std::allocator<std::basic_string<char>
> >::_M_bkt_num_key' requested here
      { return _M_bkt_num_key(_M_get_key(__obj), __n); }
               ^
/usr/include/c++/4.0.0/ext/hashtable.h:1006:31: note: in instantiation of
member function '__gnu_cxx::hashtable<std::basic_string<char>,
std::basic_string<char>, __gnu_cxx::hash<std::basic_string<char> >,
      std::_Identity<std::basic_string<char> >,
std::equal_to<std::basic_string<char> >, std::allocator<std::basic_string<char>
> >::_M_bkt_num' requested here
                          size_type __new_bucket = _M_bkt_num(__first->_M_val,
                                                   ^
/usr/include/c++/4.0.0/ext/hashtable.h:438:2: note: in instantiation of member
function '__gnu_cxx::hashtable<std::basic_string<char>,
std::basic_string<char>, __gnu_cxx::hash<std::basic_string<char> >,
      std::_Identity<std::basic_string<char> >,
std::equal_to<std::basic_string<char> >, std::allocator<std::basic_string<char>
> >::resize' requested here
        resize(_M_num_elements + 1);
        ^
In file included from test.cc:1:
/usr/include/c++/4.0.0/ext/hash_set:206:43: note: in instantiation of member
function '__gnu_cxx::hashtable<std::basic_string<char>,
std::basic_string<char>, __gnu_cxx::hash<std::basic_string<char> >,
      std::_Identity<std::basic_string<char> >,
std::equal_to<std::basic_string<char> >, std::allocator<std::basic_string<char>
> >::insert_unique' requested here
        pair<typename _Ht::iterator, bool> __p = _M_ht.insert_unique(__obj);
                                                 ^
test.cc:6:5: note: in instantiation of member function
'__gnu_cxx::hash_set<std::basic_string<char>,
__gnu_cxx::hash<std::basic_string<char> >,
std::equal_to<std::basic_string<char> >, std::allocator<std::basic_string<char>
>
      >::insert' requested here
    hs.insert("Hello world");
    ^
1 error generated.



Thoughts on what we'd like from this:

* Ideally, we should say "Please provide a specialization for
'__gnu_cxx::hash<std::string>' that provides a 'size_t operator()(const
std::string&) const' member function". I'm not sure how to get this though.

* Simply removing the definition of "template<class _Key>struct hash" in
hash_fun.h produces a better message:

In file included from test.cc:1:
In file included from /usr/include/c++/4.0.0/ext/hash_set:64:
/usr/include/c++/4.0.0/ext/hashtable.h:304:29: error: implicit instantiation of
undefined template '__gnu_cxx::hash<std::basic_string<char> >'
      hasher                _M_hash;
                            ^
In file included from test.cc:1:
/usr/include/c++/4.0.0/ext/hash_set:102:11: note: in instantiation of template
class '__gnu_cxx::hashtable<std::basic_string<char>, std::basic_string<char>,
__gnu_cxx::hash<std::basic_string<char> >,
      std::_Identity<std::basic_string<char> >,
std::equal_to<std::basic_string<char> >, std::allocator<std::basic_string<char>
> >' requested here
      _Ht _M_ht;
          ^
test.cc:5:38: note: in instantiation of template class
'__gnu_cxx::hash_set<std::basic_string<char>,
__gnu_cxx::hash<std::basic_string<char> >,
std::equal_to<std::basic_string<char> >, std::allocator<std::basic_string<char>
>
      >' requested here
    __gnu_cxx::hash_set<std::string> hs;
                                     ^
In file included from test.cc:1:
In file included from /usr/include/c++/4.0.0/ext/hash_set:64:
In file included from /usr/include/c++/4.0.0/ext/hashtable.h:71:
/usr/include/c++/4.0.0/ext/hash_fun.h:71:12: note: template is declared here
    struct hash ;
           ^
1 error generated.


... but this requires changing libstdc++ and waiting for it to roll out.

* In the "in instantiation of template class", we should avoid re-printing the
"__gnu_cxx::hashtable<std::basic_string<char>, std::basic_string<char>,
__gnu_cxx::hash<std::basic_string<char> >,
      std::_Identity<std::basic_string<char> >,
std::equal_to<std::basic_string<char> >, std::allocator<std::basic_string<char>
> >" part. Perhaps we should replace it with "..." when it's the same as the
previous class name. Maybe say "in instantiation of member function from same
class '...::_M_bkt_num'.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 21:03:16 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 21:03:16 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6874] clang cannot compile this decl yet (namespace
	af = ...)
In-Reply-To: <bug-6874-206@http.llvm.org/bugs/>
References: <bug-6874-206@http.llvm.org/bugs/>
Message-ID: <20100423020316.DD2A6312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6874

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Douglas Gregor <dgregor at apple.com> 2010-04-22 21:03:16 CDT ---
Fixed in r102160.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 21:08:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 21:08:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6830] Clang complains the qualified class name in a
 constructor call doesn't refer to a value
In-Reply-To: <bug-6830-206@http.llvm.org/bugs/>
References: <bug-6830-206@http.llvm.org/bugs/>
Message-ID: <20100423020851.92AA1312836E@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6830

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Douglas Gregor <dgregor at apple.com> 2010-04-22 21:08:51 CDT ---
I applied Eli's patch in r102161.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 21:10:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 21:10:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6344] failed assertion with 2010-02-17 trunk
In-Reply-To: <bug-6344-206@http.llvm.org/bugs/>
References: <bug-6344-206@http.llvm.org/bugs/>
Message-ID: <20100423021056.7E1A4312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6344

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #8 from Douglas Gregor <dgregor at apple.com> 2010-04-22 21:10:55 CDT ---
We can't fix a bug without preprocessing sources on which to test. Please
attach preprocessed source and re-open this bug.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 21:20:01 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 21:20:01 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6899] New: Clang cannot mangle dependent template
	names
Message-ID: <bug-6899-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6899

           Summary: Clang cannot mangle dependent template names
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


The following code causes an assertion in name mangling:

  template<typename T>
  struct supermeta {
    template<typename U>
    struct apply {
      typedef T U::*type;
    };
  };

  struct X { };

  template<typename T, typename U>
  typename supermeta<T>::template apply<U>::type f();

  void test_f() {
    f<int, X>();
  }


Assertion failed: (TD && "FIXME: Support dependent template names"), function
mangleUnresolvedScope, file
/Users/dgregor/Projects/llvm/tools/clang/lib/CodeGen/Mangle.cpp, line 477.
Stack dump:
0.    Program arguments: /Users/dgregor/Projects/llvm-xcode/bin/Debug/clang
-cc1 -emit-llvm -o -
/Users/dgregor/Projects/llvm/tools/clang/test/CodeGenCXX/mangle-template.cpp

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 21:42:38 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 21:42:38 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6885] Diagnostic for friend declaration inside of a
 namespace with same name
In-Reply-To: <bug-6885-206@http.llvm.org/bugs/>
References: <bug-6885-206@http.llvm.org/bugs/>
Message-ID: <20100423024238.601C7312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6885

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from John McCall <rjmccall at apple.com> 2010-04-22 21:42:38 CDT ---
It turns out that our handling of elaborated type specifiers was wrong in many
ways.

Fixed in r102164.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 22:32:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 22:32:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6899] Clang cannot mangle dependent template names
In-Reply-To: <bug-6899-206@http.llvm.org/bugs/>
References: <bug-6899-206@http.llvm.org/bugs/>
Message-ID: <20100423033218.0EFF2312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6899

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-22 22:32:17 CDT ---
Fixed in r102168.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 22:41:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 22:41:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6900] New: Can't compile compound expressions as
	lvalues
Message-ID: <bug-6900-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6900

           Summary: Can't compile compound expressions as lvalues
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
            Blocks: 6023


cove:clang dgregor$ cat t2.C
void f(int i, int j, int k) {
  (i += j) += k;
}
cove:clang dgregor$ clang++ t2.C
t2.C:2:4: error: cannot compile this l-value expression yet
  (i += j) += k;
   ^~~~~~
1 error generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 22 23:18:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 22 Apr 2010 23:18:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6900] Can't compile compound expressions as lvalues
In-Reply-To: <bug-6900-206@http.llvm.org/bugs/>
References: <bug-6900-206@http.llvm.org/bugs/>
Message-ID: <20100423041818.D1741312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6900

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-22 23:18:18 CDT ---
Fixed in r102170.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 01:07:26 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 01:07:26 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6860] ice: Live range not found
In-Reply-To: <bug-6860-206@http.llvm.org/bugs/>
References: <bug-6860-206@http.llvm.org/bugs/>
Message-ID: <20100423060726.6B30D312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6860

Evan Cheng <evan.cheng at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Evan Cheng <evan.cheng at apple.com> 2010-04-23 01:07:26 CDT ---
I believe this is fixed by r102027. Please verify.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 06:21:57 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 06:21:57 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6901] New: iostream compilation broken
Message-ID: <bug-6901-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6901

           Summary: iostream compilation broken
           Product: clang
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: luc_j_bourhis at mac.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


I checked out llvm and clang trunk and compiled, following instructions on the
clang web page. Then (on MacOS 10.6.3),

~> cat iostream_trivia.cpp
#include <iostream>
int main() {
  std::cout << 0.12345678 << "\n";
}

~> clang iostream_trivia.cpp
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
In file included from /usr/include/sys/select.h:78:
/usr/include/sys/_structs.h:191:40: error: elaborated type refers to a typedef
__darwin_fd_isset(int _n, const struct fd_set *_p)
                                       ^
/usr/include/sys/_structs.h:186:3: note: declared here
} fd_set;
  ^
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
/usr/include/sys/select.h:137:19: error: reference to 'fd_set' is ambiguous
int      pselect(int, fd_set * __restrict, fd_set * __restrict,
                      ^
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
In file included from /usr/include/sys/select.h:78:
/usr/include/sys/_structs.h:186:3: note: candidate found by name lookup is
'fd_set'
} fd_set;
  ^
/usr/include/sys/_structs.h:184:16: note: candidate found by name lookup is
'fd_set'
typedef struct fd_set {
               ^
/usr/include/sys/_structs.h:191:40: note: candidate found by name lookup is
'fd_set'
__darwin_fd_isset(int _n, const struct fd_set *_p)
                                       ^
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
/usr/include/sys/select.h:137:40: error: reference to 'fd_set' is ambiguous
int      pselect(int, fd_set * __restrict, fd_set * __restrict,
                                           ^
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
In file included from /usr/include/sys/select.h:78:
/usr/include/sys/_structs.h:186:3: note: candidate found by name lookup is
'fd_set'
} fd_set;
  ^
/usr/include/sys/_structs.h:184:16: note: candidate found by name lookup is
'fd_set'
typedef struct fd_set {
               ^
/usr/include/sys/_structs.h:191:40: note: candidate found by name lookup is
'fd_set'
__darwin_fd_isset(int _n, const struct fd_set *_p)
                                       ^
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
/usr/include/sys/select.h:138:3: error: reference to 'fd_set' is ambiguous
                fd_set * __restrict, const struct timespec * __restrict,
                ^
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
In file included from /usr/include/sys/select.h:78:
/usr/include/sys/_structs.h:186:3: note: candidate found by name lookup is
'fd_set'
} fd_set;
  ^
/usr/include/sys/_structs.h:184:16: note: candidate found by name lookup is
'fd_set'
typedef struct fd_set {
               ^
/usr/include/sys/_structs.h:191:40: note: candidate found by name lookup is
'fd_set'
__darwin_fd_isset(int _n, const struct fd_set *_p)
                                       ^
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
In file included from /usr/include/sys/select.h:152:
/usr/include/sys/_select.h:39:18: error: reference to 'fd_set' is ambiguous
int      select(int, fd_set * __restrict, fd_set * __restrict,
                     ^
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
In file included from /usr/include/sys/select.h:78:
/usr/include/sys/_structs.h:186:3: note: candidate found by name lookup is
'fd_set'
} fd_set;
  ^
/usr/include/sys/_structs.h:184:16: note: candidate found by name lookup is
'fd_set'
typedef struct fd_set {
               ^
/usr/include/sys/_structs.h:191:40: note: candidate found by name lookup is
'fd_set'
__darwin_fd_isset(int _n, const struct fd_set *_p)
                                       ^
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
In file included from /usr/include/sys/select.h:152:
/usr/include/sys/_select.h:39:39: error: reference to 'fd_set' is ambiguous
int      select(int, fd_set * __restrict, fd_set * __restrict,
                                          ^
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
In file included from /usr/include/sys/select.h:78:
/usr/include/sys/_structs.h:186:3: note: candidate found by name lookup is
'fd_set'
} fd_set;
  ^
/usr/include/sys/_structs.h:184:16: note: candidate found by name lookup is
'fd_set'
typedef struct fd_set {
               ^
/usr/include/sys/_structs.h:191:40: note: candidate found by name lookup is
'fd_set'
__darwin_fd_isset(int _n, const struct fd_set *_p)
                                       ^
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
In file included from /usr/include/sys/select.h:152:
/usr/include/sys/_select.h:40:3: error: reference to 'fd_set' is ambiguous
                fd_set * __restrict, struct timeval * __restrict)
                ^
In file included from iostream.cpp:1:
In file included from /usr/include/c++/4.2.1/iostream:43:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/c++config.h:41:
In file included from
/usr/include/c++/4.2.1/i686-apple-darwin10/x86_64/bits/os_defines.h:61:
In file included from /usr/include/unistd.h:519:
In file included from /usr/include/sys/select.h:78:
/usr/include/sys/_structs.h:186:3: note: candidate found by name lookup is
'fd_set'
} fd_set;
  ^
/usr/include/sys/_structs.h:184:16: note: candidate found by name lookup is
'fd_set'
typedef struct fd_set {
               ^
/usr/include/sys/_structs.h:191:40: note: candidate found by name lookup is
'fd_set'
__darwin_fd_isset(int _n, const struct fd_set *_p)
                                       ^
7 errors generated.

~> uname -a
Darwin Zeus.local 10.3.0 Darwin Kernel Version 10.3.0: Fri Feb 26 11:58:09 PST
2010; root:xnu-1504.3.12~1/RELEASE_I386 i386

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 07:36:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 07:36:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6902] New: PowerPC JIT fails to resolve
 'llvm.memset.*' Intrinsics
Message-ID: <bug-6902-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6902

           Summary: PowerPC JIT fails to resolve 'llvm.memset.*'
                    Intrinsics
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: PowerPC
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: xerxes at zafena.se
                CC: llvmbugs at cs.uiuc.edu


llvm testcase : cat jit-ppc-memset-test.ll
declare void @llvm.memset.i32(i8*, i8, i32, i32)

define i32 @main() {
 %buffer = alloca [80 x i8]
 %tmp = bitcast [80 x i8]* %buffer to i8*
 call void @llvm.memset.i32(i8* %tmp, i8 14, i32 80,i32 0)
 ret i32 0
}

FAIL when running on PPC:
./llvm-2.7-shared/Release+Debug/bin/llvm-as jit-ppc-memset-test.ll
./llvm-2.7-shared/Release+Debug/bin/lli < jit-ppc-memset-test.bc
LLVM ERROR: Cannot yet select: 0x100424e8: i32 = ExternalSymbol'memset' [ID=4]

PASSES when running on X86:
xerxes at xerxes-J464X:~$ llvm-as  jit-ppc-memset.ll 
xerxes at xerxes-J464X:~$ lli < jit-ppc-memset.bc 

tested using llvm 2.7 releasebranch

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 07:50:58 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 07:50:58 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6902] PowerPC JIT fails to resolve 'llvm.memset.*'
	Intrinsics
In-Reply-To: <bug-6902-206@http.llvm.org/bugs/>
References: <bug-6902-206@http.llvm.org/bugs/>
Message-ID: <20100423125058.15EA4312836B@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6902

Xerxes R?nby <xerxes at zafena.se> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #1 from Xerxes R?nby <xerxes at zafena.se> 2010-04-23 07:50:57 CDT ---
It turned out this bug was caused by the PowerPC workaround from bug 5201
applied on my workdir: http://llvm.org/bugs/attachment.cgi?id=4631

when removing this workaround the testcase passed on ppc.

closing as invalid.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 12:19:26 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 12:19:26 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6783] crash on invalid: (DefinitionData && "queried
 property of class with no definition)"
In-Reply-To: <bug-6783-206@http.llvm.org/bugs/>
References: <bug-6783-206@http.llvm.org/bugs/>
Message-ID: <20100423171926.249E4312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6783

Sebastian Redl <sebastian.redl at getdesigned.at> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |sebastian.redl at getdesigned.
                   |                            |at
         Resolution|                            |FIXED

--- Comment #1 from Sebastian Redl <sebastian.redl at getdesigned.at> 2010-04-23 12:19:25 CDT ---
Fixed in r102186.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 12:24:10 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 12:24:10 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6903] New: Crash on missing double-quote
Message-ID: <bug-6903-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6903

           Summary: Crash on missing double-quote
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: crash-on-invalid
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


terfin:clang dgregor$ cat t.C
void f(...);
struct X {
  X() { f("X::X()\n"); }
  ~X() { f(X::~X()\n"); }
};

terfin:clang dgregor$ clang++ t.C
t.C:4:21: error: missing terminating '"' character
  ~X() { f(X::~X()\n"); }
                    ^
t.C:4:19: error: expected ')'
  ~X() { f(X::~X()\n"); }
                  ^
t.C:4:11: note: to match this '('
  ~X() { f(X::~X()\n"); }
          ^
t.C:6:1: error: expected '}'

^
Assertion failed: (!PP.getSourceManager().isBeforeInTranslationUnit(origLoc,
Tok.getLocation()) && "We consumed more than the cached tokens!"), function
ParseLexedMethodDefs, file
/Users/buildslave/zorg/buildbot/smooshlab/slave/build.clang-x86_64-darwin10-miniswb/clang.src/clang/clang/tools/clang/lib/Parse/ParseCXXInlineMethods.cpp,
line 237.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 12:29:36 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 12:29:36 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6276] Linkage specifications are also treated as a
 storage-class-specifiers
In-Reply-To: <bug-6276-206@http.llvm.org/bugs/>
References: <bug-6276-206@http.llvm.org/bugs/>
Message-ID: <20100423172936.9D269312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6276

Sebastian Redl <sebastian.redl at getdesigned.at> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |sebastian.redl at getdesigned.
                   |                            |at
         Resolution|                            |INVALID

--- Comment #2 from Sebastian Redl <sebastian.redl at getdesigned.at> 2010-04-23 12:29:36 CDT ---
Yes it is, see C++03 7.5p7.

"""
The form of linkage-specification that contains a brace-enclosed
declaration-seq does not affect whether the contained declarations are
definitions or not (3.1); the form of linkage-specification directly containing
a single declaration is treated as an extern specifier (7.1.1) for the purpose
of determining whether the contained declaration is a definition.
"""

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 14:56:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 14:56:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6904] New: clang allows redefinition of static
	const member
Message-ID: <bug-6904-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6904

           Summary: clang allows redefinition of static const member
           Product: clang
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: akyrtzi at gmail.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


struct S {
    static const int val = 1;
};
const int S::val = 1; // or any other value


clang compiles the above sample without complaints.
It should say something like:

error: member "S::val" has already been initialized

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 15:00:13 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 15:00:13 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6905] New: Conservative Unwind Pass
Message-ID: <bug-6905-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6905

           Summary: Conservative Unwind Pass
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Transformation Utilities
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alexmac at adobe.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4723)
 --> (http://llvm.org/bugs/attachment.cgi?id=4723)
conservative unwind pass

This file implements a simple interprocedural pass which walks the call-graph,
conservatively clearing and setting the nounwind attribute of functions and
function calls.  A function is assumed to throw an exception if it contains an
unwind instruction, an indirect call, or a call to an external function. This
pass is implemented as a bottom-up traversal of the call-graph.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 15:10:04 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 15:10:04 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6906] New: VFP instructions miscompiled
Message-ID: <bug-6906-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6906

           Summary: VFP instructions miscompiled
           Product: libraries
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alexmac at adobe.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4724)
 --> (http://llvm.org/bugs/attachment.cgi?id=4724)
Fix the ARM tablegen files to output certain instructions in an alternate
syntax

The assembler currently shipping on OSX with xcode fails to compile some VFP
instructions when using a particular syntax, this patch works around those bugs
by outputting the same instructions in a syntax that the assembler can handle
correctly. I assume (although haven't verified) that gcc must not be using the
syntax llvm currently uses or it would also be producing invalid code when
compiling for armv6 with vfp enabled.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 15:12:00 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 15:12:00 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6907] New: several versions of libstdc++ confuse
	#include_next
Message-ID: <bug-6907-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6907

           Summary: several versions of libstdc++ confuse #include_next
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: marc.glisse at normalesup.org
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Hello,

I am using a precompiled version of 2.7 RC2.

I have several versions of the libstdc++ headers installed
(/usr/include/c++/4.* for a number of values of *), and all these directories
get appended in the include search path. Now when I #include <fenv.h>, it finds
the one from 4.4, which contains #include_next <fenv.h>, expecting it to be the
glibc one from /usr/include. However, clang takes the one from 4.3, which uses
the same macro guard and thus appears empty.

I suppose I can work around it with -nostdinc and appropriate -isystem. It
seems like it would be better if clang chose one libstdc++ include directory
and stuck to it. Or maybe I just missed the documentation on the best way to
handle this.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 15:12:28 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 15:12:28 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6901] iostream compilation broken
In-Reply-To: <bug-6901-206@http.llvm.org/bugs/>
References: <bug-6901-206@http.llvm.org/bugs/>
Message-ID: <20100423201228.5A441312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6901

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |rjmccall at apple.com
         Resolution|                            |FIXED

--- Comment #3 from John McCall <rjmccall at apple.com> 2010-04-23 15:12:28 CDT ---
Fixed in r102198, sorry.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 15:42:37 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 15:42:37 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6908] New: stackoverflow when using
 LegalizeAllNodesNotLeadingTo or AssignOrderingToNode on large bitcode files
Message-ID: <bug-6908-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6908

           Summary: stackoverflow when using LegalizeAllNodesNotLeadingTo
                    or AssignOrderingToNode on large bitcode files
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alexmac at adobe.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4725)
 --> (http://llvm.org/bugs/attachment.cgi?id=4725)
proposed fix

Some of the code we generate ends up so large and complex that we can run out
of stack space when running these recursive codegen functions
(LegalizeAllNodesNotLeadingTo and AssignOrderingToNode). We've altered them to
be iterative so that they work with large bitcode files.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 15:49:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 15:49:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6909] New: Path::isRootDirectory not defined for
	unix
Message-ID: <bug-6909-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6909

           Summary: Path::isRootDirectory not defined for unix
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: System Library
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alexmac at adobe.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4726)
 --> (http://llvm.org/bugs/attachment.cgi?id=4726)
add unix support for Path::isRootDirectory

Path::isRootDirectory is defined for win32 but not for unix.  I've attached a
patch that adds a unix definition, but I can't see any uses in the codebase so
perhaps it should just be removed entirely?

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 16:18:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 16:18:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 5166] verifier misses incorrect declaration of
 intrinsics (sometimes)
In-Reply-To: <bug-5166-206@http.llvm.org/bugs/>
References: <bug-5166-206@http.llvm.org/bugs/>
Message-ID: <20100423211859.5F540312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=5166

Nick Lewycky <nlewycky at google.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |nlewycky at google.com
         Resolution|                            |DUPLICATE

--- Comment #2 from Nick Lewycky <nlewycky at google.com> 2010-04-23 16:18:59 CDT ---
Yep.

*** This bug has been marked as a duplicate of bug 6491 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 16:23:32 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 16:23:32 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6903] Crash on missing double-quote
In-Reply-To: <bug-6903-206@http.llvm.org/bugs/>
References: <bug-6903-206@http.llvm.org/bugs/>
Message-ID: <20100423212332.24B15312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6903

Argiris Kirtzidis <akyrtzi at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #4 from Argiris Kirtzidis <akyrtzi at gmail.com> 2010-04-23 16:23:31 CDT ---
Fixed on r102214.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 16:25:10 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 16:25:10 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6910] New: LLVM broken with -O3,
	or -O1 -ftree-vectorize
Message-ID: <bug-6910-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6910

           Summary: LLVM broken with -O3, or -O1 -ftree-vectorize
           Product: new-bugs
           Version: unspecified
          Platform: PC
               URL: https://wwws.clamav.net/bugzilla/show_bug.cgi?id=1984
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


See
https://wwws.clamav.net/bugzilla/show_bug.cgi?id=1984

CXXFLAGS="-march=pentium4 -O1 -ftree-vectorize -m32"
This causes LLVM to crash on startup.

You can easily reproduce this, by putting this code into PassManager.cpp:
int main()
{
    printf("%d\n",__alignof(FPPassManager));
    FPPassManager *P = new FPPassManager(1);
    FPPassManager *P2 = new FPPassManager(1);
    return 0;
}

alignof is 16, but new is only guaranteed to return 8-byte aligned data.

This is because LLVM uses attribute((aligned)) in SmallVector.

See these bugs on why this is neither a gcc bug, nor a glibc bug, ...
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15795
http://sourceware.org/bugzilla/show_bug.cgi?id=206

The fix is to remove #ifdef __GNUC__ __attribute((aligned)) from SmallVector
and just rely on the unions for alignment (they create alignof 4, and not 16).

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 17:21:20 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 17:21:20 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6911] New: Function pointer with template cannot be
 passed as argument
Message-ID: <bug-6911-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6911

           Summary: Function pointer with template cannot be passed as
                    argument
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: evan at chromium.org
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4727)
 --> (http://llvm.org/bugs/attachment.cgi?id=4727)
reduction of problem

Clang r102190.

$ clang++ reduce.cc
clang: Instructions.cpp:247: void llvm::CallInst::init(llvm::Value*,
llvm::Value* const*, unsigned int): Assertion `(i >= FTy->getNumParams() ||
FTy->getParamType(i) == Params[i]->getType()) && "Calling a function with a bad
signature!"' failed.

Code below, also attached.  This came up in compiling WebKit with v8.


template <class T> class Foo;

// Callback type that involves template.
typedef void (*Callback)(Foo<int> arg);
// Function that is an instance of that type.
static void f(Foo<int> arg);
// Function that takes a callback as an argument.
void g(Callback callback);

// Pass one to the other -- boom.
void h() {
  g(&f);
}

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 20:23:21 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 20:23:21 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6912] New: Spurious null pointer dereference warning
Message-ID: <bug-6912-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6912

           Summary: Spurious null pointer dereference warning
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: evan at chromium.org
                CC: llvmbugs at cs.uiuc.edu


I just tried out scan-build on a large project and reduced many of the spurious
warnings to the attached simple test case.

The pattern is that there's a global variable along with a function that may
initialize it, and anyone who uses it tests either that the variable has been
initialized or that the initialization function succeeds.  So we always either
return or have the variable end up initialized.

This code is kind of crazy, I agree, but it's a real open source library (NSS,
from http://www.mozilla.org/projects/security/pki/nss/) that does this.

The error message is as follows:
crazy.c:19:18: warning: Dereference of null pointer (loaded from variable
'global_p')
  printf("%d\n", *global_p);

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 22:00:58 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 22:00:58 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6913] New: LLVM Segmentation Fault on Compile
Message-ID: <bug-6913-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6913

           Summary: LLVM Segmentation Fault on Compile
           Product: new-bugs
           Version: 2.6
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: MatthewRDutton.OSS at gmail.com
                CC: llvmbugs at cs.uiuc.edu


LLVM SEGFAULTs when executing the command:
llvm-gcc -c --emit-llvm -o array.bc array.c

With array.c:
#include <stdio.h>
#include <malloc.h>

int main()
{
    int x[10][10];
    int (*p)[] = x;   // <-- this line is what triggered it

    int i;

    for(i = 0; i < 10; ++i)
    {
        p[i][i] = i;
    }
}

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 23 23:23:34 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 23 Apr 2010 23:23:34 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6914] New: Instruction does not dominate all uses
Message-ID: <bug-6914-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6914

           Summary: Instruction does not dominate all uses
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu, chenyang at cs.utah.edu


Seen on x64 Linux.  I cannot repro with opt.

[regehr at gamow tmp408]$ clang -v
clang version 1.5 (trunk 102110)
Target: x86_64-unknown-linux-gnu
Thread model: posix

[regehr at gamow tmp408]$ clang -O -c -w small.c

Instruction does not dominate all uses!
  %tmp7 = load i16* %tmp6                         ; <i16> [#uses=1]
  %tmp = trunc i16 %tmp7 to i8                    ; <i8> [#uses=2]
Instruction does not dominate all uses!
  %tmp = trunc i16 %tmp7 to i8                    ; <i8> [#uses=2]
  %tmp23 = mul i8 %tmp, -2                        ; <i8> [#uses=1]
Instruction does not dominate all uses!
  %tmp = trunc i16 %tmp7 to i8                    ; <i8> [#uses=2]
  %tmp24 = mul i8 %tmp, -1                        ; <i8> [#uses=1]
Instruction does not dominate all uses!
  %1 = add i8 %0, %tmp24                          ; <i8> [#uses=1]
  %lsr.iv.next = add i8 %lsr.iv, %1               ; <i8> [#uses=1]
Broken module found, compilation aborted!
0  clang           0x00000000012582af
1  clang           0x000000000125a1b2
2  libpthread.so.0 0x00007f6410bdf190
3  libc.so.6       0x00007f640fee54b5 gsignal + 53
4  libc.so.6       0x00007f640fee8f50 abort + 384
5  clang           0x000000000120304f
6  clang           0x00000000011db66d
7  clang           0x00000000011db84b
8  clang           0x00000000011dbb0f
9  clang           0x0000000000415647
10 clang           0x0000000000415ec5
11 clang           0x00000000006102bf
12 clang           0x00000000004199b1
13 clang           0x0000000000409ccb
14 clang           0x000000000040d043 main + 2499
15 libc.so.6       0x00007f640fed0abd __libc_start_main + 253
16 clang           0x00000000004074f9
Stack dump:
0.    Program arguments:
/uusoc/exports/scratch/regehr/z/compiler-install/llvm-gcc-r102110-install/bin/clang
-cc1 -triple x86_64-unknown-linux-gnu -S -disable-free -main-file-name small.c
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -resource-dir
/uusoc/exports/scratch/regehr/z/compiler-install/llvm-gcc-r102110-install/lib/clang/1.5
-O2 -w -ferror-limit 19 -ftemplate-backtrace-limit 10 -fmessage-length 89
-fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics -o /tmp/cc-YT2vqr.s
-x c small.c 
1.    <eof> parser at end of file
2.    Code generation
3.    Running pass 'Module Verifier' on function '@int323'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 03:05:25 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 03:05:25 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6915] New: Assertion triggered by type-dependent
 enum member declaration.
Message-ID: <bug-6915-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6915

           Summary: Assertion triggered by type-dependent enum member
                    declaration.
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: zaffanella at cs.unipr.it
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


The following program causes an assertion failure:

template <typename T>
class D {
  enum T::X v;
};


$ llvm/Debug/bin/clang++ -fsyntax-only enum.cc
clang: ParseDecl.cpp:1942: void
clang::Parser::ParseEnumSpecifier(clang::SourceLocation, clang::DeclSpec&,
const clang::Parser::ParsedTemplateInfo&, clang::AccessSpecifier): Assertion
`!IsDependent && "didn't expect dependent enum"' failed.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 03:07:12 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 03:07:12 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6916] New: [meta] clang 2.7 tests on various
	applications
Message-ID: <bug-6916-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6916

           Summary: [meta] clang 2.7 tests on various applications
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I'm trying to build/test several application with clang 2.7.
I'll mark any bugs I find as blocking this one.

It would be nice to have these fixed for clang 2.8.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 05:03:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 05:03:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6917] New: clang -fstack-protector doesn't work
Message-ID: <bug-6917-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6917

           Summary: clang -fstack-protector doesn't work
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


-fstack-protector doesn't work

$ cat >x.c <<EOF
int main()
{
    char buf[4096];
    return 0;
}
EOF

$ clang -fstack-protector x.c
/tmp/cc-5HBV9O.o: In function `main':
x.c:(.text+0xf): undefined reference to `__stack_chk_guard'
x.c:(.text+0x2c): undefined reference to `__stack_chk_guard'
collect2: ld returned 1 exit status
clang: error: linker command failed with exit code 1 (use -v to see invocation)

It looks like it is using the wrong symbols, because gcc creates this:
U __stack_chk_fail
0000000000000000 T main

No reference to __stack_chk_guard.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 05:08:42 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 05:08:42 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6917] clang -fstack-protector doesn't work
In-Reply-To: <bug-6917-206@http.llvm.org/bugs/>
References: <bug-6917-206@http.llvm.org/bugs/>
Message-ID: <20100424100842.97765312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6917

Benjamin Kramer <benny.kra at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |benny.kra at gmail.com
         Resolution|                            |DUPLICATE

--- Comment #3 from Benjamin Kramer <benny.kra at gmail.com> 2010-04-24 05:08:42 CDT ---


*** This bug has been marked as a duplicate of bug 5094 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 07:25:24 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 07:25:24 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6918] New: clang++: lazy instantiation issue?
Message-ID: <bug-6918-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6918

           Summary: clang++: lazy instantiation issue?
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
            Blocks: 6916


This code compiles with g++, but not with clang++:

$ cat >test.cpp <<EOF
template <class Scalar>
inline bool equal(const Scalar &v, const Scalar &c)
{
  return ABS(v-c);
}
template < class T >
inline T
ABS( T a )
{ return (0 < a) ? a : -a ; }


int main()
{
    float a, b;
    equal(a,b);
}
EOF

$ clang++ test.cpp
x.cpp:5:10: error: use of undeclared identifier 'ABS'
  return ABS(v-c);
         ^
x.cpp:16:5: note: in instantiation of function template specialization
'equal<float>' requested here
    equal(a,b);
    ^
2 diagnostics generated.


This occurs when building blender 2.49b:
/home/edwin/clangtests/blender-2.49b/intern/elbeem/intern/ntl_vector3dim.h:998:11:
error: use of undeclared identifier 'ABS'
  return (ABS(v[0]-c[0]) +
          ^
/home/edwin/clangtests/blender-2.49b/intern/elbeem/intern/ntl_bsptree.cpp:192:7:
note: in instantiation of function template
      specialization 'equal<float>' requested here
                if( equal(triangleNormal, ntlVec3Gfx(0.0)) ) continue;
                    ^
2 diagnostics generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 10:13:08 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 10:13:08 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6918] clang++: lazy instantiation issue?
In-Reply-To: <bug-6918-206@http.llvm.org/bugs/>
References: <bug-6918-206@http.llvm.org/bugs/>
Message-ID: <20100424151308.872D2312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6918

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-24 10:13:08 CDT ---
Clang is right to reject this. ABS is looked up at template definition time
(which finds nothing), then augmented by the results of argument-dependent
lookup (which does not add anything, since "float" has no associated
namespaces).

GCC does extra lookups not permitted by the C++ standard, and therefore accepts
this ill-formed code.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 10:27:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 10:27:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6919] New: clang++: inline friend operator< symbol
 emitted as non-inline causing linker errors
Message-ID: <bug-6919-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6919

           Summary: clang++: inline friend operator< symbol emitted as
                    non-inline causing linker errors
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
            Blocks: 6916


Blender gives these linker errors:
../../lib/libbf_elbeem.a(elbeem.o): In function `operator<(ntlTriangle const&,
ntlTriangle const&)':
/home/edwin/clangtests/blender-2.49b/intern/elbeem/intern/elbeem.cpp:(.text+0x50):
multiple definition of `operator<(ntlTriangle const&, ntlTriangle const&)'
../../lib/libbf_elbeem.a(utilities.o):/home/edwin/clangtests/blender-2.49b/intern/elbeem/intern/utilities.cpp:(.text+0x2d50):
first defined here


$ cat >testcase.h <<EOF
class ntlTriangle
{
public:
    inline friend bool operator<(const ntlTriangle &lhs, const ntlTriangle
&rhs);

};

bool operator<(const ntlTriangle &lhs,const ntlTriangle &rhs)
{
    return true;
}
EOF

This header is included in 2 cpp files, so:
$ cp testcase.h t1.cpp
$ cp testcase.h t2.cpp
$ clang++ t1.cpp t2.cpp
/tmp/cc-BjvUfe.o: In function `operator<(ntlTriangle const&, ntlTriangle
const&)':
t2.cpp:(.text+0x0): multiple definition of `operator<(ntlTriangle const&,
ntlTriangle const&)'
/tmp/cc-c6oItc.o:t1.cpp:(.text+0x0): first defined here

GCC accepts this code and causes no linker errors.
As a workaround, if I add 'inline' to the definition then clang gets it right
too.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 10:53:37 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 10:53:37 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6920] New: clang++: doesn't emit inline operator
 delete from parent class
Message-ID: <bug-6920-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6920

           Summary: clang++: doesn't emit inline operator delete from
                    parent class
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Yet another blender link failure:
../../lib/libbf_converter.a(BL_BlenderDataConversion.o): In function
`btAxisSweep3Internal<unsigned int>::~btAxisSweep3Internal()':
/home/edwin/clangtests/blender-2.49b/source/gameengine/Converter/BL_BlenderDataConversion.cpp:(.gnu.linkonce.t._ZN20btAxisSweep3InternalIjED0Ev+0x27):
undefined reference to `btAxisSweep3Internal<unsigned int>::operator
delete(void*)'
../../lib/libextern_bullet.a(btAxisSweep3.o): In function
`bt32BitAxisSweep3::~bt32BitAxisSweep3()':
/home/edwin/clangtests/blender-2.49b/extern/bullet2/src/BulletCollision/BroadphaseCollision/btAxisSweep3.cpp:(.gnu.linkonce.t._ZN17bt32BitAxisSweep3D0Ev+0x27):
undefined reference to `btAxisSweep3Internal<unsigned int>::operator
delete(void*)'
collect2: ld returned 1 exit status

Here is a testcase:
class btVector3
{
};
class btOverlappingPairCache;
class btBroadphaseInterface
{
};
template < typename BP_FP_INT_TYPE > class btAxisSweep3Internal:public
  btBroadphaseInterface
{
protected:BP_FP_INT_TYPE m_bpHandleMask;
  inline void operator  delete (void *ptr)
  {
  }
public:btAxisSweep3Internal (const btVector3 & worldAabbMin,
            const btVector3 & worldAabbMax,
            BP_FP_INT_TYPE handleMask,
            BP_FP_INT_TYPE handleSentinel,
            BP_FP_INT_TYPE maxHandles =
            16384, btOverlappingPairCache * pairCache =
            0, bool disableRaycastAccelerator = false);
  virtual ~ btAxisSweep3Internal ();
};

class bt32BitAxisSweep3:public btAxisSweep3Internal < unsigned int >
{
public:bt32BitAxisSweep3 (const btVector3 & worldAabbMin,
             const btVector3 & worldAabbMax, unsigned int maxHandles =
             1500000, btOverlappingPairCache * pairCache =
             0, bool disableRaycastAccelerator = false);
};
bt32BitAxisSweep3::bt32BitAxisSweep3 (const btVector3 & worldAabbMin,
                      const btVector3 & worldAabbMax,
                      unsigned int maxHandles,
                      btOverlappingPairCache * pairCache,
                      bool disableRaycastAccelerator):
     btAxisSweep3Internal < unsigned int >(worldAabbMin, worldAabbMax,
                       0xfffffffe, 0x7fffffff, maxHandles,
                       pairCache,
                       disableRaycastAccelerator)
{
}

And here is the test script (_ZN20btAxisSweep3InternalIjEdlEP is the missing
symbol):

#!/bin/bash
/home/edwin/llvm-git/obj/Release/bin/clang++ x2.i -c -o x.o 2>/dev/null || exit
1
nm x.o | grep _ZN20btAxisSweep3InternalIjEdlEPv |grep " U " || exit 1
g++ x2.i -c -o x.o 2>/dev/null || exit 1
nm x.o | grep _ZN20btAxisSweep3InternalIjEdlEPv |grep " U " && exit 1
exit 0

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 11:29:04 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 11:29:04 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6921] clang: wrong alignment for .comm object
In-Reply-To: <bug-6921-206@http.llvm.org/bugs/>
References: <bug-6921-206@http.llvm.org/bugs/>
Message-ID: <20100424162904.375D4312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6921

T?r?k Edwin <edwintorok at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |llvmbugs at cs.uiuc.edu
          Component|Frontend                    |LLVM Codegen

--- Comment #2 from T?r?k Edwin <edwintorok at gmail.com> 2010-04-24 11:29:03 CDT ---
actually this is more of a backend problem, the IR contains align 8 in both
cases, but the assembly not.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 11:38:53 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 11:38:53 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6915] Assertion triggered by type-dependent enum
 member declaration.
In-Reply-To: <bug-6915-206@http.llvm.org/bugs/>
References: <bug-6915-206@http.llvm.org/bugs/>
Message-ID: <20100424163853.085D9312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6915

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-24 11:38:52 CDT ---
Fixed in r102247.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 11:39:02 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 11:39:02 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6649] Crash with dependent elaborated-type-specifier
In-Reply-To: <bug-6649-206@http.llvm.org/bugs/>
References: <bug-6649-206@http.llvm.org/bugs/>
Message-ID: <20100424163902.11F3A312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6649

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-24 11:39:01 CDT ---
Fixed in r102247.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 12:24:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 12:24:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6922] New: llvm demo page should feature clang demo
Message-ID: <bug-6922-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6922

           Summary: llvm demo page should feature clang demo
           Product: Website
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: General Website
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bwp at bwp.dk
                CC: llvmbugs at cs.uiuc.edu


Clang should be available as a compiler in the demo page (possibly turned on by
default, or selected in radio button/drop down box with llvm-gcc as another
option). You should be able to compile C/Obj-C/C++ code when choosing clang.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 13:07:41 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 13:07:41 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6923] New: error: definition of type 'A' conflicts
 with typedef of the same name
Message-ID: <bug-6923-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6923

           Summary: error: definition of type 'A' conflicts with typedef
                    of the same name
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ich at az2000.de
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Not exactly sure if this is valid C++ or not. But GCC (all versions I have
tried) and ICC compile this fine:


struct A;

extern "C" {
    struct A;
    typedef struct A A;
}

struct A;

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 13:17:06 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 13:17:06 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6924] New: llvm-config --cflags and --includedir
	broken
Message-ID: <bug-6924-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6924

           Summary: llvm-config --cflags and --includedir broken
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


--includedir only outputs the include dir for the sources:
/home/edwin/llvm-git/obj/../llvm/include

--cflags outputs both, but it also outputs a bunch of unneeded stuff (like -O2,
which can override an -O3 in another app, or _DEBUG which is not needed when
someone only uses the C bindings, etc.):
-I/home/edwin/llvm-git/obj/../llvm/include -I/home/edwin/llvm-git/obj/include 
-D_DEBUG -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS -O2
-fomit-frame-pointer -fPIC

Same for cppflags it outputs _DEBUG:
-I/home/edwin/llvm-git/obj/../llvm/include -I/home/edwin/llvm-git/obj/include 
-D_DEBUG -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS

So right now I had to hardcode the path to LLVM in mesa's llvmpipe to get it to
build (I didn't install LLVM as mesa expects).

llvm-config should be fixed to not output useless/bad flags, and only output
the strictly needed flags.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 13:20:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 13:20:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6925] New: "Aliasee not linked" assertion failed
Message-ID: <bug-6925-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6925

           Summary: "Aliasee not linked" assertion failed
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Linker
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: mail at exoticorn.de
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4733)
 --> (http://llvm.org/bugs/attachment.cgi?id=4733)
application.bc & ucioption.bc, triggering assertion when linked

Trying to compile the stockfish 1.7.1 chess engine to llvm .bc files (with
clang++) and linking them into one big .bc file, I ran into this assertion:

dranke at dranke-eee:~/coding/stockfish/tmp$ llvm-link -v application.bc
ucioption.bc 
Loading 'application.bc'
Loading 'ucioption.bc'
Linking in 'ucioption.bc'
llvm-link: LinkModules.cpp:761: bool LinkAlias(llvm::Module*, const
llvm::Module*, std::map<const llvm::Value*, llvm::Value*, std::less<const
llvm::Value*>, std::allocator<std::pair<const llvm::Value* const, llvm::Value*>
> >&, std::string*): Assertion `VMI != ValueMap.end() && "Aliasee not linked"'
failed.
0  llvm-link 0x081b0b08
Stack dump:
0.    Program arguments: llvm-link -v application.bc ucioption.bc 
Aborted

(I have narrowed this down to any file linked together with ucioption.bc.)

This happens with both llvm/clang trunk and the 2.7 release branch.

I'll attach both .bc files.

The original source files can be found at http://www.stockfishchess.com/.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 15:31:22 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 15:31:22 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6926] New: llvm build fails due to __func__
 conflicts in gcc 4.5.1 on Darwin 10.6.2
Message-ID: <bug-6926-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6926

           Summary: llvm build fails due to __func__ conflicts in gcc
                    4.5.1 on Darwin 10.6.2
           Product: new-bugs
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: vlad at demoninsight.com
                CC: llvmbugs at cs.uiuc.edu


Steps to reproduce:

1. build current gcc 4.5 branch (gcc 4.5.1 prerelease) on Darwin using gcc
prerequisite libraries as installed by Fink. (NOTE: this is NOT stock gcc 4.2.x
provided by Apple)

Using built-in specs.
COLLECT_GCC=gcc-4.5
COLLECT_LTO_WRAPPER=/sw/lib/gcc4.5.branch/libexec/gcc/x86_64-apple-darwin10/4.5.1/lto-wrapper
Target: x86_64-apple-darwin10
Configured with: ../src/configure --enable-checking=release
--prefix=/sw/lib/gcc4.5.branch --mandir=/sw/share/man --infodir=/sw/share/info
--enable-languages=c,c++,fortran --with-gmp=/sw --with-mpfr=/sw
--with-libiconv-prefix=/sw --with-ppl=/sw --with-cloog=/sw --with-mpc=/sw
--with-system-zlib --x-includes=/usr/X11R6/include --x-libraries=/usr/X11R6/lib
--disable-libjava-multilib --build=x86_64-apple-darwin10
--host=x86_64-apple-darwin10 --target=x86_64-apple-darwin10
Thread model: posix
gcc version 4.5.1 20100424 (prerelease) (GCC) 


2. do an SVN checkout of LLVM trunk (I reproduced this on Apr 24 2010) followed
by SVN checkout of clang trunk into <llvm dir>/tools/clang

3. export CC=... and export CXX=... as needed to set gcc 4.5 as the compiler to
be used by llvm build

4. export REQUIRES_RTTI=1

5. cd into <llvm prj dir> and do

./configure --prefix=<...> --{build,host,target}=x86_64-apple-darwin10
--enable-optimized --enable-jit --disable-expensive-checks
--enable-targets=host-only

6. subsequent make eventually fails as follows:

llvm[1]: Compiling Alarm.cpp for Release build
llvm[1]: Compiling Atomic.cpp for Release build
llvm[1]: Compiling Disassembler.cpp for Release build
llvm[1]: Compiling DynamicLibrary.cpp for Release build
...
llvm[2]: Compiling AsmMatcherEmitter.cpp for Release build
/var/folders/48/48Y65YT7FUutmsSTniSHTE+++TI/-Tmp-//ccLAMmeg.s:26321:FATAL:Symbol
__ZZN4llvm25SmallVectorTemplateCommonINS_9StringRefEEixEPvS3_jE8__func__
already defined.
make[2]: *** [/Users/.../LLVM/utils/TableGen/Release/AsmMatcherEmitter.o] Error
1
make[1]: *** [TableGen/.makeall] Error 2
make: *** [all] Error 1


NOTE: the same steps above produce valid llvm and clang libs and binaries if
used with

(a) Apple gcc 4.2.1

Using built-in specs.
Target: i686-apple-darwin10
Configured with: /var/tmp/gcc/gcc-5659~1/src/configure --disable-checking
--enable-werror --prefix=/usr --mandir=/share/man
--enable-languages=c,objc,c++,obj-c++
--program-transform-name=/^[cg][^.-]*$/s/$/-4.2/ --with-slibdir=/usr/lib
--build=i686-apple-darwin10 --program-prefix=i686-apple-darwin10-
--host=x86_64-apple-darwin10 --target=i686-apple-darwin10
--with-gxx-include-dir=/include/c++/4.2.1
Thread model: posix
gcc version 4.2.1 (Apple Inc. build 5659)

or

(b) gcc 4.4.2 as installed by Fink:

Using built-in specs.
Target: x86_64-apple-darwin10
Configured with: ../gcc-4.4.2/configure --prefix=/sw --prefix=/sw/lib/gcc4.4
--mandir=/sw/share/man --infodir=/sw/share/info
--enable-languages=c,c++,fortran,objc,java --with-gmp=/sw
--with-libiconv-prefix=/sw --with-ppl=/sw --with-cloog=/sw --with-system-zlib
--x-includes=/usr/X11R6/include --x-libraries=/usr/X11R6/lib
--disable-libjava-multilib --build=x86_64-apple-darwin10
--host=x86_64-apple-darwin10 --target=x86_64-apple-darwin10
Thread model: posix
gcc version 4.4.2 (GCC) 

Thanks,
Vlad

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 15:56:04 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 15:56:04 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6927] New: [NSAttributedString initWithString:]
 with a nil string will raise an exception in post-Leopard linked apps
Message-ID: <bug-6927-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6927

           Summary: [NSAttributedString initWithString:] with a nil string
                    will raise an exception in post-Leopard linked apps
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: alex at kempgen.de
                CC: llvmbugs at cs.uiuc.edu


Even though this is not explicitly stated in Apple's class reference, a console
log entry warns that calling [NSAttributedString initWithString:] with a nil
string will raise an exception in post-Leopard linked apps:

"-[NSConcreteAttributedString initWithString:] called with nil string argument.
This has undefined behavior and will raise an exception in post-Leopard linked
apps. This warning is displayed only once."

the static analyzer could assist finding cases where the string that's being
passed along is nil.


version = the one that shipped with xcode 3.2.2

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 18:56:50 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 18:56:50 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6928] New: Assertion with implicitly-generated copy
 constructor call with default arguments
Message-ID: <bug-6928-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6928

           Summary: Assertion with implicitly-generated copy constructor
                    call with default arguments
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


cove:clang dgregor$ cat t2.C
template <typename T>
struct X {
  X();
  X(const X&, bool = false);
};

struct Y { 
  X<int> x;
};

void f(Y);

void g() {
  Y y;
  f(y);
}
cove:clang dgregor$ clang++ t2.C
Assertion failed: (!hasUninstantiatedDefaultArg() && "Default argument is not
yet instantiated!"), function getDefaultArg, file
/Users/dgregor/Projects/llvm/tools/clang/lib/AST/Decl.cpp, line 845.

The problem here is that CodeGen is poking directly at the default arguments of
the ParmVarDecls in the copy constructor when it generates the copy-constructor
call, rather than letting Sema do that work.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 21:12:27 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 21:12:27 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6922] llvm demo page should feature clang demo
In-Reply-To: <bug-6922-206@http.llvm.org/bugs/>
References: <bug-6922-206@http.llvm.org/bugs/>
Message-ID: <20100425021228.0C398312836B@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6922

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-24 21:12:27 CDT ---
Yes, it should definitely be improved.  We're looking for someone web-savy to
do this, it shouldn't be that hard.  The code for the demo script is in the
llvm-www repository.

*** This bug has been marked as a duplicate of bug 1440 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sat Apr 24 23:00:14 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sat, 24 Apr 2010 23:00:14 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6929] New: possible wrong code bug
Message-ID: <bug-6929-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6929

           Summary: possible wrong code bug
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu, chenyang at cs.utah.edu


Created an attachment (id=4734)
 --> (http://llvm.org/bugs/attachment.cgi?id=4734)
failure-inducing C code

We think "1" is the right answer.

[regehr at gamow tmp409]$ clang -O small.c -o small
[regehr at gamow tmp409]$ ./small
1
[regehr at gamow tmp409]$ clang -Os small.c -o small
[regehr at gamow tmp409]$ ./small
0
[regehr at gamow tmp409]$ clang -v
clang version 1.5 (trunk 102243)
Target: x86_64-unknown-linux-gnu
Thread model: posix

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 00:31:07 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 00:31:07 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6921] clang: wrong alignment for .comm object
In-Reply-To: <bug-6921-206@http.llvm.org/bugs/>
References: <bug-6921-206@http.llvm.org/bugs/>
Message-ID: <20100425053107.1D54A312836D@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6921

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |clattner at apple.com
         Resolution|                            |FIXED

--- Comment #3 from Chris Lattner <clattner at apple.com> 2010-04-25 00:31:06 CDT ---
This should be fixed in r102300, please verify.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 00:52:21 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 00:52:21 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6913] LLVM Segmentation Fault on Compile
In-Reply-To: <bug-6913-206@http.llvm.org/bugs/>
References: <bug-6913-206@http.llvm.org/bugs/>
Message-ID: <20100425055221.CEBD2312836B@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6913

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
          Component|new bugs                    |llvm-gcc
         Resolution|                            |FIXED
            Product|new-bugs                    |tools

--- Comment #2 from Chris Lattner <clattner at apple.com> 2010-04-25 00:52:21 CDT ---
Ouch.  Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20100419/100354.html

Testcase here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20100419/100355.html

Thanks!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 03:39:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 03:39:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6930] New: clang: invalid input constraint "p" in
	asm
Message-ID: <bug-6930-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6930

           Summary: clang: invalid input constraint "p" in asm
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 4068


An error when compiling the Linux kernel:

In file included from arch/x86/kernel/asm-offsets.c:4:                          
In file included from arch/x86/kernel/asm-offsets_64.c:8:                       
In file included from include/linux/crypto.h:20:                                
In file included from
/home/edwin/builds/linux-2.6/arch/x86/include/asm/atomic.h:6:  
In file included from
/home/edwin/builds/linux-2.6/arch/x86/include/asm/processor.h:15:
/home/edwin/builds/linux-2.6/arch/x86/include/asm/current.h:14:9: error:
invalid input constraint 'p' in asm
        return percpu_read_stable(current_task);                                
               ^                                                                
In file included from arch/x86/kernel/asm-offsets.c:4:                          
In file included from arch/x86/kernel/asm-offsets_64.c:8:                       
In file included from include/linux/crypto.h:20:                                
In file included from
/home/edwin/builds/linux-2.6/arch/x86/include/asm/atomic.h:6:                   
In file included from
/home/edwin/builds/linux-2.6/arch/x86/include/asm/processor.h:15:               
In file included from
/home/edwin/builds/linux-2.6/arch/x86/include/asm/current.h:5:                  
/home/edwin/builds/linux-2.6/arch/x86/include/asm/percpu.h:203:34: note:
instantiated from:                                   
#define percpu_read_stable(var)         percpu_from_op("mov", var, "p"
(&(var)))                                              
                                        ^                                       
In file included from arch/x86/kernel/asm-offsets.c:4:                          
In file included from arch/x86/kernel/asm-offsets_64.c:8:                       
In file included from include/linux/crypto.h:20:                                
In file included from
/home/edwin/builds/linux-2.6/arch/x86/include/asm/atomic.h:6:                   
In file included from
/home/edwin/builds/linux-2.6/arch/x86/include/asm/processor.h:15:               
/home/edwin/builds/linux-2.6/arch/x86/include/asm/current.h:14:9: note:
instantiated from:                                    
        return percpu_read_stable(current_task);                                
               ^                                                  

The rationale for using "p" is described very nicely in the git commit:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=ed8d9adf357ec331603fa1049510399812cea7e5

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 04:21:19 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 04:21:19 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6931] New: Substandard performance of fibonacci
	example code
Message-ID: <bug-6931-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6931

           Summary: Substandard performance of fibonacci example code
           Product: new-bugs
           Version: 2.6
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: yuri at tsoft.com
                CC: llvmbugs at cs.uiuc.edu


I compiled this simple example:

#include <stdlib.h>
#include <stdio.h>

int fib(int AnArg) {
 if (AnArg <= 2) return (1);
 return (fib(AnArg-1)+fib(AnArg-2));
}

int main(int argc, char* argv[]) {
 int n = atoi(argv[1]);
 printf("fib(%i)=%i\n", n, fib(n));
}

with llvm-2.6 with gcc frontend.

Runtime with argument 45 is ~4.3sec.
Runtime of mainstream gcc compiled code is always below 4 sec (3.4-3.9sec).

This is a very basic example and should produce a very optimal code.

Intel(R) Core(TM) i7 CPU       @ 9200 @ 2.67GHz

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 04:39:54 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 04:39:54 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6934] New: clang: should have a -Wno-* flag to turn
 off "initializer overrides prior initialization of this subobject"
Message-ID: <bug-6934-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6934

           Summary: clang: should have a -Wno-* flag to turn off
                    "initializer overrides prior initialization of this
                    subobject"
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Driver
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 4068


Testcase from Linux kernel, there should be a way to turn off this warning,
since it serves no useful purpose here:
typedef void (*sys_call_ptr_t)(void);
extern void sys_ni_syscall(void);
#define __SYSCALL(nr, call) [nr] = (call),
#define __NR_syscall_max 299
extern void sys_read(void) ;
const sys_call_ptr_t sys_call_table[__NR_syscall_max+1] = {
        [0 ... __NR_syscall_max] = &sys_ni_syscall,
#define __NR_read                               0
__SYSCALL(__NR_read, sys_read)
};

$ clang p3.c
p3.c:9:1: warning: initializer overrides prior initialization of this subobject
__SYSCALL(__NR_read, sys_read)
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
p3.c:3:36: note: instantiated from:
#define __SYSCALL(nr, call) [nr] = (call),
                                   ^~~~~~
p3.c:7:29: note: previous initialization is here
        [0 ... __NR_syscall_max] = &sys_ni_syscall,
                                   ^~~~~~~~~~~~~~~
2 diagnostics generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 04:46:43 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 04:46:43 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6935] New: clang: -pedantic warnings by default and
 no way to turn off
Message-ID: <bug-6935-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6935

           Summary: clang: -pedantic warnings by default and no way to
                    turn off
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Driver
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 4068


Another testcase from Linux kernel. It warns. It shouldn't.

struct ext4_group_info {
        unsigned long   bb_state;
        int     bb_counters[];
};
struct sg {
                struct ext4_group_info info;
                int counters[16];
} sg;


p4.c:6:26: warning: field 'info' with variable sized type 'struct
ext4_group_info' not at the end of a struct or class is a
      GNU extension [-pedantic]
                struct ext4_group_info info;
                                       ^
1 diagnostic generated.

All these warn, none should since I didn't specify -pedantic:
clang p4.c -c
clang p4.c -std=gnu99 -c
clang p4.c -std=gnu89 -c
clang p4.c -no-pedantic -c

clang should have a way to turn off these warnings since they are useless for
the Linux kernel (we already know it uses GNU extensions, and there are no
reasons to change that).

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 04:56:55 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 04:56:55 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6936] New: clang -E -P -C -include doesn't work
Message-ID: <bug-6936-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6936

           Summary: clang -E -P -C -include doesn't work
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 4068


Command-line:
  clang -E -MD -MF arch/x86/kernel/acpi/realmode/.wakeup.lds.d  -nostdinc
-isystem /usr/local/bin/../lib/clang/1.0/include
-I/home/edwin/builds/linux-2.6/arch/x86/include -Iinclude  -include
include/generated/autoconf.h -D__KERNEL__   -P -C -P -C -Ux86_64 -D__ASSEMBLY__
-DLINKER_SCRIPT -o arch/x86/kernel/acpi/realmode/wakeup.lds
arch/x86/kernel/acpi/realmode/wakeup.lds.S


Output:
/*
 * Automatically generated C config: don't edit
 * Linux kernel version: 2.6.34-rc5
 * Sun Apr 25 12:54:09 2010
 */# 1 "<built-in>" 2

Which makes ld reject the linker script (# 1 is outside comment).

Small testcase:
$ touch empty.S
$ cat >foo.h <<EOF
/*
 * Automatically generated C config: don't edit
 * Linux kernel version: 2.6.34-rc5
 * Sun Apr 25 12:54:09 2010
 */
EOF

$ clang -E -P -C -include foo.h empty.S
/*
 * Automatically generated C config: don't edit
 * Linux kernel version: 2.6.34-rc5
 * Sun Apr 25 12:54:09 2010
 */# 1 "<built-in>" 2

$ gcc -E -P -C -include foo.h empty.S
/*
 * Automatically generated C config: don't edit
 * Linux kernel version: 2.6.34-rc5
 * Sun Apr 25 12:54:09 2010
 */

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 05:19:27 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 05:19:27 -0500 (CDT)
Subject: [LLVMbugs] [Bug 4628] clang doesn't support '-mcmodel=kernel'
In-Reply-To: <bug-4628-206@http.llvm.org/bugs/>
References: <bug-4628-206@http.llvm.org/bugs/>
Message-ID: <20100425101927.8E2E3312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=4628

T?r?k Edwin <edwintorok at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #19 from T?r?k Edwin <edwintorok at gmail.com> 2010-04-25 05:19:26 CDT ---
I don't see the linker warnings anymore.

(In reply to comment #18)
> Edwin is this still broken?  I thought this was done.

Sorry, I must've missed your question, and only noticed now when going through
the kernel bugs.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 07:36:17 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 07:36:17 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6834] clang installs empty directories like
 usr/include/CMakeFiles
In-Reply-To: <bug-6834-206@http.llvm.org/bugs/>
References: <bug-6834-206@http.llvm.org/bugs/>
Message-ID: <20100425123617.21396312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6834

Markus Rothe <markus at unixforces.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME

--- Comment #1 from Markus Rothe <markus at unixforces.net> 2010-04-25 07:36:16 CDT ---
I cannot reproduce this as of today (yesterday's build had empty directories).

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 09:46:15 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 09:46:15 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6916] [meta] clang 2.7 tests on various applications
In-Reply-To: <bug-6916-206@http.llvm.org/bugs/>
References: <bug-6916-206@http.llvm.org/bugs/>
Message-ID: <20100425144615.69384312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6916

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |LATER

--- Comment #4 from Chris Lattner <clattner at apple.com> 2010-04-25 09:46:15 CDT ---
Edwin, you can do this by doing a search for all bugs in the clang component. 
You can filter by subcomponent and severity as well.  I believe that all
miscompiles are marked as 'normal' severity.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 09:55:34 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 09:55:34 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6938] New: clang:crashes when compiling nmi.c
Message-ID: <bug-6938-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6938

           Summary: clang:crashes when compiling nmi.c
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 4068


Created an attachment (id=4735)
 --> (http://llvm.org/bugs/attachment.cgi?id=4735)
preprocessed source

'clang -Os -c nmi.i' crashes, see below.
clang -Os -c nmi.i -w works.

clang -c nmi.i crashes too. Looks like something when it wants to emit
warnings/analyze code for warnings.

clang: /home/edwin/llvm-git/clang/lib/Analysis/CFG.cpp:331:
bool<unnamed>::CFGBuilder::FinishBlock(clang::CFGBlock*): Assertion `B' failed. 
0  clang           0x000000000118753f                                           
1  clang           0x0000000001187d4d                                           
2  libpthread.so.0 0x0000003a06a0ea60                                           
3  libc.so.6       0x0000003a05e31f05 gsignal + 53
4  libc.so.6       0x0000003a05e34d10 abort + 384
5  libc.so.6       0x0000003a05e2afaa __assert_fail + 234
6  clang           0x0000000000835bb6
7  clang           0x000000000083ec99
8  clang           0x000000000083adb7
9  clang           0x000000000083e10c
10 clang           0x000000000083acc5
11 clang           0x000000000083c174
12 clang           0x000000000083a86d
13 clang           0x000000000083cec0
14 clang           0x0000000000831bb4
15 clang           0x000000000060f0c2
16 clang           0x000000000060fc84
17 clang           0x00000000006357ed
18 clang           0x0000000000635ef7
19 clang           0x000000000090e284
20 clang           0x0000000000919867
21 clang           0x000000000091eee7
22 clang           0x00000000009181f5
23 clang           0x0000000000918688
24 clang           0x000000000091a3c9
25 clang           0x000000000091a89a
26 clang           0x00000000005fc8eb
27 clang           0x0000000000415d99
28 clang           0x00000000004095c9
29 clang           0x000000000040bf8b main + 1675
30 libc.so.6       0x0000003a05e1eabd __libc_start_main + 253
31 clang           0x0000000000407099
Stack dump:
0.      Program arguments: /usr/local/bin/clang -cc1 -triple
x86_64-unknown-linux-gnu -S -disable-free -main-file-name nmi.i
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -resource-dir /usr/local/lib/clang/1.1 -Os
-fmessage-length 126 -fgnu-runtime -fdiagnostics-show-option
-fcolor-diagnostics -o /tmp/cc-MyYG0K.s -x cpp-output nmi.i
1.      arch/x86/kernel/apic/nmi.c:475:1: current parser token 'static'
2.      arch/x86/kernel/apic/nmi.c:393:1: parsing function body
'nmi_watchdog_tick'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 14:14:10 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 14:14:10 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6935] clang: -pedantic warnings by default and no
 way to turn off
In-Reply-To: <bug-6935-206@http.llvm.org/bugs/>
References: <bug-6935-206@http.llvm.org/bugs/>
Message-ID: <20100425191410.732D2312836C@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6935

T?r?k Edwin <edwintorok at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from T?r?k Edwin <edwintorok at gmail.com> 2010-04-25 14:14:10 CDT ---
Lets close it then, I was testing 2.7

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Sun Apr 25 16:21:47 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Sun, 25 Apr 2010 16:21:47 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6905] Conservative Unwind Pass
In-Reply-To: <bug-6905-206@http.llvm.org/bugs/>
References: <bug-6905-206@http.llvm.org/bugs/>
Message-ID: <20100425212147.C77D43128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6905

Nick Lewycky <nicholas at mxc.ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #5 from Nick Lewycky <nicholas at mxc.ca> 2010-04-25 16:21:47 CDT ---
Sounds good!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 07:15:04 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 07:15:04 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6939] New: Compilation fails for clang: missing
 function EmitAnyExprToMem()
Message-ID: <bug-6939-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6939

           Summary: Compilation fails for clang: missing function
                    EmitAnyExprToMem()
           Product: clang
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: reeuwijk at few.vu.nl
                CC: llvmbugs at cs.uiuc.edu


When compiling clang, I get the following error message:

llvm-trunk/tools/clang/lib/CodeGen/CGException.cpp: In function 'void
EmitAnyExprToExn(clang::CodeGen::CodeGenFunction&, const clang::Expr*,
llvm::Value*)':
llvm-trunk/tools/clang/lib/CodeGen/CGException.cpp:179: error: 'class
clang::CodeGen::CodeGenFunction' has no member named 'EmitAnyExprToMem'

Grepping around in the trees shows that there is indeed no such function in the
current source tree, and there is no function that comes anywhere close in
name.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 09:58:32 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 09:58:32 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6940] New: SCCP miscompiles sitofp(undef)
Message-ID: <bug-6940-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6940

           Summary: SCCP miscompiles sitofp(undef)
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


On this testcase:

define double @foo() {
  %t = sitofp i32 undef to double
  ret double %t
}

SCCP changes the sitofp operand to undef. This is over-aggressive, since there
exist double values which sitofp cannot produce. For comparison, instcombine on
this testcase changes the operand to 0.0.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 10:32:41 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 10:32:41 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6941] New: miscompilation - partial result
	optimized away
Message-ID: <bug-6941-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6941

           Summary: miscompilation - partial result optimized away
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


on amd64:

pes ~$ cat buf.c
typedef unsigned int __attribute__((__mode__(__DI__))) u_int64_t;

unsigned int
snd_xbytes(unsigned int v, unsigned int from, unsigned int to)
{
 unsigned int w, x, y;

 x = from;
 y = to;

 while (y != 0) {
    w = x % y;
    x = y;
    y = w;
 }
                            // [1]
 from /= x;
 to /= x;

 // printf("v = %u, from = %u, to = %u\n", v, from, to); // [2]

 return (unsigned int)(((u_int64_t)v * to) / from);
}

int main(int argc, char **argv) {
   printf("%u\n", snd_xbytes(atoi(argv[1]), atoi(argv[2]), atoi(argv[3])));
}


pes ~$ clang -O1 buf.c -o a.out.clang && ./a.out.clang 25050482 2094268276
1210547630

    ...warnings...

1

pes ~$ gcc -O1 buf.c -o a.out.gcc && ./a.out.gcc 25050482 2094268276 1210547630

    ...warnings...

14479903

the problem is that clang -O1 compiles the code starting at [1] to:

.LBB0_5:
        movl    %ecx, %eax
        xorl    %edx, %edx
        divl    %r8d
        movl    %edi, %ecx
        imulq   %rcx, %rax        // [3]
        movl    %esi, %eax        // [4]
        xorl    %edx, %edx
        divl    %r8d
        movl    %eax, %ecx
        xorl    %edx, %edx
        divq    %rcx
        popq    %rbp
        ret

the imulq at [3] does basically %rax = %rax * %rcx but the movl at [4] rewrites
the value in %rax which is then never used. Clang -O0 does not exhibit this 
behaviour. When I uncomment the printf at [2] the miscompilation is gone.

I guess this is backend issue as opt -O2 still shows the multiplication as live

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 11:11:44 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 11:11:44 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6906] VFP instructions miscompiled
In-Reply-To: <bug-6906-206@http.llvm.org/bugs/>
References: <bug-6906-206@http.llvm.org/bugs/>
Message-ID: <20100426161144.417A8312800A@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6906

Bob Wilson <bob.wilson at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #4 from Bob Wilson <bob.wilson at apple.com> 2010-04-26 11:11:43 CDT ---
The patch will be useful as a workaround until a more recent cctools is
released, but after doing all the work to switch to unified syntax, we don't
want to go back.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 12:31:43 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 12:31:43 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6939] Compilation fails for clang: missing function
 EmitAnyExprToMem()
In-Reply-To: <bug-6939-206@http.llvm.org/bugs/>
References: <bug-6939-206@http.llvm.org/bugs/>
Message-ID: <20100426173143.1B2B73128042@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6939

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-26 12:31:42 CDT ---
Clang builds fine for me, perhaps you have an out of date/inconsistent source
tree or something?

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 13:21:35 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 13:21:35 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6940] SCCP miscompiles sitofp(undef)
In-Reply-To: <bug-6940-206@http.llvm.org/bugs/>
References: <bug-6940-206@http.llvm.org/bugs/>
Message-ID: <20100426182135.87BE1312800A@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6940

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-26 13:21:35 CDT ---
Fixed in r102358, thanks.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 13:37:53 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 13:37:53 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6921] clang: wrong alignment for .comm object
In-Reply-To: <bug-6921-206@http.llvm.org/bugs/>
References: <bug-6921-206@http.llvm.org/bugs/>
Message-ID: <20100426183753.AEBDC3128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6921

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|CLOSED                      |REOPENED
         Resolution|FIXED                       |

--- Comment #5 from Chris Lattner <clattner at apple.com> 2010-04-26 13:37:53 CDT ---
My fix broke a bunch of objc code and had to be reverted. Will fix it the other
way.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 13:47:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 13:47:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6921] clang: wrong alignment for .comm object
In-Reply-To: <bug-6921-206@http.llvm.org/bugs/>
References: <bug-6921-206@http.llvm.org/bugs/>
Message-ID: <20100426184718.8BA703128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6921

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #7 from Chris Lattner <clattner at apple.com> 2010-04-26 13:47:18 CDT ---
This should be refixed in r102365, please verify again. thanks :)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 14:54:46 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 14:54:46 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6942] New: __vector bool int
Message-ID: <bug-6942-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6942

           Summary: __vector bool int
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: John.Thompson.JTSoftware at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Apparently types like the following need to be supported for AltiVec:

__vector bool char
__vector bool short
__vector bool int

The error text:

(file): error: cannot combine with previous '_Bool' declaration specifier

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 16:51:19 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 16:51:19 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6914] Instruction does not dominate all uses
In-Reply-To: <bug-6914-206@http.llvm.org/bugs/>
References: <bug-6914-206@http.llvm.org/bugs/>
Message-ID: <20100426215119.BB1EE3128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6914

Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Dan Gohman <gohman at apple.com> 2010-04-26 16:51:19 CDT ---
Fixed in 102389.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 17:08:33 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 17:08:33 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6936] clang -E -P -C -include doesn't work
In-Reply-To: <bug-6936-206@http.llvm.org/bugs/>
References: <bug-6936-206@http.llvm.org/bugs/>
Message-ID: <20100426220833.00C2A3128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6936

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-26 17:08:32 CDT ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20100426/029786.html

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 17:22:00 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 17:22:00 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6943] New: Crash on invalid code
Message-ID: <bug-6943-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6943

           Summary: Crash on invalid code
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: crash-on-invalid
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com




-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 18:03:57 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 18:03:57 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6944] New: Calling a non-sret function with an sret
 convention corrupts the stack
Message-ID: <bug-6944-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6944

           Summary: Calling a non-sret function with an sret convention
                    corrupts the stack
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: csdavec at swan.ac.uk
                CC: llvmbugs at cs.uiuc.edu


[theraven at EtoileVM ~/tmp]$ cat fpret.c
typedef struct
{
    double x,y;
} rect;

typedef rect (*aRet)(void);

int foo(void)
{
    return 0;
}

int main(void)
{
    rect a;
    aRet f = (aRet)foo;
    a = f();
    return 0;
}
$ clang fpret.c && ./a.out
Segmentation fault: 11 (core dumped)
$ gcc fpret.c && ./a.out

This is a reduced test case - the problem is exhibited most commonly when
calling an Objective-C method that returns a structure on nil.  The code
generated by GCC returns some undefined value, as does the code generated by
clang, however the code generated by clang crashes (somewhat confusingly) when
the calling function returns.  In this case, when main() returns.  

The assembly produced by clang is very different from that produced by gcc. 
GCC produces this:

main:
    leal    4(%esp), %ecx
    andl    $-16, %esp
    pushl   -4(%ecx)
    pushl   %ebp
    movl    %esp, %ebp
    pushl   %ecx
    subl    $52, %esp
    movl    $foo, -12(%ebp)
    leal    -32(%ebp), %eax
    movl    %eax, (%esp)
    movl    -12(%ebp), %eax
    call    *%eax
    subl    $4, %esp
    movl    $0, %eax
    movl    -4(%ebp), %ecx
    leave
    leal    -4(%ecx), %esp
    ret


Clang produces this:

main:
    pushl   %ebp
    movl    %esp, %ebp
    subl    $32, %esp
    movl    $0, -4(%ebp)
    leal    foo, %eax
    movl    %eax, -28(%ebp)
    movl    -28(%ebp), %eax
    movl    %esp, %ecx
    leal    -24(%ebp), %edx
    movl    %edx, (%ecx)
    call    *%eax
    subl    $4, %esp
    movl    $0, -4(%ebp)
    movl    -4(%ebp), %eax
    addl    $32, %esp
    popl    %ebp
    ret

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 18:24:30 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 18:24:30 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6945] New: Failed template argument deduction
 should mention argument types
Message-ID: <bug-6945-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6945

           Summary: Failed template argument deduction should mention
                    argument types
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: quality-of-implementation
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


And, ideally, say which ones failed to match the parameters.

As of r102340, std::max(different, types) produces an error message that
doesn't help the user figure out what they did wrong.


$ cat test.cc
#include <algorithm>

long foo(int arg1, long arg2) {
    return std::max(arg1, arg2);
}

$ clang -c test.cc
test.cc:4:12: error: no matching function for call to 'max'
    return std::max(arg1, arg2);
           ^~~~~~~~
In file included from test.cc:1:
In file included from /usr/include/c++/4.0.0/algorithm:64:
/usr/include/c++/4.0.0/bits/stl_algobase.h:206:5: note: candidate template
ignored: failed template argument deduction
    max(const _Tp& __a, const _Tp& __b)
    ^
/usr/include/c++/4.0.0/bits/stl_algobase.h:248:5: note: candidate template
ignored: failed template argument deduction
    max(const _Tp& __a, const _Tp& __b, _Compare __comp)
    ^
1 error generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 18:50:38 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 18:50:38 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6694] lowerinvoke breaks alloca instructions if
 they do not occur in the entry basic block
In-Reply-To: <bug-6694-206@http.llvm.org/bugs/>
References: <bug-6694-206@http.llvm.org/bugs/>
Message-ID: <20100426235038.957BA3128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6694

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #7 from Chris Lattner <clattner at apple.com> 2010-04-26 18:50:38 CDT ---
Applied here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20100426/100413.html

In the future, please don't merge two patches like this.  It would be better to
apply them as two separate patches so that they are tracked independently in
revision control.

Thanks for the patch!

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Mon Apr 26 19:10:04 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Mon, 26 Apr 2010 19:10:04 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6946] New: Clang should suggest a header to include
 for more standard names
Message-ID: <bug-6946-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6946

           Summary: Clang should suggest a header to include for more
                    standard names
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: quality-of-implementation
          Severity: enhancement
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Right now, clang suggests to include <stdio.h> when printf is used but not
defined in C, but not in C++. Clang also fails to suggest <stddef.h> for size_t
and ptrdiff_t in either language.

Clang should probably have a table of all the names commonly used from C, C++,
Posix, and Win32, and suggest appropriate headers for them when they're missing
a declaration. This is all relatively low priority.


$ cat test.cc
void foo() {
    size_t s;
    ptrdiff_t p;
    printf("Hello World\n");
}

$ clang -x c -c test.cc
test.cc:2:5: error: use of undeclared identifier 'size_t'; did you mean
'sizeof'?
    size_t s;
    ^
test.cc:3:5: error: use of undeclared identifier 'ptrdiff_t'
    ptrdiff_t p;
    ^
test.cc:4:5: warning: implicitly declaring C library function 'printf' with
type 'int (char const *, ...)'
    printf("Hello World\n");
    ^
test.cc:4:5: note: please include the header <stdio.h> or explicitly provide a
declaration for 'printf'
1 warning and 2 errors generated.

$ clang -x c++ -c test.cc
test.cc:2:5: error: use of undeclared identifier 'size_t'; did you mean
'sizeof'?
    size_t s;
    ^
test.cc:3:5: error: use of undeclared identifier 'ptrdiff_t'
    ptrdiff_t p;
    ^
test.cc:4:5: error: use of undeclared identifier 'printf'
    printf("Hello World\n");
    ^
3 errors generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 03:20:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 03:20:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6947] New: Assertion 'Passed null for Decl param'
Message-ID: <bug-6947-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6947

           Summary: Assertion 'Passed null for Decl param'
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chris at bubblescope.net
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


The following code causes a 'Passed null for Decl param' assertion. It is
derived from boost code. This code is (I believe) minimal. The 'private:' and
the unused method 'make_processor_holder' are necessary to trigger it.

template< class Scheduler > 
struct processor_container {
  int create_processor( )      
  {
    typedef void ( processor_container::*impl_fun_ptr )( );
    impl_fun_ptr pImpl = &processor_container::template
create_processor_impl1<int>;
  }
 private:                  
    int make_processor_holder() {
  }
  template< class Processor>                  
  void create_processor_impl1( )                 
  {
  }
};

char GetKey() {
  processor_container<int> pc;
  pc.create_processor();
}

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 03:40:07 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 03:40:07 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6897] clang has a hardcoded version 1.0 (which
 should be 1.1 instead)
In-Reply-To: <bug-6897-206@http.llvm.org/bugs/>
References: <bug-6897-206@http.llvm.org/bugs/>
Message-ID: <20100427084007.779093128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6897

Tanya Lattner <tonic at nondot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #4 from Tanya Lattner <tonic at nondot.org> 2010-04-27 03:40:06 CDT ---
Actually, I just built this in linux and get:
llvm:/opt/clang-releases/llvm-2.7/bin> ./clang -v
clang version 1.1 (branches/release_27)
Target: x86_64-unknown-linux-gnu
Thread model: posix

I think this is fine.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 03:40:28 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 03:40:28 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6948] New: Assertion 'queried property of class
 with no definition' on invalid code
Message-ID: <bug-6948-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6948

           Summary: Assertion 'queried property of class with no
                    definition' on invalid code
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chris at bubblescope.net
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


The following invalid code makes clang++ segfault with:

clang:
/home/caj/llvm/tools/clang/lib/CodeGen/../../include/clang/AST/DeclCXX.h:317:
const clang::CXXRecordDecl::DefinitionData& clang::CXXRecordDecl::data() const:
Assertion `DefinitionData && "queried property of class with no definition"'
failed.



 template<typename _CharT>     class basic_string;

       int main() {
          basic_string<char> str (read_from_file());
       }

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 05:24:15 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 05:24:15 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6949] New: --enable-debug-runtime configure option
 doesn't enable '-g' flag in compile
Message-ID: <bug-6949-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6949

           Summary: --enable-debug-runtime configure option doesn't enable
                    '-g' flag in compile
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: yuri at tsoft.com
                CC: llvmbugs at cs.uiuc.edu


The only thing that --enable-debug-runtime is not stripping. To be useful, it
also needs to enable debug info.

Tried in llvm-2.7. But 2.7 isn't in the version list, so reporting against the
trunk.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 06:03:57 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 06:03:57 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6950] New: scan-build for configure script that
 uses AC_LANG_WERROR does not work as expected
Message-ID: <bug-6950-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6950

           Summary: scan-build for configure script that uses
                    AC_LANG_WERROR does not work as expected
           Product: clang
           Version: 2.6
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: simon at comsys.ntu-kpi.kiev.ua
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4741)
 --> (http://llvm.org/bugs/attachment.cgi?id=4741)
Configure script that uses AC_LANG_WERROR

clang-2.6_3, llvm-2.6_2, autoconf-2.62 on FreeBSD 8.0-STABLE, everything
installed from FreeBSD Ports.

There is the configure script that uses AC_LANG_WERROR to verify whether the C
compiler shows any warnings about code inside AC_COMPILE_IFELSE(). Actually it
checks whether a compiler understands attributes like "noreturn", "format",
"unused", etc.

This script correctly works with GCC, but when it runs under scan-build control
all checks after AC_LANG_WERROR failed. As I understand these checks failed
since scan-build outputs "ANALYZE: ..." strings for the input file.

There is a simple configure.ac in the attachment and this is corresponding
config.log for it:

configure:3389: checking attribute "noreturn"
configure:3416: /usr/local/bin/ccc-analyzer -c -g -O2  conftest.c >&5
ANALYZE: conftest.c bar
configure:3423: $? = 0
configure: failed program was:
| /* confdefs.h.  */
| #define PACKAGE_NAME "test"
| #define PACKAGE_TARNAME "test"
| #define PACKAGE_VERSION "1"
| #define PACKAGE_STRING "test 1"
| #define PACKAGE_BUGREPORT "me at home"
| #define PACKAGE "test"
| #define VERSION "1"
| /* end confdefs.h.  */
| 
| #include <stdlib.h>
| 
| void bar(void) __attribute__ ((noreturn));
| 
| void
| bar(void)
| {
|     exit(0);
| }
| 
configure:3438: result: no

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 08:02:30 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 08:02:30 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6951] New: bugpoint crashes when debugging a
	miscompilation
Message-ID: <bug-6951-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6951

           Summary: bugpoint crashes when debugging a miscompilation
           Product: new-bugs
           Version: trunk
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: kalle.raiskila at nokia.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4742)
 --> (http://llvm.org/bugs/attachment.cgi?id=4742)
bugpoint output

Bugpoint crashes when debugging a llc-miscompilation. The crash appeared in
LLVM r99160.

Crash is verified to happen only on CellBE-linux (Linux
2.6.25-2-powerpc64-smp), when targeting the SPU. (This might not be platform
specific: SPU backend just seems to have a lot of miscompilations.)
LLVM is compiled with 'gcc (Ubuntu 4.3.2-1ubuntu12) 4.3.2'

bugpoint command line used: 
'bugpoint bug.ll -run-llc --tool-args --march=cellspu'
Full output and failing testcase as attachment. The slightly non-default gcc
used by bugpoint is a link to /usr/bin/spu-gcc.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 09:30:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 09:30:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6952] New: Clang allows redefinition of
 namespace-scope function via friend instantiation
Message-ID: <bug-6952-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6952

           Summary: Clang allows redefinition of namespace-scope function
                    via friend instantiation
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: accepts-invalid
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
            Blocks: 6023


Created an attachment (id=4744)
 --> (http://llvm.org/bugs/attachment.cgi?id=4744)
Test case

We do not diagnose [temp.friend]p4, where a friend definition inside a class
template that is instantiated twice is a redefinition error. Instead, we just
see that the friend definition was instantiated and skip it.

The attached test case attempts to illustrate the intended behavior for C++98
and C++0x (there were changes in [temp.friend]p4). Clang type-checks it without
complain.

Boost.Units relies on this check for some of its static type safety.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 09:47:07 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 09:47:07 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6953] New: __builtin_types_compatible_p does not
 take in account noreturn attributes
Message-ID: <bug-6953-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6953

           Summary: __builtin_types_compatible_p does not take in account
                    noreturn attributes
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: abramobagnara at tin.it
                CC: llvmbugs at cs.uiuc.edu


This code compiles in gcc, but not in clang due to different result.

The two types are considered incompatible by gcc and compatible by clang.

typedef void (*f1)();
typedef void (*f2)() __attribute__ ((noreturn));

int p() {
  int a[1 - __builtin_types_compatible_p(f1, f2) * 2];
}

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 09:53:19 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 09:53:19 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6954] New: Instantiating friend declaration injects
 friend into namespace scope
Message-ID: <bug-6954-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6954

           Summary: Instantiating friend declaration injects friend into
                    namespace scope
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: accepts-invalid
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


If a class template contains a friend function declaration (whose type does not
depend on the class template's template parameters), and we instantiate that
class template twice, the friend function becomes visible when it shouldn't:

cove:clang dgregor$ cat t.cpp
template<typename T>
struct X {
  operator int();
  friend void f(int x);
};

int array0[sizeof(X<int>)]; 
int array1[sizeof(X<float>)];

void g() {
  f(5); // expected-error{{undeclared identifier 'f'}}
}
cove:clang dgregor$ clang++ -fsyntax-only t.cpp 
cove:clang dgregor$ g++ -fsyntax-only t.cpp 
t.cpp: In function ?void g()?:
t.cpp:11: error: ?f? was not declared in this scope

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 10:18:38 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 10:18:38 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6955] New: clang: crashes when compiling
 arch/x86/kernel/cpu/common.c from linux kernel
Message-ID: <bug-6955-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6955

           Summary: clang: crashes when compiling
                    arch/x86/kernel/cpu/common.c from linux kernel
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: stefan at datenfreihafen.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4747)
 --> (http://llvm.org/bugs/attachment.cgi?id=4747)
preprocessed code

Latest trunk of clang and latest linus HEAD gives me the following:

  CC      arch/x86/kernel/cpu/common.o
clang: CGExprConstant.cpp:82:
bool<unnamed>::ConstStructBuilder::AppendField(const clang::FieldDecl*,
uint64_t, llvm::Constant*): Assertion `NextFieldOffsetInBytes <=
FieldOffsetInBytes && "Field offset mismatch!"' failed.
0  clang           0x000000000127458f
1  clang           0x0000000001276482
stefan at excalibur:~/Projekte/kernel/linux-2.6$ make
  CHK     include/linux/version.h
  CHK     include/generated/utsrelease.h
  CALL    scripts/checksyscalls.sh
  CHK     include/generated/compile.h
  CC      arch/x86/kernel/cpu/common.o
clang: CGExprConstant.cpp:82:
bool<unnamed>::ConstStructBuilder::AppendField(const clang::FieldDecl*,
uint64_t, llvm::Constant*): Assertion `NextFieldOffsetInBytes <=
FieldOffsetInBytes && "Field offset mismatch!"' failed.
0  clang           0x000000000127458f
1  clang           0x0000000001276482
2  libpthread.so.0 0x00002ae1fc11ba60
3  libc.so.6       0x00002ae1fcd0bf05 gsignal + 53
4  libc.so.6       0x00002ae1fcd0ed10 abort + 384
5  libc.so.6       0x00002ae1fcd04faa __assert_fail + 234
6  clang           0x00000000005806ee
7  clang           0x000000000058130d
8  clang           0x000000000057e39b
9  clang           0x000000000057f467
10 clang           0x000000000058130d
11 clang           0x000000000057e39b
12 clang           0x000000000057f467
13 clang           0x000000000058130d
14 clang           0x000000000057e39b
15 clang           0x0000000000581199
16 clang           0x000000000057e39b
17 clang           0x000000000057f467
18 clang           0x000000000058130d
19 clang           0x000000000057e39b
20 clang           0x000000000051d61b
21 clang           0x000000000051fb44
22 clang           0x0000000000520287
23 clang           0x00000000005205b7
24 clang           0x0000000000517381
25 clang           0x0000000000412884
26 clang           0x0000000000612665
27 clang           0x0000000000418d11
28 clang           0x0000000000409a16
29 clang           0x000000000040cb73 main + 2371
30 libc.so.6       0x00002ae1fccf8abd __libc_start_main + 253
31 clang           0x0000000000407399
Stack dump:
0.      Program arguments: /usr/local/bin/clang -cc1 -triple
x86_64-unknown-linux-gnu -S -disable-free -main-file-name common.c
-mrelocation-model static -mconstructor-aliases -mcode-model kernel
-disable-red-zone -target-cpu x86-64 -target-feature -sse -target-feature -mmx
-target-feature -sse2 -target-feature -3dnow -nostdinc -resource-dir
/usr/local/lib/clang/1.5 -dependency-file arch/x86/kernel/cpu/.common.o.d -MT
arch/x86/kernel/cpu/common.o -sys-header-deps -isystem include -include
include/generated/autoconf.h -D__KERNEL__ -DCONFIG_AS_CFI=1
-DCONFIG_AS_CFI_SIGNAL_FRAME=1 -DKBUILD_STR(s)=#s
-DKBUILD_BASENAME=KBUILD_STR(common) -DKBUILD_MODNAME=KBUILD_STR(common)
-I/home/stefan/Projekte/kernel/linux-2.6/arch/x86/include -Iinclude -O2 -Wall
-Wundef -Wstrict-prototypes -Wno-trigraphs
-Werror-implicit-function-declaration -Wno-format-security -Wno-unused-value
-Wno-sign-compare -Wdeclaration-after-statement -Wno-pointer-sign -ferror-limit
19 -ftemplate-backtrace-limit 10 -fmessage-length 106 -fgnu-runtime -fno-common
-fdiagnostics-show-option -fcolor-diagnostics -o /tmp/cc-nckBgP.s -x c
arch/x86/kernel/cpu/common.c 
1.      arch/x86/kernel/cpu/common.c:138:36: current parser token ';'
2.      arch/x86/kernel/cpu/common.c:86:1
<Spelling=arch/x86/kernel/cpu/common.c:86:46>: LLVM IR generation of
declaration 'gdt_page'
3.      arch/x86/kernel/cpu/common.c:86:1
<Spelling=arch/x86/kernel/cpu/common.c:86:46>: Generating code for declaration
'gdt_page'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)
make[3]: *** [arch/x86/kernel/cpu/common.o] Error 250
make[2]: *** [arch/x86/kernel/cpu] Error 2
make[1]: *** [arch/x86/kernel] Error 2
make: *** [arch/x86] Error 2

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 11:10:49 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 11:10:49 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6947] Assertion 'Passed null for Decl param'
In-Reply-To: <bug-6947-206@http.llvm.org/bugs/>
References: <bug-6947-206@http.llvm.org/bugs/>
Message-ID: <20100427161049.F20C43128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6947

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-27 11:10:49 CDT ---
Fixed in r102434. Which Boost library did this affect?

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 11:20:27 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 11:20:27 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6948] Assertion 'queried property of class with no
 definition' on invalid code
In-Reply-To: <bug-6948-206@http.llvm.org/bugs/>
References: <bug-6948-206@http.llvm.org/bugs/>
Message-ID: <20100427162027.142393128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6948

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-27 11:20:26 CDT ---
Fixed in r102436.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 11:27:19 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 11:27:19 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6923] error: definition of type 'A' conflicts with
 typedef of the same name
In-Reply-To: <bug-6923-206@http.llvm.org/bugs/>
References: <bug-6923-206@http.llvm.org/bugs/>
Message-ID: <20100427162719.781C73128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6923

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Douglas Gregor <dgregor at apple.com> 2010-04-27 11:27:19 CDT ---
Fixed in r102437

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 13:12:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 13:12:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6266] clang should compile libstdc++
In-Reply-To: <bug-6266-206@http.llvm.org/bugs/>
References: <bug-6266-206@http.llvm.org/bugs/>
Message-ID: <20100427181251.220AE3128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6266

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Douglas Gregor <dgregor at apple.com> 2010-04-27 13:12:50 CDT ---
Clang does compile libstdc++.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 16:07:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 16:07:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6953] __builtin_types_compatible_p does not take in
 account noreturn attributes
In-Reply-To: <bug-6953-206@http.llvm.org/bugs/>
References: <bug-6953-206@http.llvm.org/bugs/>
Message-ID: <20100427210756.646603128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6953

Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |sharparrow1 at yahoo.com
         Resolution|                            |FIXED

--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com> 2010-04-27 16:07:56 CDT ---
Fixed in r102464.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 16:25:19 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 16:25:19 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6956] New: Array operations
Message-ID: <bug-6956-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6956

           Summary: Array operations
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Keywords: new-feature
          Severity: enhancement
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bearophile at mailas.com
                CC: llvmbugs at cs.uiuc.edu


This is a low-priority enhancement request. Seeing how many things are done by
LLVM I am asking if it can be positive for LLVM to perform something already
done by D language front-end.

In all D1 and D2 language compilers there is native support array operations,
like:

c[] = a[] * b[];
c[] *= a[] * 20 + b[] * 6;

Where a, b, c are arrays of the same length, of a built-in type like int, real,
double, unsigned byte, etc.

Both dynamic arrays (allocated on the heap) and static arrays (allocated on the
stack, with a length known at compile-time) can be used with the same syntax.

In D static arrays are given to functions by value (unlike C language). (This
can be useful for very short arrays, that can be kept in SSE registers).

Currently such vector ops are performed by the D front-end, for example the LDC
compiler that uses LLVM as back-end. But such operations are currently not
efficient.

Currently some arrays ops are implemented with functions that contain
handwritten asm, for example relative to arrays of signed integers:
http://www.dsource.org/projects/phobos/browser/branches/phobos-1.x/phobos/internal/arrayint.d

Or arrays of doubles:
http://www.dsource.org/projects/phobos/browser/branches/phobos-1.x/phobos/internal/arraydouble.d?rev=811

While the other ones are translated to normal loops.

They are not efficient because they are always translated to a function call,
even when the product of two static arrays of 4 floats can be done with a
single SSE instruction.

This feature is not the same thing as loop vectorization, because there is no
need for LLVM to analyse loops and infer things, the front-end can just convert
the D vector ops to defined LLVM instructions (both in the dynamic and static
case. In the static case LLVM can produce better code).

Then LLVM can compile the ops well, using SSE or future wider AVX registers,
partially unrolling loops (even in the dynamic case), and eventually even
splitting the computation to different CPUs because each item in the array has
no dependences to other elements, it's a pure parallel loop.

Such vector ops can be useful for other languages and compilers too that use
LLVM as back-end (like valarray operator in C++).

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 16:25:26 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 16:25:26 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6957] New: specialization of virtual function
 member trigger a compilation error
Message-ID: <bug-6957-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6957

           Summary: specialization of virtual function member trigger a
                    compilation error
           Product: clang
           Version: 2.7
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: cyb at lepi.org
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4748)
 --> (http://llvm.org/bugs/attachment.cgi?id=4748)
test case

In case a template class contains two virtual members, the specialization of
the second virtual function trigger the following error:

test31.cpp:18:16: error: explicit specialization of 'hello2' after
instantiation
void A<float>::hello2( )
               ^
test31.cpp:12:15: note: implicit instantiation first required here
int A<float>::hello()
              ^
2 diagnostics generated.


In attachment, there is a code snippet that show the problem (that code
compiles with gcc and msvc).

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 17:36:09 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 17:36:09 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6056] clang ignores -fno-constant-cfstrings on OS X
In-Reply-To: <bug-6056-206@http.llvm.org/bugs/>
References: <bug-6056-206@http.llvm.org/bugs/>
Message-ID: <20100427223609.CF9183128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6056

Fariborz Jahanian <fjahanian at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #6 from Fariborz Jahanian <fjahanian at apple.com> 2010-04-27 17:36:09 CDT ---
Addition of patch: http://llvm.org/viewvc/llvm-project?view=rev&revision=102431
(by Daniel) completes this PR.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 18:35:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 18:35:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6958] New: Performance issue.
Message-ID: <bug-6958-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6958

           Summary: Performance issue.
           Product: clang
           Version: 2.7
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ldm at thorgal.homelinux.org
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Everything I read on LLVM's site indicate that clang is markedly faster than
g++ (http://clang.llvm.org/performance.html). I develop a large C++ project
(close to 1M LOC)  with g++ on MacOS and Linux and MSDev on windows. The code
is portable and true C++. When I saw the release of 2.7 with (apparently) good
support for C++, I thought I'd give it a spin on a file or two to see the
performance level. I compiled one C++ source file (LocRuntime.C)

-rw-r--r--@ 1 ldm  ldm  66790 Apr  4 18:13 LocRuntime.C

On a MacOS 10.6.3 system

Darwin tanatlocwl.linux.home 10.3.0 Darwin Kernel Version 10.3.0: Fri Feb 26
11:58:09 PST 2010; root:xnu-1504.3.12~1/RELEASE_I386 i386

The compilation options are for debug. Nothing fancy. Maybe a few thousand
class prototypes and a few dozen class compiled here. 

I timed the stock g++ and clang++. Here is the comparison.

tanatlocwl:runtime ldm$ time g++-4.0 -msse2 -msse -c -D_REENTRANT
-DCOMETVER='"2.1.0"' -DCOMETBUILD='"0eb0b297a134c494febffb68405ee0efce2bcde8"' 
-fPIC -g  -Woverloaded-virtual   -DX11GUI=1 -I. -I../include -I../data
-I../localizer -I/sw/include -I../include/data LocRuntime.C 
real    0m2.497s
user    0m2.267s
sys    0m0.224s

And

tanatlocwl:runtime ldm$ time
~ldm/Downloads/Applications/clang+llvm-2.7-x86_64-apple-darwin10/bin/clang++
-msse2 -msse -c -D_REENTRANT -DCOMETVER='"2.1.0"'
-DCOMETBUILD='"0eb0b297a134c494febffb68405ee0efce2bcde8"'  -fPIC -g 
-Woverloaded-virtual   -DX11GUI=1 -I. -I../include -I../data -I../localizer
-I/sw/include -I../include/data LocRuntime.C

real    0m3.785s
user    0m3.552s
sys    0m0.193s


This is with the binaries for Darwin from the llvm site (for 2.7). 

So the comparison gives gcc-4.0 as the clear winner with  2.5s realtime vs.
3.8s for clang++ for the simplest compilation options that are in use during
the development cycle (when one wishes the compiler to be really fast). 

What am I missing? 

(I'd love to switch to a faster compiler for everyday use!)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Tue Apr 27 19:00:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Tue, 27 Apr 2010 19:00:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6911] Function pointer with incomplete type cannot
 be passed as argument
In-Reply-To: <bug-6911-206@http.llvm.org/bugs/>
References: <bug-6911-206@http.llvm.org/bugs/>
Message-ID: <20100428000051.35CBC3128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6911

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from John McCall <rjmccall at apple.com> 2010-04-27 19:00:50 CDT ---
Fixed in r102473.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 03:31:36 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 03:31:36 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6960] New: x86 codegen: vectors don't work on
	non-SSE CPUs
Message-ID: <bug-6960-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6960

           Summary: x86 codegen: vectors don't work on non-SSE CPUs
           Product: libraries
           Version: 2.7
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4751)
 --> (http://llvm.org/bugs/attachment.cgi?id=4751)
bitcode

Reported on #dri-devel, that softpipe/llvmpipe in mesa crashes like this, on a
CPU without SSE or SSE2:
#
glxgears: X86ISelLowering.cpp:6058: llvm::SDValue
llvm::X86TargetLowering::LowerVSETCC(llvm::SDValue, llvm::SelectionDAG&):
Assertion `VT0 == MVT::v4f32 || VT0 == MVT::v2f64' failed.
#
Stack dump:
#
0.      Running pass 'X86 DAG->DAG Instruction Selection' on function
'@draw_llvm_shader'

I dumped the module generated by mesa, and attached it here.

You can reproduce the bug like this (on llvm 2.7):
$ llc l.bc -march=x86 -mattr=-sse
UNREACHABLE executed!
0  llc 0x08631db9
Stack dump:
0.      Program arguments: /home/edwin/llvm-git/obj32/Release-Asserts/bin/llc
l.bc -march=x86 -mattr=-sse
1.      Running pass 'X86 DAG->DAG Instruction Selection' on function
'@draw_llvm_shader'
Aborted

And like this:
$ llc l.bc -march=x86 -mattr=-sse2
LLVM ERROR: Cannot yet select: 0x9d86fe4: v2i32 = X86ISD::CMPPD 0x9d98964,
0x9d7eccc, 0x9d7308c [ID=518]
  0x9d98964: v2f32 = bit_convert 0x9d988e0 [ID=512]
    0x9d988e0: v1i64,ch = load 0x9d98754, 0x9d987d8, 0x9d7d344<LD8[<unknown>]>
[ID=506]
      0x9d98754: ch = store 0x9d54080, 0x9d88918, 0x9d986d0, 
.....

/proc/cpuinfo:
#
processor       : 0
#
vendor_id       : AuthenticAMD
#
cpu family      : 6
#
model           : 3
#
model name      : AMD Duron(tm) Processor
#
stepping        : 1
#
cpu MHz         : 950.024
#
cache size      : 64 KB
#
fdiv_bug        : no
#
hlt_bug         : no
#
f00f_bug        : no
#
coma_bug        : no
#
fpu             : yes
#
fpu_exception   : yes
#
cpuid level     : 1
#
wp              : yes
#
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov
pat pse36 mmx fxsr syscall mmxext 3dnowext 3dnow up
#
bogomips        : 1900.04
#
clflush size    : 32
#
cache_alignment : 32
#
address sizes   : 36 bits physical, 32 bits virtual
#
power management:

For reference here is how I dumped the module:
diff --git a/src/gallium/auxiliary/draw/draw_llvm.c
b/src/gallium/auxiliary/draw/draw_llvm.c
index 2738322..8e0a939 100644
--- a/src/gallium/auxiliary/draw/draw_llvm.c
+++ b/src/gallium/auxiliary/draw/draw_llvm.c
@@ -179,6 +179,7 @@ draw_llvm_create(struct draw_context *draw)

    llvm->pass = LLVMCreateFunctionPassManager(llvm->provider);
    LLVMAddTargetData(llvm->target, llvm->pass);
+   fprintf(stderr, "%s\n", LLVMCopyStringRepOfTargetData(llvm->target));
    /* These are the passes currently listed in llvm-c/Transforms/Scalar.h,
     * but there are more on SVN. */
    /* TODO: Add more passes */
@@ -694,6 +695,7 @@ draw_llvm_generate(struct draw_llvm *llvm, struct
draw_llvm_variant *variant)
       debug_printf("\n");
    }
    variant->jit_func =
(draw_jit_vert_func)LLVMGetPointerToGlobal(llvm->draw->engine,
variant->function);
+   LLVMDumpModule(llvm->module);

    if (0)
       lp_disassemble(variant->jit_func);

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 03:40:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 03:40:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6961] New: Incorrect prototype missing void in
 argument (easy to fix).
Message-ID: <bug-6961-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6961

           Summary: Incorrect prototype missing void in argument (easy to
                    fix).
           Product: new-bugs
           Version: 2.7
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: Lars.Rasmusson at sics.se
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4753)
 --> (http://llvm.org/bugs/attachment.cgi?id=4753)
Patch that makes the function prototype correct.

In include/llvm-c/Target.h function prototypes are generated with a macro.

The function prototypes look like

void fun();

which when compiling with gcc gives the warning

warning: function declaration isn?t a prototype

The reason for this is according to
http://stackoverflow.com/questions/42125/function-declaration-isnt-a-prototype

In C foo() and foo(void) are different functions. foo() accepts an infinite
number of arguments, while foo(void) accepts 0 arguments. In C++ they mean the
same thing. I suggest that you use void consistently when you mean no
arguments.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 04:06:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 04:06:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6962] New: A Concrete class inside a template class
 is identified as abstract
Message-ID: <bug-6962-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6962

           Summary: A Concrete class inside a template class is identified
                    as abstract
           Product: clang
           Version: 2.7
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: etienne.pierre at c-s.fr
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4755)
 --> (http://llvm.org/bugs/attachment.cgi?id=4755)
Source to reproduce the bug

Attached is a simple code example (complied with clang++ -c bug.cpp) that
reproduces the problem.

Inside a template class (template <class T> class TemplateClass) are 2 inner
classes, the first one is abstract (Interface), the second inherits from it and
is concrete (Concrete). If I try to add a member to TemplateClass of type
Concrete, the compiler gives me the following message :

bug.cpp:21:14: error: field type 'Concrete' is an abstract class
    Concrete m_c;

bug.cpp:10:22: note: pure virtual function 'Write'
        virtual void Write(const char * a_pBuf) = 0;

The funny thing is that, if you remove line 21 (Concrete m_c;), the compiler
doesn't complain on line 26 (which uses TemplateClass<char>::Concrete but
outside TemplateClass).


Thanks for the great work on clang.

?tienne PIERRE

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 04:30:20 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 04:30:20 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6963] New: Crash from boost code
Message-ID: <bug-6963-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6963

           Summary: Crash from boost code
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chris at bubblescope.net
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4756)
 --> (http://llvm.org/bugs/attachment.cgi?id=4756)
Reduced test case

TThe attached instance is reduced from a boost::spirit example. I have tried to
reduce this testcase further, but have failed. Sorry. The test makes clang
crash, not assert. It compiles correctly in g++.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 05:47:57 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 05:47:57 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6964] New: Assertion in boost testcase
Message-ID: <bug-6964-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6964

           Summary: Assertion in boost testcase
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chris at bubblescope.net
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


The attached testcase, reduced from boost, causes an assertion in clang++. It
compiles fine in g++.

t.cc:1:32: warning: non-type template argument value '9223372036854775807'
truncated to '-1'
      for template parameter of type 'int'
template <typename ,int, int = 9223372036854775807L >
                               ^~~~~~~~~~~~~~~~~~~~
t.cc:1:30: note: template parameter is declared here                            
template <typename ,int, int = 9223372036854775807L >
                             ^
t.cc:1:32: warning: non-type template argument value '9223372036854775807'
truncated to '-1' 
      for template parameter of type 'int'
template <typename ,int, int = 9223372036854775807L >
                               ^~~~~~~~~~~~~~~~~~~~
clang: SemaTemplate.cpp:5521: std::string
clang::Sema::getTemplateArgumentBindingsText(const
clang::TemplateParameterList*, const clang::TemplateArgument*, unsigned int):
Assertion `false && "No expressions in deduced template arguments!"' failed.
0  clang           0x0000000001490e82
1  clang           0x0000000001490d69
2  libpthread.so.0 0x00007f2d64aca8f0
3  libc.so.6       0x00007f2d63dbaa75 gsignal + 53
4  libc.so.6       0x00007f2d63dbe5c0 abort + 384
5  libc.so.6       0x00007f2d63db3941 __assert_fail + 241
6  clang           0x00000000007d5e35
7  clang           0x00000000007f17aa
8  clang           0x000000000068eb00
9  clang           0x00000000007cc31d
10 clang           0x00000000007c8162
11 clang           0x00000000007c927a
12 clang           0x00000000007cfb2f
13 clang           0x0000000000a6270a
14 clang           0x0000000000a5714e
15 clang           0x0000000000a7b498
16 clang           0x0000000000a7b2f2
17 clang           0x0000000000a7ae67
18 clang           0x0000000000a53d72
19 clang           0x0000000000a4e2af
20 clang           0x0000000000a4dc79
21 clang           0x000000000068be08
22 clang           0x000000000043705c
23 clang           0x0000000000436cc7
24 clang           0x00000000004212eb
25 clang           0x0000000000409172
26 clang           0x000000000040dce2 main + 259
27 libc.so.6       0x00007f2d63da5c4d __libc_start_main + 253
28 clang           0x0000000000407b49
Stack dump:
0.      Program arguments: /usr/local/bin/clang -cc1 -triple
x86_64-unknown-linux-gnu -S -disable-free -main-file-name t.cc
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -resource-dir /usr/local/lib/clang/1.5
-ferror-limit 19 -ftemplate-backtrace-limit 10 -fmessage-length 93 -fexceptions
-fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics -o /tmp/cc-ycf7Js.s
-x c++ t.cc 
1.      t.cc:6:1: current parser token '{'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 09:08:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 09:08:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6965] New: Alignment for ByVal arguments does not
 work on x86_64
Message-ID: <bug-6965-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6965

           Summary: Alignment for ByVal arguments does not work on x86_64
           Product: new-bugs
           Version: 2.6
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: arushi987 at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4758)
 --> (http://llvm.org/bugs/attachment.cgi?id=4758)
test case.

Adding an alignment attribute to a ByVal argument, does not align it correctly.

This is a simple example, with no optimizations run on it.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 10:47:33 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 10:47:33 -0500 (CDT)
Subject: [LLVMbugs] =?utf-8?q?=5BBug_6966=5D_New=3A_llvm/include/llvm/Supp?=
 =?utf-8?b?b3J0L0FsbG9jYXRvci5oOjEwMzoxODogbm90ZTogIOKAmGxsdm0gOjpCdW1w?=
 =?utf-8?q?PtrAllocator=3A=3A_Allocator=E2=80=99_should_be_initialized?=
Message-ID: <bug-6966-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6966

           Summary: llvm/include/llvm/Support/Allocator.h:103:18: note:
                    ?llvm::BumpPtrAllocator::Allocator? should be
                    initialized
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gryzman at gmail.com
                CC: llvmbugs at cs.uiuc.edu


make[4]: Entering directory
`/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend'                     
llvm[4]: Compiling GeneratePCH.cpp for Release build                            
In file included from
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/Decl.h:17:0, 
                 from
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/Frontend/PCHWriter.h:18, 
                 from GeneratePCH.cpp:16:                                       
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?clang::APValue::APSInt& clang::APValue::getInt()?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:111:29:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?clang::APValue::APFloat& clang::APValue::getFloat()?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:119:30:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?clang::APValue& clang::APValue::getVectorElt(unsigned
int)?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:127:31:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?const clang::APValue& clang::APValue::getVectorElt(unsigned
int) const?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:131:43:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?unsigned int clang::APValue::getVectorLength() const?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:135:44:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?clang::APValue::APSInt&
clang::APValue::getComplexIntReal()?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:140:41:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?clang::APValue::APSInt&
clang::APValue::getComplexIntImag()?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:148:41:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?clang::APValue::APFloat&
clang::APValue::getComplexFloatReal()?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:156:42:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?clang::APValue::APFloat&
clang::APValue::getComplexFloatImag()?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:164:42:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?void clang::APValue::setInt(const
clang::APValue::APSInt&)?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:175:22:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?void clang::APValue::setFloat(const
clang::APValue::APFloat&)?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:179:23:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?void clang::APValue::setVector(const clang::APValue*,
unsigned int)?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:183:24:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:184:24:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:186:26:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?void clang::APValue::setComplexInt(const
clang::APValue::APSInt&, const clang::APValue::APSInt&)?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:192:34:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:193:34:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:
In member function ?void clang::APValue::setComplexFloat(const
clang::APValue::APFloat&, const clang::APValue::APFloat&)?:
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:199:35:
warning: dereferencing type-punned pointer will break strict-aliasing rules
/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/../../include/clang/AST/APValue.h:200:35:
warning: dereferencing type-punned pointer will break strict-aliasing rules
GeneratePCH.cpp: In constructor ?<unnamed>::PCHGenerator::PCHGenerator(const
clang::Preprocessor&, const char*, llvm::raw_ostream*)?:
GeneratePCH.cpp:52:19: error: uninitialized reference member in ?class
clang::MemorizeStatCalls? using ?new?
/home/gjaskie/Projects/llvm/llvm/include/llvm/Support/Allocator.h:103:18: note:
?llvm::BumpPtrAllocator::Allocator? should be initialized
/bin/rm: cannot remove
`/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/Release/GeneratePCH.d.tmp':
No such file or directory
make[4]: ***
[/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend/Release/GeneratePCH.o]
Error 1
make[4]: Leaving directory
`/home/gjaskie/Projects/llvm/llvm/tools/clang/lib/Frontend'

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 12:59:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 12:59:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6967] New: Friend doesn't work for template method
	declaration.
Message-ID: <bug-6967-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6967

           Summary: Friend doesn't work for template method declaration.
           Product: clang
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dnljms at gmail.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4759)
 --> (http://llvm.org/bugs/attachment.cgi?id=4759)
A test case.

In the attached test, the methods can't access foo::type at instantiation
although I think they should since access is a friend of foo.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 13:48:04 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 13:48:04 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6968] New: add -N to options.td and -N/-S to
 freebsd linker driver
Message-ID: <bug-6968-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6968

           Summary: add -N to options.td and -N/-S to freebsd linker
                    driver
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P
         Component: Driver
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4760)
 --> (http://llvm.org/bugs/attachment.cgi?id=4760)
patch

the attached patch adds -N to options.td and -N/-S to freebsd linker driver.

this is necessary for bootloader in freebsd to fit into the specified space

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 13:53:04 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 13:53:04 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6969] New: Getting wint_t from <wctype.h> does not
	work
Message-ID: <bug-6969-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6969

           Summary: Getting wint_t from <wctype.h> does not work
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Headers
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: pipping.elias at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Hi,

I'm trying to compile lsof on linux. wint_t ends up undefined. lsof #includes
<wctype.h>, which, according to e.g.
http://www.opengroup.org/onlinepubs/007908799/xsh/wctype.h.html should define
wint_t.

The header looks like this:

[..]
/* Get wint_t from <stddef.h>.  */
# define __need_wint_t
# include <stddef.h>
[..]

clang comes with its own copy of <stddef.h> however.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 14:46:51 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 14:46:51 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6969] Getting wint_t from <wctype.h> does not work
In-Reply-To: <bug-6969-206@http.llvm.org/bugs/>
References: <bug-6969-206@http.llvm.org/bugs/>
Message-ID: <20100428194651.A39F83128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6969

Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |asl at math.spbu.ru
         Resolution|                            |DUPLICATE

--- Comment #2 from Anton Korobeynikov <asl at math.spbu.ru> 2010-04-28 14:46:51 CDT ---


*** This bug has been marked as a duplicate of bug 6691 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 15:16:47 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 15:16:47 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6112] llvm-gcc: cast of pointer to constant value
 fails in llvm BitcodeWriter.cpp: 714
In-Reply-To: <bug-6112-206@http.llvm.org/bugs/>
References: <bug-6112-206@http.llvm.org/bugs/>
Message-ID: <20100428201648.26D1D312800A@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6112

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #7 from Chris Lattner <clattner at apple.com> 2010-04-28 15:16:47 CDT ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20100426/100501.html

CallbackVH to the rescue :-)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 15:24:55 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 15:24:55 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6961] Incorrect prototype missing void in argument
 (easy to fix).
In-Reply-To: <bug-6961-206@http.llvm.org/bugs/>
References: <bug-6961-206@http.llvm.org/bugs/>
Message-ID: <20100428202455.8489A3128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6961

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-28 15:24:55 CDT ---
Thanks, applied in r102523

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 17:50:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 17:50:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 5880] clang c++ crash instantiating
	__builtin_offsetof
In-Reply-To: <bug-5880-206@http.llvm.org/bugs/>
References: <bug-5880-206@http.llvm.org/bugs/>
Message-ID: <20100428225003.16BD73128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=5880

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Douglas Gregor <dgregor at apple.com> 2010-04-28 17:50:02 CDT ---
Fixed in r102542.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 19:35:35 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 19:35:35 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6967] Friend doesn't work for template method
	declaration.
In-Reply-To: <bug-6967-206@http.llvm.org/bugs/>
References: <bug-6967-206@http.llvm.org/bugs/>
Message-ID: <20100429003535.84F2E3128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6967

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from John McCall <rjmccall at apple.com> 2010-04-28 19:35:35 CDT ---
Fixed in r102572.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 19:53:41 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 19:53:41 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6970] New: Clang does not convert "ptr to const
 member function" to "ptr to member function"
Message-ID: <bug-6970-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6970

           Summary: Clang does not convert "ptr to const member function"
                    to "ptr to member function"
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: matthewbg at google.com
                CC: llvmbugs at cs.uiuc.edu, nlewycky at google.com,
                    dgregor at apple.com


The C++ standard doesn't explicitly allow conversion of "(T::*)() const" to
"(T::*)()". While this will hopefully be fixed at some point, in the interim it
would be great if Clang would allow this behavior as an extension. Currently
(see attached const_member_function.cc):
% clang++ -fsyntax-only const_member_function.cc 
const_member_function.cc:10:3: error: no matching function for call to 'g'
  g(&X::foo);
  ^
const_member_function.cc:6:6: note: candidate function not viable: no known
      conversion from 'void (X::*)() const' to 'void (X::*)()' for 1st argument
void g(void (X::*ptr_to_member)(void)) {
     ^
1 error generated.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 19:57:48 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 19:57:48 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6970] Clang does not convert "ptr to const member
 function" to "ptr to member function"
In-Reply-To: <bug-6970-206@http.llvm.org/bugs/>
References: <bug-6970-206@http.llvm.org/bugs/>
Message-ID: <20100429005748.2DADB3128034@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6970

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #2 from Douglas Gregor <dgregor at apple.com> 2010-04-28 19:57:47 CDT ---
I am against adding this extension to Clang, because adding implicit
conversions has a huge effect on overloading, template argument deduction, etc.
and will make Clang non-conforming in ways that will be hard to detect or
explain. No other compiler supports such an extension, so we aren't doing users
any favors by allowing this.

The proper way to get this change is to work through the C++ standards
committee to get the change into an upcoming standard, and Clang will implement
it then.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 19:58:17 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 19:58:17 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6970] Clang does not convert "ptr to const member
 function" to "ptr to member function"
In-Reply-To: <bug-6970-206@http.llvm.org/bugs/>
References: <bug-6970-206@http.llvm.org/bugs/>
Message-ID: <20100429005817.2F3573128034@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6970

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|INVALID                     |WONTFIX

--- Comment #3 from Douglas Gregor <dgregor at apple.com> 2010-04-28 19:58:16 CDT ---
Oops. WONTFIX, not INVALID.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 20:10:42 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 20:10:42 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6938] clang:crashes when compiling nmi.c
In-Reply-To: <bug-6938-206@http.llvm.org/bugs/>
References: <bug-6938-206@http.llvm.org/bugs/>
Message-ID: <20100429011042.5D0F23128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6938

Ted Kremenek <kremenek at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |kremenek at apple.com
         Resolution|                            |FIXED

--- Comment #3 from Ted Kremenek <kremenek at apple.com> 2010-04-28 20:10:41 CDT ---
Fixed in r102576.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 20:39:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 20:39:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6971] New: Code generation for non-constant
	__builtin_offsetof
Message-ID: <bug-6971-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6971

           Summary: Code generation for non-constant __builtin_offsetof
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu


__builtin_offsetof is represented by an OffsetOfExpr in C++, and we would like
to use that representation in C as well to eliminate UnaryOperator::OffsetOf
(which is a hack). However, we need code generation support for non-constant
offsetof expressions, such as the one in this example:

  struct sockaddr_un {
   unsigned char sun_len;
   char sun_path[104];
  };

  int test(int len) {
    return __builtin_offsetof(struct sockaddr_un, sun_path[len+1]);
  }

The UnaryOperator::OffsetOf implementation handles this by creating an lvalue,
then applying various array-index and member-reference operations to that value
and subtracting the original lvalue's address from the resulting lvalue's
address. We'll likely need to do the same for OffsetOfExpr. Once implemented,
we can kill off UnaryOperator::OffsetOf.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 22:33:31 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 22:33:31 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6972] New: virtual derived constructor with
 "throw()" is seen as too lax when templates are involved
Message-ID: <bug-6972-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6972

           Summary: virtual derived constructor with "throw()" is seen as
                    too lax when templates are involved
           Product: clang
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: peter.goodman at gmail.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4764)
 --> (http://llvm.org/bugs/attachment.cgi?id=4764)
Code used to test this problem.

I have 3 classes. The first, Foo, is a template with two parameters. The second
is a template class, Bar, with one parameter T that extends Foo<T,T>. The third
class, Baz, extends Bar<int> where int can be replaced with concrete type. All
classes have virtual destructors with throw() set. When attempting to compile
this, I am issued the following error:

--
    ./clang-test-throw-destructor.cpp:16:13: error: exception specification of
    overriding function is more lax than base version
        virtual ~Baz(void) throw() { }
                ^
    ./clang-test-throw-destructor.cpp:11:13: note: overridden virtual function
    is here
        virtual ~Bar(void) throw() { }
                ^
    1 error generated.
--


I *think* that this is incorrect as all three destructors have the same
exception specification.

Specs:
clang version 1.5 (trunk 101948)
Apple Model Id: MacBookPro5,5
Processor: Intel Core 2 Duo
OS: Mac OS X 10.6.3 (10D573)
Kernel Version: Darwin 10.3.0
Xcode version: 3.2.1 (1613)
Compiler flags: -O0 -g -pedantic -pedantic-errors -Wall -Werror -Wextra -ansi

Attached is the code that I tested with, and that the above errors reported.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 23:31:21 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 23:31:21 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6963] Crash from boost code
In-Reply-To: <bug-6963-206@http.llvm.org/bugs/>
References: <bug-6963-206@http.llvm.org/bugs/>
Message-ID: <20100429043121.9FE1D3128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6963

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-28 23:31:21 CDT ---
I fixed a crasher that looked just like this yesterday, and this test is no
longer crashing for me. Please re-open if it's still crashing for you.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Wed Apr 28 23:55:26 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Wed, 28 Apr 2010 23:55:26 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6964] Assertion in boost testcase
In-Reply-To: <bug-6964-206@http.llvm.org/bugs/>
References: <bug-6964-206@http.llvm.org/bugs/>
Message-ID: <20100429045526.243C83128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6964

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Douglas Gregor <dgregor at apple.com> 2010-04-28 23:55:25 CDT ---
Fixed in r102595.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 01:43:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 01:43:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6910] LLVM broken with -O3, or -O1 -ftree-vectorize
In-Reply-To: <bug-6910-206@http.llvm.org/bugs/>
References: <bug-6910-206@http.llvm.org/bugs/>
Message-ID: <20100429064356.2CAB2312800A@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6910

T?r?k Edwin <edwintorok at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from T?r?k Edwin <edwintorok at gmail.com> 2010-04-29 01:43:55 CDT ---
Fixed in r102604.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 03:11:42 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 03:11:42 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6973] New: Assertion in boost::statechart
Message-ID: <bug-6973-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6973

           Summary: Assertion in boost::statechart
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chris at bubblescope.net
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4768)
 --> (http://llvm.org/bugs/attachment.cgi?id=4768)
Asserting reduced test case

The attached test case, reduced from a boost::statechart test, causes the
following assertion:

caj at nyanko:~/clang-temp$ clang++ t.cc -w
clang: CodeGenFunction.h:1366: void
clang::CodeGen::CodeGenFunction::EmitCallArgs(clang::CodeGen::CallArgList&,
const T*, clang::ConstExprIterator, clang::ConstExprIterator) [with T =
clang::FunctionProtoType]: Assertion
`getContext().getCanonicalType(ArgType.getNonReferenceType()). getTypePtr() ==
getContext().getCanonicalType(Arg->getType()).getTypePtr() && "type mismatch in
call argument!"' failed.
0  clang           0x00000000014995da
1  clang           0x00000000014994c1
2  libpthread.so.0 0x00007fddcaaa48f0
3  libc.so.6       0x00007fddc9d94a75 gsignal + 53
4  libc.so.6       0x00007fddc9d985c0 abort + 384
5  libc.so.6       0x00007fddc9d8d941 __assert_fail + 241
6  clang           0x00000000005d305a
7  clang           0x00000000005daea7
8  clang           0x00000000005db5ed
9  clang           0x00000000005ce0e9
10 clang           0x00000000005f18b8
11 clang           0x00000000005fbd17
12 clang           0x00000000005fb338
13 clang           0x00000000005f9f3d
14 clang           0x00000000005c5e48
15 clang           0x0000000000640c1d
16 clang           0x0000000000641166
17 clang           0x0000000000640f23
18 clang           0x0000000000640af4
19 clang           0x000000000066c60b
20 clang           0x000000000066cc7d
21 clang           0x000000000056e42d
22 clang           0x000000000056c3d0
23 clang           0x000000000056b694
24 clang           0x0000000000569cac
25 clang           0x0000000000568a29
26 clang           0x000000000041aad1
27 clang           0x000000000068fe06
28 clang           0x000000000043705c
29 clang           0x0000000000436cc7
30 clang           0x00000000004212eb
31 clang           0x0000000000409172
32 clang           0x000000000040dce2 main + 259
33 libc.so.6       0x00007fddc9d7fc4d __libc_start_main + 253
34 clang           0x0000000000407b49
Stack dump:
0.      Program arguments: /usr/local/bin/clang -cc1 -triple
x86_64-unknown-linux-gnu -S -disable-free -main-file-name t.cc
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -resource-dir /usr/local/lib/clang/1.5 -w
-ferror-limit 19 -ftemplate-backtrace-limit 10 -fmessage-length 162
-fexceptions -fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics -o
/tmp/cc-0sLw0R.s -x c++ t.cc 
1.      <eof> parser at end of file
2.      Per-file LLVM IR generation
3.      t.cc:316:30: Generating code for declaration 'Entry'
4.      t.cc:316:76: LLVM IR generation of compound statement ('{}')
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 03:23:15 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 03:23:15 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6974] New: -fno-rtti and 'try' / 'catch' causes
 clang to SIGSEGV (or assertion failure)
Message-ID: <bug-6974-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6974

           Summary: -fno-rtti and 'try' / 'catch' causes clang to SIGSEGV
                    (or assertion failure)
           Product: clang
           Version: 2.7
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: slyfox at inbox.ru
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Removing '-fno-rtti' allows this file to build properly.

g++ compiles file below successfully (-fexceptions and -fno-rtti was used in
original code to allow exception handling within modules, but disable things,
like 'dynamic_cast', 'typeinfo' and relevant strings in generated object files)

// $ cat try_error.cc 
/*
 * clang++ -O0 -fexceptions -fno-rtti try_error.cc -emit-llvm -c
 */

class err {};

void f (void)
{
    try {}
    catch( err & /* error */ )
    {}
}

Trying to build it:
$ clang++ -O0 -fexceptions -fno-rtti try_error.cc -emit-llvm -c            
Instruction referencing instruction not embedded in a basic block!
  %_rethrow = alloca i8*                          ; <i8**> [#uses=4]
  store i8* %exc1, i8** %_rethrow
Instruction referencing instruction not embedded in a basic block!
  %exc1 = call i8* @llvm.eh.exception()           ; <i8*> [#uses=3]
  store i8* %exc1, i8** %_rethrow
Broken module found, compilation aborted!
0  clang           0x000000000116281f
1  clang           0x000000000116303d
2  libpthread.so.0 0x00007f74bc948120
3  libc.so.6       0x00007f74bbc661a5 gsignal + 53
4  libc.so.6       0x00007f74bbc675d0 abort + 384
5  clang           0x0000000001113f5b
6  clang           0x00000000010f03d6
7  clang           0x00000000010f05b1
8  clang           0x00000000010f07de
9  clang           0x000000000041b8a5
10 clang           0x00000000005f83f4
11 clang           0x000000000041ef49
12 clang           0x00000000004123d9
13 clang           0x0000000000414e5b main + 1675
14 libc.so.6       0x00007f74bbc52bbd __libc_start_main + 253
15 clang           0x000000000040ff59
Stack dump:
0.      Program arguments: /usr/bin/clang -cc1 -triple x86_64-pc-linux-gnu
-emit-llvm-bc -disable-free -main-file-name try_error.cc -mrelocation-model
static -mdisable-fp-elim -mconstructor-aliases -munwind-tables -target-cpu
x86-64 -resource-dir /usr/lib/clang/1.1 -O0 -fmessage-length 127 -fexceptions
-fno-rtti -fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics -o
try_error.o -x c++ try_error.cc 
1.      <eof> parser at end of file
2.      Per-function optimization
3.      Running pass 'Module Verifier' on function '@_Z1fv'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 04:18:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 04:18:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6961] Incorrect prototype missing void in argument
 (easy to fix).
In-Reply-To: <bug-6961-206@http.llvm.org/bugs/>
References: <bug-6961-206@http.llvm.org/bugs/>
Message-ID: <20100429091840.44AD1312800A@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6961

Lars R <Lars.Rasmusson at sics.se> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |
         AssignedTo|unassignedbugs at nondot.org   |Lars.Rasmusson at sics.se

--- Comment #2 from Lars R <Lars.Rasmusson at sics.se> 2010-04-29 04:18:39 CDT ---
Created an attachment (id=4769)
 --> (http://llvm.org/bugs/attachment.cgi?id=4769)
patch to add void to the declaration of inline functions

(I forgot that) also inline function arguments need to be declared as void to
prevent gcc from complaining.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 05:11:07 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 05:11:07 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6056] clang ignores -fno-constant-cfstrings on OS X
In-Reply-To: <bug-6056-206@http.llvm.org/bugs/>
References: <bug-6056-206@http.llvm.org/bugs/>
Message-ID: <20100429101107.B30713128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6056

Jonathan Schleifer <js-llvm-bugzilla at webkeks.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |

--- Comment #7 from Jonathan Schleifer <js-llvm-bugzilla at webkeks.org> 2010-04-29 05:11:07 CDT ---
Now it always references NSConstantString and ignores what's specified by
-fconstant-string-class, leading to not even compiling it now, as it can't find
the interface for NSConstantString.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 12:06:36 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 12:06:36 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6975] New: Crash in boost::lexical_cast test
Message-ID: <bug-6975-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6975

           Summary: Crash in boost::lexical_cast test
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chris at bubblescope.net
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4770)
 --> (http://llvm.org/bugs/attachment.cgi?id=4770)
Reduced testcase from lexical_cast

The attached file is from a reduced test case from boost::lexical_cast. The
code compiles correctly in g++ 4.4.



/home/caj/Dropbox/lex_test.cpp:49:1: warning: control may reach end of non-void
function [-Wreturn-type]
}
^
/home/caj/Dropbox/lex_test.cpp:37:2: warning: control reaches end of non-void
function [-Wreturn-type]                                                        
        }
        ^
/home/caj/Dropbox/lex_test.cpp:44:3: note: in instantiation of function
template specialization 'boost::lexical_cast<int, char const [1]>' requested
here                                                                            
                lexical_cast<int>("");
                ^
/home/caj/Dropbox/lex_test.cpp:31:3: warning: control reaches end of non-void
function [-Wreturn-type]                                                        
                }
                ^
/home/caj/Dropbox/lex_test.cpp:36:11: note: in instantiation of function
template specialization 'boost::detail::lexical_cast<int>' requested here       
                detail::lexical_cast( buf);
                        ^
/home/caj/Dropbox/lex_test.cpp:44:3: note: in instantiation of function
template specialization 'boost::lexical_cast<int, char const [1]>' requested
here                                                                            
                lexical_cast<int>("");
                ^
0  clang           0x00000000014995da                                           
1  clang           0x00000000014994c1
2  libpthread.so.0 0x00007f2d289068f0
3  clang           0x00000000005bdb66
4  clang           0x00000000005bd428
5  clang           0x00000000005bd010
6  clang           0x00000000005bd075
7  clang           0x00000000005c4c6a
8  clang           0x00000000005bc60b
9  clang           0x00000000005b99c3
10 clang           0x00000000005b98be
11 clang           0x0000000000642935
12 clang           0x0000000000640f46
13 clang           0x0000000000640af4
14 clang           0x0000000000641166
15 clang           0x0000000000640f23
16 clang           0x0000000000640af4
17 clang           0x000000000066c60b
18 clang           0x000000000066cc7d
19 clang           0x000000000056e42d
20 clang           0x000000000056c3d0
21 clang           0x000000000056b694
22 clang           0x0000000000569cac
23 clang           0x0000000000568a29
24 clang           0x000000000041aad1
25 clang           0x000000000068fe06
26 clang           0x000000000043705c
27 clang           0x0000000000436cc7
28 clang           0x00000000004212eb
29 clang           0x0000000000409172
30 clang           0x000000000040dce2 main + 259
31 libc.so.6       0x00007f2d27be1c4d __libc_start_main + 253
32 clang           0x0000000000407b49
Stack dump:
0.      Program arguments: /usr/local/bin/clang -cc1 -triple
x86_64-unknown-linux-gnu -S -disable-free -main-file-name lex_test.cpp
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -resource-dir /usr/local/lib/clang/1.5
-ferror-limit 19 -ftemplate-backtrace-limit 10 -fmessage-length 251
-fexceptions -fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics -o
/tmp/cc-OQCr02.s -x c++ /home/caj/Dropbox/lex_test.cpp 
1.      <eof> parser at end of file
2.      Per-file LLVM IR generation
3.      /home/caj/Dropbox/lex_test.cpp:20:43: Generating code for declaration
'boost::detail::lexical_stream_limited_src<int>::operator>>'
4.      /home/caj/Dropbox/lex_test.cpp:21:4: LLVM IR generation of compound
statement ('{}')
clang: error: compiler command failed due to signal 11 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 12:44:21 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 12:44:21 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6976] New: autoconf test for <stdbool.h> fails
Message-ID: <bug-6976-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6976

           Summary: autoconf test for <stdbool.h> fails
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Headers
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: pipping.elias at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Here's the error:

conftest.c:84:16: error: initializer element is not a compile-time constant
         int xlcbug = 1 / (&(digs + 5)[-2 + (bool) 1] == &digs[4] ? 1 : -1);

here's where it comes from:


configure: failed program was:
| /* confdefs.h */
| #define PACKAGE_NAME "strace"
| #define PACKAGE_TARNAME "strace"
| #define PACKAGE_VERSION "4.5.20"
| #define PACKAGE_STRING "strace 4.5.20"
| #define PACKAGE_BUGREPORT ""
| #define PACKAGE_URL ""
| #define PACKAGE "strace"
| #define VERSION "4.5.20"
| #define LINUX 1
| #define X86_64 1
| #define STDC_HEADERS 1
| #define HAVE_SYS_TYPES_H 1
| #define HAVE_SYS_STAT_H 1
| #define HAVE_STDLIB_H 1
| #define HAVE_STRING_H 1
| #define HAVE_MEMORY_H 1
| #define HAVE_STRINGS_H 1
| #define HAVE_INTTYPES_H 1
| #define HAVE_STDINT_H 1
| #define HAVE_UNISTD_H 1
| #define __EXTENSIONS__ 1
| #define _ALL_SOURCE 1
| #define _GNU_SOURCE 1
| #define _POSIX_PTHREAD_SEMANTICS 1
| #define _TANDEM_SOURCE 1
| #define STDC_HEADERS 1
| /* end confdefs.h.  */
| 
| #include <stdbool.h>
| #ifndef bool
|  "error: bool is not defined"
| #endif
| #ifndef false
|  "error: false is not defined"
| #endif
| #if false
|  "error: false is not 0"
| #endif
| #ifndef true
|  "error: true is not defined"
| #endif
| #if true != 1
|  "error: true is not 1"
| #endif
| #ifndef __bool_true_false_are_defined
|  "error: __bool_true_false_are_defined is not defined"
| #endif
| 
|       struct s { _Bool s: 1; _Bool t; } s;
| 
|       char a[true == 1 ? 1 : -1];
|       char b[false == 0 ? 1 : -1];
|       char c[__bool_true_false_are_defined == 1 ? 1 : -1];

|       char d[(bool) 0.5 == true ? 1 : -1];
|       bool e = &s;
|       char f[(_Bool) 0.0 == false ? 1 : -1];
|       char g[true];
|       char h[sizeof (_Bool)];
|       char i[sizeof s.t];
|       enum { j = false, k = true, l = false * true, m = true * 256 };
|       /* The following fails for
|          HP aC++/ANSI C B3910B A.05.55 [Dec 04 2003]. */
|       _Bool n[m];
|       char o[sizeof n == m * sizeof n[0] ? 1 : -1];
|       char p[-1 - (_Bool) 0 < 0 && -1 - (bool) 0 < 0 ? 1 : -1];
| #     if defined __xlc__ || defined __GNUC__
|        /* Catch a bug in IBM AIX xlc compiler version 6.0.0.0
|           reported by James Lemley on 2005-10-05; see
|          
http://lists.gnu.org/archive/html/bug-coreutils/2005-10/msg00086.html
|           This test is not quite right, since xlc is allowed to
|           reject this program, as the initializer for xlcbug is
|           not one of the forms that C requires support for.
|           However, doing the test right would require a runtime
|           test, and that would make cross-compilation harder.
|           Let us hope that IBM fixes the xlc bug, and also adds
|           support for this kind of constant expression.  In the
|           meantime, this test will reject xlc, which is OK, since
|           our stdbool.h substitute should suffice.  We also test
|           this with GCC, where it should work, to detect more
|           quickly whether someone messes up the test in the
|           future.  */
|        char digs[] = "0123456789";
|        int xlcbug = 1 / (&(digs + 5)[-2 + (bool) 1] == &digs[4] ? 1 : -1);
| #     endif
|       /* Catch a bug in an HP-UX C compiler.  See
|          http://gcc.gnu.org/ml/gcc-patches/2003-12/msg02303.html

|         
http://lists.gnu.org/archive/html/bug-coreutils/2005-11/msg00161.html
|        */
|       _Bool q = true;
|       _Bool *pq = &q;
| 
| int
| main ()
| {
| 
|       *pq |= q;
|       *pq |= ! q;
|       /* Refer to every declared value, to avoid compiler optimizations.  */
|       return (!a + !b + !c + !d + !e + !f + !g + !h + !i + !!j + !k + !!l
|               + !m + !n + !o + !p + !q + !pq);
| 
|   ;
|   return 0;
| }

I'm not sure who's at fault here. This does mean that clang can't compile
strace right away, though, because configure (generated by autoconf 2.65)
considers clang's <stdbool.h> useless and the type bool is used in the code
(albeit for exactly one variable).

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 13:27:56 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 13:27:56 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6973] Assertion in boost::statechart
In-Reply-To: <bug-6973-206@http.llvm.org/bugs/>
References: <bug-6973-206@http.llvm.org/bugs/>
Message-ID: <20100429182756.576CD3128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6973

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from Douglas Gregor <dgregor at apple.com> 2010-04-29 13:27:56 CDT ---
Fixed in r102650.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 15:32:20 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 15:32:20 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6977] New: crash (EXC_BAD_ACCESS) in
 clang::InitializationSequence::Diagnose
Message-ID: <bug-6977-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6977

           Summary: crash (EXC_BAD_ACCESS) in
                    clang::InitializationSequence::Diagnose
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ich at az2000.de
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Process:         clang [4220]
Path:            /Users/az/Programmierung/CppIDE/llvm-src/Debug/bin/clang
Identifier:      clang
Version:         ??? (???)
Code Type:       X86 (Native)
Parent Process:  clang++ [4219]

Interval Since Last Report:          173890 sec
Crashes Since Last Report:           2
Per-App Interval Since Last Report:  0 sec
Per-App Crashes Since Last Report:   3

Date/Time:       2010-04-29 22:29:12.421 +0200
OS Version:      Mac OS X 10.5.8 (9L30)
Report Version:  6
Anonymous UUID:  C99325D9-953E-44CB-A0FE-88CB8BF0B517

Exception Type:  EXC_BAD_ACCESS (SIGBUS)
Exception Codes: KERN_PROTECTION_FAILURE at 0x0000000000000000
Crashed Thread:  0

Application Specific Information:
Stack dump:
0.    Program arguments:
/Users/az/Programmierung/CppIDE/llvm-src/Debug/bin/clang -cc1 -triple
i386-apple-darwin9.0.0 -S -disable-free -main-file-name test3.cpp -pic-level 1
-mdisable-fp-elim -target-cpu yonah -g -resource-dir
/Users/az/Programmierung/CppIDE/llvm-src/Debug/lib/clang/1.5 -DDEBUG=1
-D_AI_DEBUG -D SYSTEM_DATA_DIR="/usr/share/games" -Iinclude -Wall -ferror-limit
19 -ftemplate-backtrace-limit 10 -fmessage-length 86 -pthread -fexceptions
-fdiagnostics-show-option -fcolor-diagnostics -o
/var/folders/P9/P9wWY+Z8HDeOPixVmDZ0J++++TI/-Tmp-/cc-U45hoK.s -x c++ test3.cpp 
1.    include/LieroX.h:49:43: current parser token '{'
2.    include/LieroX.h:46:1: parsing struct/union/class body 'lierox_t'


Thread 0 Crashed:
0   clang                             0x00308492
clang::InitializationSequence::Diagnose(clang::Sema&, clang::InitializedEntity
const&, clang::InitializationKind const&, clang::Expr**, unsigned int) + 2606
1   clang                             0x0030e0c5
clang::InitializationSequence::Perform(clang::Sema&, clang::InitializedEntity
const&, clang::InitializationKind const&,
clang::ASTMultiPtr<&(clang::ActionBase::DeleteExpr(void*))>, clang::QualType*)
+ 101
2   clang                             0x002a4212
BuildImplicitMemberInitializer(clang::Sema&, clang::CXXConstructorDecl*,
ImplicitInitializerKind, clang::FieldDecl*,
clang::CXXBaseOrMemberInitializer*&) + 1446
3   clang                             0x002a4cd7
clang::Sema::SetBaseOrMemberInitializers(clang::CXXConstructorDecl*,
clang::CXXBaseOrMemberInitializer**, unsigned int, bool) + 1625
4   clang                             0x002a5269
clang::Sema::ActOnMemInitializers(clang::OpaquePtr<0>, clang::SourceLocation,
void**, unsigned int, bool) + 543
5   clang                             0x005d86f3
clang::Parser::ParseConstructorInitializer(clang::OpaquePtr<0>) + 459
6   clang                             0x005c5eda
clang::Parser::ParseLexedMethodDefs(clang::Parser::ParsingClass&) + 1072
7   clang                             0x005dab19
clang::Parser::ParseCXXMemberSpecification(clang::SourceLocation, unsigned int,
clang::OpaquePtr<0>) + 1563
8   clang                             0x005dbecf
clang::Parser::ParseClassSpecifier(clang::tok::TokenKind,
clang::SourceLocation, clang::DeclSpec&, clang::Parser::ParsedTemplateInfo
const&, clang::AccessSpecifier, bool) + 4879
9   clang                             0x005ce641
clang::Parser::ParseDeclarationSpecifiers(clang::DeclSpec&,
clang::Parser::ParsedTemplateInfo const&, clang::AccessSpecifier,
clang::Parser::DeclSpecContext) + 7169
10  clang                             0x0060d1af
clang::Parser::ParseDeclarationOrFunctionDefinition(clang::Parser::ParsingDeclSpec&,
clang::AttributeList*, clang::AccessSpecifier) + 85 (Parser.cpp:539)
11  clang                             0x0060d590
clang::Parser::ParseDeclarationOrFunctionDefinition(clang::AttributeList*,
clang::AccessSpecifier) + 66 (Parser.cpp:590)
12  clang                             0x0060e370
clang::Parser::ParseExternalDeclaration(clang::CXX0XAttributeList) + 2314
(Parser.cpp:479)
13  clang                             0x0060e4ae
clang::Parser::ParseTopLevelDecl(clang::OpaquePtr<1>&) + 296 (Parser.cpp:351)
14  clang                             0x0022cbaa
clang::ParseAST(clang::Preprocessor&, clang::ASTConsumer*, clang::ASTContext&,
bool, bool, clang::CodeCompleteConsumer*) + 426
15  clang                             0x0003d594
clang::ASTFrontendAction::ExecuteAction() + 266
16  clang                             0x0003d484
clang::FrontendAction::Execute() + 278
17  clang                             0x0001f9b7
clang::CompilerInstance::ExecuteAction(clang::FrontendAction&) + 879
18  clang                             0x000031e3 cc1_main(char const**, char
const**, char const*, void*) + 1755
19  clang                             0x000069bf main + 233 (driver.cpp:181)
20  clang                             0x00001f36 start + 54

Thread 0 crashed with X86 Thread State (32-bit):
  eax: 0x00000000  ebx: 0x00307a75  ecx: 0x00000000  edx: 0x00000028
  edi: 0x002a504a  esi: 0x00000006  ebp: 0xbfffb5f8  esp: 0xbfffb200
   ss: 0x0000001f  efl: 0x00010202  eip: 0x00308492   cs: 0x00000017
   ds: 0x0000001f   es: 0x0000001f   fs: 0x00000000   gs: 0x00000037
  cr2: 0x00000000

Binary Images:
    0x1000 -  0x117ef1f +clang ??? (???) <e5bfc526f15f802e7f3d5c303c6ce501>
/Users/az/Programmierung/CppIDE/llvm-src/Debug/bin/clang
0x8fe00000 - 0x8fe2db43  dyld 97.1 (???) <458eed38a009e5658a79579e7bc26603>
/usr/lib/dyld
0x92737000 - 0x9273efe9  libgcc_s.1.dylib ??? (???)
<f53c808e87d1184c0f9df63aef53ce0b> /usr/lib/libgcc_s.1.dylib
0x946e4000 - 0x94741ffb  libstdc++.6.dylib ??? (???)
<04b812dcec670daa8b7d2852ab14be60> /usr/lib/libstdc++.6.dylib
0x94c50000 - 0x94db7ff3  libSystem.B.dylib ??? (???)
<c8f52e158bf540cc000146ca8a705958> /usr/lib/libSystem.B.dylib
0x95237000 - 0x9523bfff  libmathCommon.A.dylib ??? (???)
/usr/lib/system/libmathCommon.A.dylib
0xffff0000 - 0xffff1780  libSystem.B.dylib ??? (???) /usr/lib/libSystem.B.dylib

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 15:51:50 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 15:51:50 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6978] New: lit should not unset CCACHE_DIR
Message-ID: <bug-6978-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6978

           Summary: lit should not unset CCACHE_DIR
           Product: Test Suite
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: lit
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: pipping.elias at gmail.com
                CC: llvmbugs at cs.uiuc.edu, daniel at zuster.org


The gcc that I have in my path is the one provided by ccache. To work properly,
CCACHE_DIR needs to be set or else ccache will try to use $HOME/.ccache as its
cache location.

When, from inside tools/clang/test, I run

  CCACHE_DIR=/var/tmp/ccache \
  PATH=/usr/libexec/ccache/:$PATH \
  ../../../utils/lit/lit.py Driver

the variable CCACHE_DIR ends up unset for some reason, causing

  tools/clang/test/Driver/Output/hello.c.tmp.log

to contain

  "/usr/libexec/ccache/gcc" -c -m64 -o /tmp/cc-lRfkeD.o -x assembler
/tmp/cc-wRxkGW.s
ccache: failed to create [..]/.ccache (Permission denied)

if $HOME/.ccache is not writable (which with automated building, a dedicated
user account and proper sandboxing will be the case), causing the test to fail.
In such a scenario, not adding ccache's gcc wrapper to the PATH isn't a
feasible solution (because it's usually added globally). Instead, ccache is
usually disabled by setting CCACHE_DISABLE. Again, this variable does not get
through, so this doesn't work either.

Does lit.py need to strip CCACHE_* from the environment? If not, could it be
made not to?

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 15:52:32 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 15:52:32 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6979] New: error: no type named ...
Message-ID: <bug-6979-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6979

           Summary: error: no type named ...
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ich at az2000.de
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


macbook-7:openlierox az$ ../CppIDE/llvm-src/Debug/bin/clang++ foo.cpp 
In file included from /Users/az/Programmierung/openlierox/src/main.cpp:1:
In file included from /Users/az/Programmierung/openlierox/src/main.cpp:21:
In file included from
/Users/az/Programmierung/openlierox/./include/CClient.h:24:
In file included from
/Users/az/Programmierung/openlierox/./include/FastVector.h:14:
/Users/az/Programmierung/openlierox/./include/Event.h:68:25: error: no type
named
      'HandlerList' in 'Event<_Data>'
  const typename Event::HandlerList& get() { return base->m_handlers; }
        ~~~~~~~~~~~~~~~~^~~~~~~~~~~
1 error generated.


This seems invalid. This is the code:



template< typename _Data = EventData >
class Event : public _Event {
public:
    class Handler {
    public:
        typedef _Data Data;
        virtual ~Handler() {}
        virtual void operator()(_Data data) = 0;
        virtual bool operator==(const Handler& hndl) = 0;
        virtual Handler* copy() const = 0;
    };

    typedef std::list< Ref<Handler> > HandlerList;

...


I was not able though to reconstruct a test case because I always get the crash
of bug #6977. But I will attach the preprocessed file.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 16:49:09 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 16:49:09 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6953] __builtin_types_compatible_p does not take in
 account noreturn attributes
In-Reply-To: <bug-6953-206@http.llvm.org/bugs/>
References: <bug-6953-206@http.llvm.org/bugs/>
Message-ID: <20100429214909.2088D3128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6953

Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |

--- Comment #2 from Eli Friedman <sharparrow1 at yahoo.com> 2010-04-29 16:49:08 CDT ---
Fix was backed out.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 18:28:26 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 18:28:26 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6961] Incorrect prototype missing void in argument
 (easy to fix).
In-Reply-To: <bug-6961-206@http.llvm.org/bugs/>
References: <bug-6961-206@http.llvm.org/bugs/>
Message-ID: <20100429232826.957403128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6961

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #3 from Chris Lattner <clattner at apple.com> 2010-04-29 18:28:26 CDT ---
Committed in r102667.

However, that patch isn't valid and there is no reason to reopen the PR. 
Please send (valid :) patches to llvm-commits with [PATCH] in the subject line
in the future.   Thanks!

-Chris

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 18:35:45 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 18:35:45 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6980] New: Assertion in boost::exception test
Message-ID: <bug-6980-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6980

           Summary: Assertion in boost::exception test
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chris at bubblescope.net
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4775)
 --> (http://llvm.org/bugs/attachment.cgi?id=4775)
Test case which causes assertion

The attached code, reduced from a boost testcase, causes the following
assertion:

caj at nyanko:~/clang-temp2$ clang++ t.cc -c
t.cc:26:1: warning: control may reach end of non-void function [-Wreturn-type]
}
^
While deleting: label %init.check                                               
Use still stuck around after Def is destroyed:  br i1 %tobool, label
%init.check, label %init.end
clang: Value.cpp:75: virtual llvm::Value::~Value(): Assertion `use_empty() &&
"Uses remain when a value is destroyed!"' failed.
0  clang           0x000000000149aab6
1  clang           0x000000000149a99d
2  libpthread.so.0 0x00007f8d22a4b8f0
3  libc.so.6       0x00007f8d21d3ba75 gsignal + 53
4  libc.so.6       0x00007f8d21d3f5c0 abort + 384
5  libc.so.6       0x00007f8d21d34941 __assert_fail + 241
6  clang           0x0000000001427d82
7  clang           0x0000000001366481
8  clang           0x00000000011da1c4
9  clang           0x00000000011d98af
10 clang           0x00000000013665b6
11 clang           0x00000000010f5bea
12 clang           0x0000000001402fd3
13 clang           0x0000000001402c9d
14 clang           0x000000000140294d
15 clang           0x000000000041bdb0
16 clang           0x000000000041ab09
17 clang           0x000000000068fe86
18 clang           0x0000000000437070
19 clang           0x0000000000436cdb
20 clang           0x00000000004212ff
21 clang           0x0000000000409172
22 clang           0x000000000040dce2 main + 259
23 libc.so.6       0x00007f8d21d26c4d __libc_start_main + 253
24 clang           0x0000000000407b49
Stack dump:
0.      Program arguments: /usr/local/bin/clang -cc1 -triple
x86_64-unknown-linux-gnu -S -disable-free -main-file-name t.cc
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -resource-dir /usr/local/lib/clang/2.0
-ferror-limit 19 -ftemplate-backtrace-limit 10 -fmessage-length 122
-fexceptions -fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics -o
/tmp/cc-A7bEI7.s -x c++ t.cc 
1.      <eof> parser at end of file
2.      Code generation
3.      Running pass 'Remove unreachable blocks from the CFG' on function
'@_ZN5array17failed_rangecheckEv'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 18:40:00 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 18:40:00 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6981] New: JIT crashes on code containing debug info
Message-ID: <bug-6981-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6981

           Summary: JIT crashes on code containing debug info
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Common Code Generator Code
        AssignedTo: dpatel at apple.com
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu


1 crash seems to be about leaving debug intrinsics in code for too long

$ cat test.c
int main(void) {
  int i =3;
}
$ Debug/bin/clang -g test.c -emit-llvm -c -o test.bc

$ Debug/bin/lli test.bc
psuedo instructions should be removed before code emission
UNREACHABLE executed at
/Users/jyasskin/src/llvm/clang/src/lib/Target/X86/X86CodeEmitter.cpp:609!
0  libLLVM-2.8svn.dylib 0xe08fcee4 llvm::SearchForAddressOfSpecialSymbol(char
const*) + 478
1  libLLVM-2.8svn.dylib 0xe08fd32a llvm::sys::SetInterruptFunction(void (*)())
+ 436
2  libSystem.B.dylib    0x9065e2bb _sigtramp + 43
3  libSystem.B.dylib    0xffffffff _sigtramp + 1872371055
4  libSystem.B.dylib    0x906d223a raise + 26
5  libSystem.B.dylib    0x906de679 abort + 73
6  libLLVM-2.8svn.dylib 0xe08d0563 llvm::llvm_unreachable_internal(char const*,
char const*, unsigned int) + 201
7  libLLVM-2.8svn.dylib 0xe09964d8 (anonymous
namespace)::Emitter<llvm::JITCodeEmitter>::emitInstruction(llvm::MachineInstr
const&, llvm::TargetInstrDesc const*) + 1772
8  libLLVM-2.8svn.dylib 0xe09979f0 (anonymous
namespace)::Emitter<llvm::JITCodeEmitter>::runOnMachineFunction(llvm::MachineFunction&)
+ 576
9  libLLVM-2.8svn.dylib 0xe034e609
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) + 77
10 libLLVM-2.8svn.dylib 0xe0500ce9
llvm::FPPassManager::runOnFunction(llvm::Function&) + 287
11 libLLVM-2.8svn.dylib 0xe0500f0a
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 80
12 libLLVM-2.8svn.dylib 0xe050105f
llvm::FunctionPassManager::run(llvm::Function&) + 195
13 libLLVM-2.8svn.dylib 0xe05cbf63
llvm::JIT::runJITOnFunctionUnlocked(llvm::Function*, llvm::MutexGuard const&) +
123
14 libLLVM-2.8svn.dylib 0xe05cc2f6
llvm::JIT::getPointerToFunction(llvm::Function*) + 622
15 libLLVM-2.8svn.dylib 0xe05cd1c9 llvm::JIT::runFunction(llvm::Function*,
std::vector<llvm::GenericValue, std::allocator<llvm::GenericValue> > const&) +
99
16 libLLVM-2.8svn.dylib 0xe054e147
llvm::ExecutionEngine::runFunctionAsMain(llvm::Function*,
std::vector<std::string, std::allocator<std::string> > const&, char const*
const*) + 1135
17 lli                  0x00002c7d main + 2313
18 lli                  0x00002306 start + 54
Stack dump:
0.    Program arguments: Debug/bin/lli test.bc 
1.    Running pass 'X86 Machine Code Emitter' on function '@main'
Abort trap


The other seems to be about metadata that isn't materialized from bitcode soon
enough:

$ cat test2.c
int fib_left(int);
int fib_right(int);

int fib_left(int i) {
  if (i < 2)
    return 1;
  return fib_left(i-1) + fib_right(i-2);
}
int fib_right(int i) {
  if (i < 2)
    return 1;
  return fib_left(i-1) + fib_right(i-2);
}
int fib(int i) {
  if (i < 2)
    return 1;
  return fib_left(i-1) + fib_right(i-2);
}

int main() {
  int i = 40;
  fib(i);
}
$ Debug/bin/clang -g test2.c -emit-llvm -c -o test2.bc
$ Debug/bin/lli test2.bc
Assertion failed: (isa<X>(Val) && "cast_or_null<Ty>() argument of incompatible
type!"), function cast_or_null, file
/Users/jyasskin/src/llvm/clang/src/include/llvm/Support/Casting.h, line 213.
0  libLLVM-2.8svn.dylib 0xe08fcee4 llvm::SearchForAddressOfSpecialSymbol(char
const*) + 478
1  libLLVM-2.8svn.dylib 0xe08fd32a llvm::sys::SetInterruptFunction(void (*)())
+ 436
2  libSystem.B.dylib    0x9065e2bb _sigtramp + 43
3  libSystem.B.dylib    0xffffffff _sigtramp + 1872371055
4  libSystem.B.dylib    0x906d223a raise + 26
5  libSystem.B.dylib    0x906de679 abort + 73
6  libSystem.B.dylib    0x906d33db __assert_rtn + 101
7  libLLVM-2.8svn.dylib 0xe048eec7 llvm::cast_retty<llvm::MDNode,
llvm::Value*>::ret_type llvm::cast_or_null<llvm::MDNode,
llvm::Value>(llvm::Value*) + 107
8  libLLVM-2.8svn.dylib 0xe04de085 llvm::DbgDeclareInst::getAddress() const +
33
9  libLLVM-2.8svn.dylib 0xe083f31f
llvm::SelectionDAGBuilder::visitIntrinsicCall(llvm::CallInst const&, unsigned
int) + 4805
10 libLLVM-2.8svn.dylib 0xe0842967
llvm::SelectionDAGBuilder::visitCall(llvm::CallInst const&) + 231
11 libLLVM-2.8svn.dylib 0xe0843c3e llvm::SelectionDAGBuilder::visit(unsigned
int, llvm::User const&) + 1328
12 libLLVM-2.8svn.dylib 0xe085ca1b
llvm::SelectionDAGBuilder::visit(llvm::Instruction const&) + 111
13 libLLVM-2.8svn.dylib 0xe08749ed
llvm::SelectionDAGISel::SelectBasicBlock(llvm::MachineBasicBlock*,
llvm::BasicBlock const*, llvm::ilist_iterator<llvm::Instruction const>,
llvm::ilist_iterator<llvm::Instruction const>, bool&) + 45
14 libLLVM-2.8svn.dylib 0xe0875082
llvm::SelectionDAGISel::SelectAllBasicBlocks(llvm::Function const&) + 1510
15 libLLVM-2.8svn.dylib 0xe087538e
llvm::SelectionDAGISel::runOnMachineFunction(llvm::MachineFunction&) + 668
16 libLLVM-2.8svn.dylib 0xe034e609
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) + 77
17 libLLVM-2.8svn.dylib 0xe0500ce9
llvm::FPPassManager::runOnFunction(llvm::Function&) + 287
18 libLLVM-2.8svn.dylib 0xe0500f0a
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 80
19 libLLVM-2.8svn.dylib 0xe050105f
llvm::FunctionPassManager::run(llvm::Function&) + 195
20 libLLVM-2.8svn.dylib 0xe05cbf63
llvm::JIT::runJITOnFunctionUnlocked(llvm::Function*, llvm::MutexGuard const&) +
123
21 libLLVM-2.8svn.dylib 0xe05cc2f6
llvm::JIT::getPointerToFunction(llvm::Function*) + 622
22 libLLVM-2.8svn.dylib 0xe05cd1c9 llvm::JIT::runFunction(llvm::Function*,
std::vector<llvm::GenericValue, std::allocator<llvm::GenericValue> > const&) +
99
23 libLLVM-2.8svn.dylib 0xe054e147
llvm::ExecutionEngine::runFunctionAsMain(llvm::Function*,
std::vector<std::string, std::allocator<std::string> > const&, char const*
const*) + 1135
24 lli                  0x00002c7d main + 2313
25 lli                  0x00002306 start + 54
Stack dump:
0.    Program arguments: Debug/bin/lli test2.bc 
1.    Running pass 'X86 DAG->DAG Instruction Selection' on function '@main'
Abort trap


We seem to need more tests for this.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 19:07:07 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 19:07:07 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6975] Crash in boost::lexical_cast test
In-Reply-To: <bug-6975-206@http.llvm.org/bugs/>
References: <bug-6975-206@http.llvm.org/bugs/>
Message-ID: <20100430000707.5A51A3128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6975

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #2 from John McCall <rjmccall at apple.com> 2010-04-29 19:07:07 CDT ---
Fixed in r102677.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 19:48:33 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 19:48:33 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6982] New: Poor diagnostic for non-template used as
	template
Message-ID: <bug-6982-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6982

           Summary: Poor diagnostic for non-template used as template
           Product: clang
           Version: 2.7
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: me22.ca at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4776)
 --> (http://llvm.org/bugs/attachment.cgi?id=4776)
Source file (as pasted in description)

I came across this from a pasto on my part.  Given this code:

    template <int UnitBits>
    struct big_unit_big_bit {};
    typedef big_unit_big_bit<8> big_octet_big_bit;

    template <typename Endianness>
    struct imploder;
    template <int UnitBits>
    struct imploder<big_octet_big_bit<UnitBits> > {};

I get the following from clang++:

    clang.cc:9:34: error: expected '>'
    struct imploder<big_octet_big_bit<UnitBits> > {};
                                     ^
    clang.cc:9:1: error: declaration of anonymous struct must be a definition
    struct imploder<big_octet_big_bit<UnitBits> > {};
    ^
    clang.cc:9:1: error: declaration does not declare anything
    struct imploder<big_octet_big_bit<UnitBits> > {};
    ^
    3 diagnostics generated.

For comparison, g++ (I tried 3.4, 4.1, 4.3, and 4.4) gives this:

    clang.cc:9: error: ?big_octet_big_bit? is not a template
    clang.cc:9: error: template parameters not used in partial specialization:
    clang.cc:9: error:         ?UnitBits?

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 20:15:44 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 20:15:44 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6974] -fno-rtti and 'try' / 'catch' causes clang to
 SIGSEGV (or assertion failure)
In-Reply-To: <bug-6974-206@http.llvm.org/bugs/>
References: <bug-6974-206@http.llvm.org/bugs/>
Message-ID: <20100430011544.9272F3128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6974

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from John McCall <rjmccall at apple.com> 2010-04-29 20:15:44 CDT ---
Fixed in r102684.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Thu Apr 29 21:52:55 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Thu, 29 Apr 2010 21:52:55 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6681] __VERSION__ is missing clang/LLVM version
 information, __clang__ and __llvm_ have no useful number
In-Reply-To: <bug-6681-206@http.llvm.org/bugs/>
References: <bug-6681-206@http.llvm.org/bugs/>
Message-ID: <20100430025255.870F53128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6681

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |dgregor at apple.com
         Resolution|                            |FIXED

--- Comment #2 from Douglas Gregor <dgregor at apple.com> 2010-04-29 21:52:55 CDT ---
Implemented in Clang r102686.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 00:58:05 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 00:58:05 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6622] Assertion failed: Invalid GetElementPtrInst
 indices for type!
In-Reply-To: <bug-6622-206@http.llvm.org/bugs/>
References: <bug-6622-206@http.llvm.org/bugs/>
Message-ID: <20100430055805.376E53128026@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6622

John McCall <rjmccall at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #10 from John McCall <rjmccall at apple.com> 2010-04-30 00:58:04 CDT ---
Fixed in r102692.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 02:11:34 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 02:11:34 -0500 (CDT)
Subject: [LLVMbugs] [Bug 2066] [LSR] Suboptimal code for factorial function
In-Reply-To: <bug-2066-206@http.llvm.org/bugs/>
References: <bug-2066-206@http.llvm.org/bugs/>
Message-ID: <20100430071134.E21BE3128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=2066

Owen Anderson <resistor at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |resistor at mac.com
         Resolution|                            |FIXED

--- Comment #4 from Owen Anderson <resistor at mac.com> 2010-04-30 02:11:34 CDT ---
This appears to be fixed in HEAD.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 02:16:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 02:16:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6983] New: False positive
Message-ID: <bug-6983-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6983

           Summary: False positive
           Product: clang
           Version: 2.6
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: amitshah at gmx.net
                CC: llvmbugs at cs.uiuc.edu, amitshah at gmx.net


clang-2.6-0.5.pre1.fc12.x86_64


http://amitshah.fedorapeople.org/clang-output/2010-04-30-1/report-uvVsds.html#EndPath

The snippet is below.

114    static void do_flush_queued_data(VirtIOSerialPort *port, VirtQueue *vq,
115    VirtIODevice *vdev, bool_Bool discard)
116    {
117    VirtQueueElement elem;
118    
119    assert(port || discard)((port || discard) ? (void) (0) : __assert_fail
("port || discard"
, "/home/amit/src/qemu/hw/virtio-serial-bus.c", 119, __PRETTY_FUNCTION__
));

1
    Within the expansion of the macro 'assert':
a
    Assuming pointer value is null
120    
121    while ((discard || !port->throttled) && virtqueue_pop(vq, &elem)) {

2
    Dereference of null pointer
122    uint8_t *buf;

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 02:22:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 02:22:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6984] New: False positive in condition checks
	(switch)
Message-ID: <bug-6984-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6984

           Summary: False positive in condition checks (switch)
           Product: clang
           Version: 2.6
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: amitshah at gmx.net
                CC: llvmbugs at cs.uiuc.edu, amitshah at gmx.net


http://amitshah.fedorapeople.org/clang-output/2010-04-30-1/report-AQMFoC.html#EndPath

false positive in switch statement checks:

268     if (len < sizeof(cpkt)) {

1
    Taking false branch
269    /* The guest sent an invalid control packet */
270    return;
271    }
272    
273    cpkt.event = lduw_p(&gcpkt->event)lduw_be_p(&gcpkt->event);
274    cpkt.value = lduw_p(&gcpkt->value)lduw_be_p(&gcpkt->value);
275    
276    port = find_port_by_id(vser, ldl_p(&gcpkt->id)ldl_be_p(&gcpkt->id));
277    if (!port && cpkt.event != VIRTIO_CONSOLE_DEVICE_READY0)

2
    Assuming pointer value is null

3
    Taking false branch
278    return;
279    
280    switch(cpkt.event) {

4
    Control jumps to 'case 6:' at line 343
281    case VIRTIO_CONSOLE_DEVICE_READY0:
282    if (!cpkt.value) {
283    error_report("virtio-serial-bus: Guest failure in adding device %s\n",
284    vser->bus->qbus.name);
285    break;
286    }
287    /*
288    * The device is up, we can now tell the device about all the
289    * ports we have here.
290    */
291    QTAILQ_FOREACH(port, &vser->ports, next)for ((port) =
((&vser->ports)->tqh_first); (port); (
port) = ((port)->next.tqe_next)) {
292    send_control_event(port, VIRTIO_CONSOLE_PORT_ADD1, 1);
293    }
294    break;
295    
296    case VIRTIO_CONSOLE_PORT_READY3:
297    if (!cpkt.value) {
298    error_report("virtio-serial-bus: Guest failure in adding port %u for
device %s\n",
299    port->id, vser->bus->qbus.name);
300    break;
301    }
302    /*
303    * Now that we know the guest asked for the port name, we're
304    * sure the guest has initialised whatever state is necessary
305    * for this port. Now's a good time to let the guest know if
306    * this port is a console port so that the guest can hook it
307    * up to hvc.
308    */
309    if (port->is_console) {
310    send_control_event(port, VIRTIO_CONSOLE_CONSOLE_PORT4, 1);
311    }
312    
313    if (port->name) {
314    stw_p(&cpkt.event, VIRTIO_CONSOLE_PORT_NAME)stw_be_p(&cpkt.event, 7);
315    stw_p(&cpkt.value, 1)stw_be_p(&cpkt.value, 1);
316    
317    buffer_len = sizeof(cpkt) + strlen(port->name) + 1;
318    buffer = qemu_malloc(buffer_len);
319    
320    memcpy(buffer, &cpkt, sizeof(cpkt));
321    memcpy(buffer + sizeof(cpkt), port->name, strlen(port->name));
322    buffer[buffer_len - 1] = 0;
323    
324    send_control_msg(port, buffer, buffer_len);
325    qemu_free(buffer);
326    }
327    
328    if (port->host_connected) {
329    send_control_event(port, VIRTIO_CONSOLE_PORT_OPEN6, 1);
330    }
331    
332    /*
333    * When the guest has asked us for this information it means
334    * the guest is all setup and has its virtqueues
335    * initialised. If some app is interested in knowing about
336    * this event, let it know.
337    */
338    if (port->info->guest_ready) {
339    port->info->guest_ready(port);
340    }
341    break;
342    
343    case VIRTIO_CONSOLE_PORT_OPEN6:
344    port->guest_connected = cpkt.value;

5
    Dereference of null pointer

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 02:25:15 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 02:25:15 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6985] New: RFE: scan-build should take env
 variables for repository name, sha1, etc.
Message-ID: <bug-6985-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6985

           Summary: RFE: scan-build should take env variables for
                    repository name, sha1, etc.
           Product: clang
           Version: 2.6
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: amitshah at gmx.net
                CC: llvmbugs at cs.uiuc.edu


I run scan-build nightly on the qemu repository. Recording the sha-1 of the
working tree and the repository name is nice. I carry this patch locally for
scan-build doing this:

--- /usr/bin/scan-build 2009-08-04 22:35:18.000000000 +0530
+++ /home/amit/bin/scan-build   2010-01-12 22:37:37.871876419 +0530
@@ -22,6 +22,8 @@
 use Cwd qw/ getcwd abs_path /;
 use Sys::Hostname;

+my $RealBin = "/usr/lib64/clang-analyzer";
+
 my $Verbose = 0;       # Verbose output from this script.
 my $Prog = "scan-build";
 my $BuildName;
@@ -36,6 +38,9 @@
 my $HostName = HtmlEscape(hostname() || 'unknown');
 my $CurrentDir = HtmlEscape(getcwd());
 my $CurrentDirSuffix = basename($CurrentDir);
+my $RepoName = $ENV{'CLANG_REPONAME'};
+my $BranchName = $ENV{'CLANG_BRANCHNAME'};
+my $CommitHash = $ENV{'CLANG_COMMITHASH'};

 my $CmdArgs;

@@ -560,8 +565,9 @@
 <h1>${HtmlTitle}</h1>

 <table>
-<tr><th>User:</th><td>${UserName}\@${HostName}</td></tr>
-<tr><th>Working Directory:</th><td>${CurrentDir}</td></tr>
+<tr><th>Repository:</th><td>${RepoName}</td></tr>
+<tr><th>Branch:</th><td>${BranchName}</td></tr>
+<tr><th>Commit Hash:</th><td>${CommitHash}</td></tr>
 <tr><th>Command Line:</th><td>${CmdArgs}</td></tr>
 <tr><th>Date:</th><td>${Date}</td></tr>
 ENDTEXT

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 03:44:41 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 03:44:41 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6986] New: Assert in boost::multi_index
Message-ID: <bug-6986-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6986

           Summary: Assert in boost::multi_index
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chris at bubblescope.net
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4777)
 --> (http://llvm.org/bugs/attachment.cgi?id=4777)
reduced boost test case

The attached code, reduced from libs/multi_index/test/test_key_extractors.cpp ,
compiles fine in g++ and asserts clang++ as follows:

clang: SemaTemplate.cpp:3066:
clang::ASTOwningResult<&clang::ActionBase::DeleteExpr>
clang::Sema::BuildExpressionFromDeclTemplateArgument(const
clang::TemplateArgument&, clang::QualType, clang::SourceLocation): Assertion
`!RefExpr.isInvalid() && Context.hasSameType(((Expr*)
RefExpr.get())->getType(), ParamType)' failed.             
0  clang           0x000000000149a70a
1  clang           0x000000000149a5f1
2  libpthread.so.0 0x00007f7ee4afb8f0
3  libc.so.6       0x00007f7ee3deba75 gsignal + 53
4  libc.so.6       0x00007f7ee3def5c0 abort + 384
5  libc.so.6       0x00007f7ee3de4941 __assert_fail + 241
6  clang           0x00000000007d2e9b
7  clang           0x00000000007f85b7
8  clang           0x00000000007f8717
9  clang           0x00000000007fd0d2
10 clang           0x00000000007ffda8
11 clang           0x0000000000803b1e
12 clang           0x00000000007fc0f4
13 clang           0x00000000007fb886
14 clang           0x00000000007ff8b9
15 clang           0x0000000000803b1e
16 clang           0x00000000007fc0f4
17 clang           0x00000000007fb886
18 clang           0x00000000007fb9c4
19 clang           0x00000000007ff1b1
20 clang           0x0000000000804184
21 clang           0x00000000007fc1ab
22 clang           0x00000000007fb886
23 clang           0x00000000007fb9c4
24 clang           0x00000000007f8e6a
25 clang           0x00000000007f9581
26 clang           0x00000000007f9a90
27 clang           0x00000000007fa554
28 clang           0x000000000083a817
29 clang           0x000000000083ac8b
30 clang           0x00000000006dd8b3
31 clang           0x0000000000a5c989
32 clang           0x0000000000a5bc30
33 clang           0x0000000000a5b934
34 clang           0x0000000000a5b756
35 clang           0x0000000000a49b88
36 clang           0x0000000000a4b8af
37 clang           0x0000000000a508c2
38 clang           0x0000000000a56833
39 clang           0x0000000000a5bb4f
40 clang           0x0000000000a5627e
41 clang           0x0000000000a562e7
42 clang           0x0000000000a55c76
43 clang           0x0000000000a5554d
44 clang           0x000000000068fc00
45 clang           0x0000000000437014
46 clang           0x0000000000436c7f
47 clang           0x00000000004212ff
48 clang           0x0000000000409172
49 clang           0x000000000040dce2 main + 259
50 libc.so.6       0x00007f7ee3dd6c4d __libc_start_main + 253
51 clang           0x0000000000407b49
Stack dump:
0.      Program arguments: /usr/local/bin/clang -cc1 -triple
x86_64-unknown-linux-gnu -S -disable-free -main-file-name
test_key_extractors.cpp -mrelocation-model static -mdisable-fp-elim
-mconstructor-aliases -munwind-tables -target-cpu x86-64 -resource-dir
/usr/local/lib/clang/2.0 -I . -ferror-limit 19 -ftemplate-backtrace-limit 10
-fmessage-length 167 -fexceptions -fgnu-runtime -fdiagnostics-show-option
-fcolor-diagnostics -o /tmp/cc-FKim8c.s -x c++
libs/multi_index/test/test_key_extractors.cpp 
1.      libs/multi_index/test/test_key_extractors.cpp:171:18: current parser
token ';'
2.      libs/multi_index/test/test_key_extractors.cpp:167:1: parsing function
body 'test_key_extractors'
3.      libs/multi_index/test/test_key_extractors.cpp:167:1: in compound
statement ('{}')
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 04:40:40 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 04:40:40 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6987] New: Missing builtins prevent compiling
 <cmath> in g++ 4.4.3
Message-ID: <bug-6987-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6987

           Summary: Missing builtins prevent compiling <cmath> in g++
                    4.4.3
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chris at bubblescope.net
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Compiling <cmath> on g++ 4.4.3 requires the intrinsics:

__builtin_isfinite
__builtin_fpclassify

__builtin_isfinite is, I hope, self-explanitory. __builtin_fpclassify is
exactly what is required to implement fpclassify.

If you have trouble reproducing this, make sure your 'bits' directory is set up
correctly. In 4.4 it includes:

#define _GLIBCXX_USE_C99_MATH 1
#define _GLIBCXX_USE_C99_MATH_TR1 1

I discovered this as it also blocks boost on my system, which uses <cmath> in
some places.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 07:44:16 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 07:44:16 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6988] New: Calling template function makes
 assertion "!asLValue" fail
Message-ID: <bug-6988-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6988

           Summary: Calling template function makes assertion "!asLValue"
                    fail
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P
         Component: Static Analyzer
        AssignedTo: kremenek at apple.com
        ReportedBy: magnus.reftel at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The following small testcase makes the static analyzer crash

--

template<typename T> void foo(const T&);

void Fail()
{
        int i;
        foo(&i);
}

--

The error message the analyzer prints is

--

/opt/llvm-r102704/bin/clang++ --analyze testcase.cpp
assertion "!asLValue" failed: file "GRExprEngine.cpp", line 2749, function:
void clang::GRExprEngine::VisitUnaryOperator(clang::UnaryOperator*,
clang::ExplodedNode*, clang::ExplodedNodeSet&, bool)
Stack dump:
0.      Program arguments: /opt/llvm-r102704/bin/clang -cc1 -triple
i386-pc-cygwin -analyze -disable-free -main-file-name testcase.cpp
-analyzer-store=region -analyzer-opt-analyze-nested-blocks
-analyzer-check-dead-stores -analyzer-check-objc-mem -analyzer-eagerly-assume
-analyzer-check-objc-methodsigs -analyzer-check-objc-unused-ivars
-analyzer-output plist -w -mrelocation-model static -mdisable-fp-elim
-mconstructor-aliases -target-cpu pentium4 -resource-dir
/opt/llvm-r102704/lib/clang/2.0 -ferror-limit 19 -ftemplate-backtrace-limit 10
-fmessage-length 80 -fexceptions -fgnu-runtime -fdiagnostics-show-option
-fcolor-diagnostics -o testcase.plist -x c++ testcase.cpp
1.      <eof> parser at end of file
2.      testcase.cpp:6:2: Error evaluating statement
3.      testcase.cpp:6:2: Error evaluating statement
4.      testcase.cpp:6:6: Error evaluating statement
clang: error: analyzer command failed due to signal 6 (use -v to see
invocation)

--

I've tested with various versions from r98628 to r102704. They all have the
same behavior.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 08:07:10 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 08:07:10 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6989] New: Problematic using declarations
Message-ID: <bug-6989-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6989

           Summary: Problematic using declarations
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chris at bubblescope.net
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


The following problem comes from boost. Consider the following code:

template<typename T>
struct A {
  typedef int type;
};

template<typename T>
struct B : A<T> {
    using A<T>::type;

};

B<int> b;


In clang, the 'using' line must be changed to: using typename A<T>::type;

Now, I believe this is technically correct. However, with g++ version does not
accept typename, and the clang version does not work without it, meaning it is
impossible to write this code in such a way that it compiles in both g++ and
clang!

Comeau compiles both with and without the typename, in both strict and relaxed
mode.

I believe the only thing which can be placed after that 'using' is a type.
Given that, would clang consider allowing the typename to be missing? (Possibly
with a warning?)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 10:48:28 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 10:48:28 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6989] Problematic using declarations
In-Reply-To: <bug-6989-206@http.llvm.org/bugs/>
References: <bug-6989-206@http.llvm.org/bugs/>
Message-ID: <20100430154828.3C07E3128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6989

Christopher Jefferson <chris at bubblescope.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID

--- Comment #1 from Christopher Jefferson <chris at bubblescope.net> 2010-04-30 10:48:27 CDT ---
Just in case anyone comes searching at this bug report.

Using 'typedef typename A<T>::type type' produces the same effect, and works on
all compilers.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 10:55:45 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 10:55:45 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6987] Missing builtins prevent compiling <cmath> in
	g++ 4.4.3
In-Reply-To: <bug-6987-206@http.llvm.org/bugs/>
References: <bug-6987-206@http.llvm.org/bugs/>
Message-ID: <20100430155545.862F73128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6987

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-30 10:55:45 CDT ---


*** This bug has been marked as a duplicate of bug 6083 ***

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 11:44:05 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 11:44:05 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6991] New: extern "C" typedef ....
Message-ID: <bug-6991-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6991

           Summary: extern "C" typedef ....
           Product: clang
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: Fons.Rademakers at cern.ch
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


In file included from core/editline/src/TTermManip.cxx:12:
core/editline/src/TTermManip.h:19:12: error: cannot combine with previous
      'extern' declaration specifier
extern "C" typedef int (*PutcFunc_t)(int);
           ^

This construct is flagged as an error while it is accepted by all other know
C++ compilers.

Cheers, Fons.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 11:44:01 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 11:44:01 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6990] New: Assert in boost::lexer
Message-ID: <bug-6990-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6990

           Summary: Assert in boost::lexer
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: chris at bubblescope.net
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Created an attachment (id=4779)
 --> (http://llvm.org/bugs/attachment.cgi?id=4779)
Lexer asserting test

The attached test case is reduced from boost::lexer.

clang: Sema.h:3664: void
clang::Sema::LocalInstantiationScope::InstantiatedLocal(const clang::Decl*,
clang::Decl*): Assertion `(!Stored || Stored == Inst) && "Already instantiated
this local"' failed.
0  clang           0x000000000149a70a
1  clang           0x000000000149a5f1
2  libpthread.so.0 0x00007f5554c9e8f0
3  libc.so.6       0x00007f5553f8ea75 gsignal + 53
4  libc.so.6       0x00007f5553f925c0 abort + 384
5  libc.so.6       0x00007f5553f87941 __assert_fail + 241
6  clang           0x0000000000821c57
7  clang           0x00000000007f93a0
8  clang           0x00000000007f8984
9  clang           0x00000000007fb6b5
10 clang           0x00000000007f892d
11 clang           0x00000000008025d8
12 clang           0x00000000007fbe18
13 clang           0x00000000007fb886
14 clang           0x00000000007f8daa
15 clang           0x00000000007cd0c3
16 clang           0x00000000007ce79c
17 clang           0x00000000007cb849
18 clang           0x0000000000819309
19 clang           0x0000000000803be3
20 clang           0x00000000007fc0f4
21 clang           0x00000000007fb886
22 clang           0x00000000007f8daa
23 clang           0x000000000082891f
24 clang           0x00000000008306af
25 clang           0x000000000082cff0
26 clang           0x00000000007f9b23
27 clang           0x00000000007fa554
28 clang           0x000000000083a817
29 clang           0x000000000083abbe
30 clang           0x00000000006fd22f
31 clang           0x00000000006fd6fd
32 clang           0x0000000000a6aa3a
33 clang           0x0000000000a6a6b1
34 clang           0x0000000000a6c070
35 clang           0x0000000000a6a35c
36 clang           0x0000000000a5e9da
37 clang           0x0000000000a55ea1
38 clang           0x0000000000a562e7
39 clang           0x0000000000a55c76
40 clang           0x0000000000a5554d
41 clang           0x000000000068fc00
42 clang           0x0000000000437014
43 clang           0x0000000000436c7f
44 clang           0x00000000004212ff
45 clang           0x0000000000409172
46 clang           0x000000000040dce2 main + 259
47 libc.so.6       0x00007f5553f79c4d __libc_start_main + 253
48 clang           0x0000000000407b49
Stack dump:
0.      Program arguments: /usr/local/bin/clang -cc1 -triple
x86_64-unknown-linux-gnu -S -disable-free -main-file-name t.cc
-mrelocation-model static -mdisable-fp-elim -mconstructor-aliases
-munwind-tables -target-cpu x86-64 -resource-dir /usr/local/lib/clang/2.0
-ferror-limit 19 -ftemplate-backtrace-limit 10 -fmessage-length 166
-fexceptions -fgnu-runtime -fdiagnostics-show-option -fcolor-diagnostics -o
/tmp/cc-xpRayb.s -x c++ t.cc 
1.      t.cc:17:1: current parser token '{'
2.      t.cc:16:1: parsing struct/union/class body 'r'
clang: error: compiler command failed due to signal 6 (use -v to see
invocation)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 11:46:18 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 11:46:18 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6992] New: Unfortunate code placement for PHI
	lowering
Message-ID: <bug-6992-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6992

           Summary: Unfortunate code placement for PHI lowering
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: gohman at apple.com
                CC: llvmbugs at cs.uiuc.edu


In the testcase from PR2066:

define i32 @power(i32 %X) nounwind  {
entry:
        %cmp3 = icmp eq i32 %X, 0
        br i1 %cmp3, label %ifthen, label %ifend

ifthen:         ; preds = %ifend, %entry
        %accumulator.tr.lcssa = phi i32 [ 1, %entry ], [ %mul, %ifend ]
        ret i32 %accumulator.tr.lcssa

ifend:          ; preds = %ifend, %entry
        %indvar = phi i32 [ %indvar.next, %ifend ], [ 0, %entry ]
        %accumulator.tr1 = phi i32 [ %mul, %ifend ], [ 1, %entry ]
        %X.tr2 = sub i32 %X, %indvar
        %mul = mul i32 %X.tr2, %accumulator.tr1
        %indvar.next = add i32 %indvar, 1
        %exitcond = icmp eq i32 %indvar.next, %X
        br i1 %exitcond, label %ifthen, label %ifend
}

llc gets the loop right, but the rest of the function is hairy:

_power:                                 ## @power
## BB#0:                                ## %entry
        testl   %edi, %edi
        je      LBB0_3
## BB#1:                                ## %entry.ifend_crit_edge
        movl    $1, %eax
        .align  4, 0x90
LBB0_2:                                 ## %ifend
                                        ## =>This Inner Loop Header: Depth=1
        imull   %edi, %eax
        decl    %edi
        jne     LBB0_2
        jmp     LBB0_4
LBB0_3:                                 ## %entry.ifthen_crit_edge
        movl    $1, %eax
LBB0_4:                                 ## %ifthen
        ret

This code doesn't really need two "movl $1, %eax" instructions; just one in
BB#0 would suffice. Also, that would allow the loop exit to just fall through
into the return block rather than jumping over the extra movl.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 12:43:59 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 12:43:59 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6993] New: Extra load inside sparse matrix
 multiplication kernel
Message-ID: <bug-6993-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6993

           Summary: Extra load inside sparse matrix multiplication kernel
           Product: new-bugs
           Version: 2.7
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Keywords: code-quality
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bearophile at mailas.com
                CC: llvmbugs at cs.uiuc.edu


The SciMark2 benchmark shows two points where llvm-gcc 2.7 generates suboptimal
asm code (they can be found by visual inspection of the asm and comparing the
timings with gcc 4.5).

The less significant of those two points is in the "Sparse matmult" part of
SciMark2 (the other more important problem is left to another bug report). A
good thing of this point is that it's easy to spot and it's compact.

In attach there is a reduced version of SciMark2 in C language that performs
only the Sparse matmult.

Timings, NLOOPS=400_000, seconds:
  gcc 4.5:       5.49
  llvm-gcc 2.7:  5.69

CPU Celeron 2.13 GHz, Windows Vista 32 bit.

In both cases code compiled with:
-O3 -s -fomit-frame-pointer -msse3 -march=core2


The situation can also be seen inspecting the asm generated by the two
compilers:

gcc, inner loop of SparseCompRow_matmult():
L45:
    movl    (%esi,%eax,4), %edx
    fldl    (%edi,%edx,8)
    fmull   (%ebx,%eax,8)
    incl    %eax
    faddp   %st, %st(1)
    cmpl    %eax, %ecx
    jg  L45


llvm-gcc, inner loop of SparseCompRow_matmult():
LBB3_6:
    movl    (%edi), %edx
    movl    52(%esp), %ebp
    addl    $4, %edi
    movsd   (%ebp,%edx,8), %xmm1
    mulsd   (%ebx), %xmm1
    addl    $8, %ebx
    decl    %esi
    addsd   %xmm1, %xmm0
    jne LBB3_6


llvm-gcc here has four loads from memory instead of the necessary three.
Normally one extra load is not an important difference, but inside the inner
loops of numerical FP kernels even few extra asm instructions can be
significant.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 13:56:03 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 13:56:03 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6990] Assert in boost::lexer
In-Reply-To: <bug-6990-206@http.llvm.org/bugs/>
References: <bug-6990-206@http.llvm.org/bugs/>
Message-ID: <20100430185603.2D1D83128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6990

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-30 13:56:02 CDT ---
Fixed in r102732

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 14:21:14 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 14:21:14 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6994] New: Warning on use of implicit this
Message-ID: <bug-6994-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6994

           Summary: Warning on use of implicit this
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: jyasskin at google.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


http://code.google.com/p/unladen-swallow/wiki/StyleGuide recommends always
specifying this-> to access member variables. It would be nice to be able to
turn on a warning to enforce this.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 15:18:16 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 15:18:16 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6976] autoconf test for <stdbool.h> fails
In-Reply-To: <bug-6976-206@http.llvm.org/bugs/>
References: <bug-6976-206@http.llvm.org/bugs/>
Message-ID: <20100430201816.99A093128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6976

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX

--- Comment #1 from Chris Lattner <clattner at apple.com> 2010-04-30 15:18:16 CDT ---
This autoconf test is generating non-compliant code, and emulating GCC's bug
would be non-trivial.  This:

  &(diags+5)[-1] == &diags[4]

is not an integer constant expression, therefore it can't be used to initialize
a global in C.  Perhaps the xlcbug case can be ifdef'd with an xlc-specific
macro or something.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 15:31:55 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 15:31:55 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6995] New: C++ classes get lowered to gross IR types
Message-ID: <bug-6995-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6995

           Summary: C++ classes get lowered to gross IR types
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


In some thing like the example below, we compile the struct types to this:

%0 = type { i8*, i8* }
%1 = type { i8*, i8*, i32, i32, i8*, i64, i8*, i64 }
%2 = type { i8*, i8*, i8* }
%struct.A = type { i8** }
%struct.B = type { i8** }
%struct.C = type { [16 x i8] }
%struct.D = type { [16 x i8] }

This is really nasty, llvm-gcc produces something much more friendly to the
optimizers:

%struct.A = type { i32 (...)** }
%struct.B = type { i32 (...)** }
%struct.C = type { %struct.A, %struct.A }
%struct.D = type { %struct.C }


struct A {
  virtual int f() { return 1; }
};

struct B {
  virtual int f() { return 2; }
};

struct C : A, B {
  virtual int f() { return 3; }
};

struct D : C {
  virtual int f() { return 4; }
};

static int f(D* d) {
  B* b = d;
  return b->f();
};

int g() {
  D d;

  return f(&d);
}

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 16:09:25 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 16:09:25 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6996] New: vtables marked weak_odr instead of
	linkonce_odr
Message-ID: <bug-6996-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6996

           Summary: vtables marked weak_odr instead of linkonce_odr
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


In this testcase:

struct A {
  virtual int f() { return 1; }
};

struct B {
  virtual int f() { return 2; }
};

struct C : A, B {
  virtual int f() { return 3; }
};

struct D : C {
  virtual int f() { return 4; }
};
//}

static int f(D* d) {
  B* b = d;
  return b->f();
};

int g() {
  D d;
  return f(&d);
}

The vtables and related gunk are emitted as weak_odr instead of linkonce_odr. 
Because the vtables will be emitted in each translation unit that could
possibly use them (they don't have a key function) they can be linkonce_odr. 
When llvm devirtualizes this example, we end up emitting all the vtables etc
even though they aren't used.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 16:46:50 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 16:46:50 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6986] Assert in boost::multi_index
In-Reply-To: <bug-6986-206@http.llvm.org/bugs/>
References: <bug-6986-206@http.llvm.org/bugs/>
Message-ID: <20100430214650.B83E23128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6986

Douglas Gregor <dgregor at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Douglas Gregor <dgregor at apple.com> 2010-04-30 16:46:50 CDT ---
Fixed in r102777.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 17:21:44 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 17:21:44 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6997] New: Bus Error compiling Boost.Uuid
Message-ID: <bug-6997-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6997

           Summary: Bus Error compiling Boost.Uuid
           Product: clang
           Version: trunk
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: jfreeman at cse.tamu.edu
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Compiling source files that include Boost.Uuid result in bus errors. The error
message is:

clang: error: compiler command failed due to signal 10 (use -v to see
invocation)

Sometimes, this error is preceded by a stack dump:

Stack dump:
0.    Program arguments: /path/to/bin/clang -cc1 -triple i386-apple-darwin9.0.0
-S -disable-free -disable-llvm-verifier -main-file-name Session.cpp -pic-level
1 -mdisable-fp-elim -target-cpu yonah -g -resource-dir /path/to/lib/clang/2.0
-Isource -Iinclude -I/opt/local/include -O0 -Wunused-parameter
-Wunused-variable -ferror-limit 19 -ftemplate-backtrace-limit 10
-fmessage-length 235 -fexceptions -fdiagnostics-show-option -fcolor-diagnostics
-o /var/folders/4k/4ki6lTM0Hl02Xo7FxM1LC++++TI/-Tmp-/cc-tbwlL0.s -x c++
source/Session.cpp 
1.    <eof> parser at end of file
2.    /opt/local/include/boost/uuid/uuid_generators.hpp:82:10: instantiating
function definition 'boost::uuids::string_generator::operator()'


System information:

MacBook Pro
Intel processor
Mac OS X 10.5.8

Boost version 1.42, installed through MacPorts
Clang and LLVM built from trunk, version 102712

(test case coming as soon as I get it distilled)

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 18:29:37 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 18:29:37 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6998] New: clang poor template instantiation
	performance
Message-ID: <bug-6998-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6998

           Summary: clang poor template instantiation performance
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: alp at nuanti.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com


Taken from my notes in PR6958:

Here's a quick comparative test putting g++ against clang on the template
instantiation test found in the paper "Effects of Metaprogramming Style on
Compilation Time"
(http://users.rcn.com/abrahams/instantiation_speed/index.html). 

This is a TOT release build, no debug symbols, codegen enabled.

By N=1000, gcc is ~10 times faster, and the growth ratio doesn't look too
healthy. Good news is, resource use seems constant and more stable than gcc,
and some simple profiling suggests the problem is localised to Sema, with
CodeGen/IRGen doing fine, and the parser performing very well.

Given the heavy templating in the previously attached Comet test case, this
could be a good starting point to investigate. The paper is worth reading and
I'd be interested to hear about other profiling work going on and the tools
being used.

http://nuanti.com/tmp/clang-tmpl-2010-04-29.png

                          Template instantiation timings
  CPU time (seconds)
    30 ++------------------------------------------------------------------+B
       |    gcc 4.4.3 **A***                                              ##|
       |clang r102560 ##B###                                            #B  |
    25 ++                                                             ##   ++
       |                                                            #B      |
       |                                                         #B#        |
    20 ++                                                      ##          ++
       |                                                     #B             |
       |                                                 ##B#               |
    15 ++                                              B#                  ++
       |                                             ##                     |
       |                                          ##B                       |
    10 ++                                      #B#                         ++
       |                                ##B##B#                             |
       |                             #B#                                    |
     5 ++                        ##B#                                      ++
       |                  ##B##B#                                      **A**A
       +      + #B###B##B#  +      +    **A**A**A***A**A***A**A***A**A*     +
     0 ++-A***A**A***A**A***A**A***A**A*--+-----+------+------+------+-----++
       0     100    200    300    400    500   600    700    800    900    1000
                                    -DREPEATS=n

# N g++ clang++
50 0.14 0.26
100 0.19 0.51
150 0.25 0.93
200 0.32 1.47
250 0.40 2.14
300 0.50 2.95
350 0.60 3.91
400 0.72 4.98
450 0.83 6.22
500 0.97 7.56
550 1.10 9.08
600 1.26 10.68
650 1.42 12.44
700 1.58 14.48
750 1.78 16.66
800 1.96 18.61
850 2.16 20.93
900 2.46 23.39
950 2.69 26.22
1000 3.03 29.87

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 19:06:20 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 19:06:20 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6999] New: [C++ CodeGen] Reimplement synthesis of
 implicit copy assignment operators
Message-ID: <bug-6999-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=6999

           Summary: [C++ CodeGen] Reimplement synthesis of implicit copy
                    assignment operators
           Product: clang
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: C++
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: dgregor at apple.com
                CC: llvmbugs at cs.uiuc.edu, dgregor at apple.com
            Blocks: 6023


terfin:clang dgregor$ cat t.cpp 
template<typename T>
CodeGen is picking the wrong copy-assignment operator for bases/members when it
synthesizes the implicit copy assignment operator. Here's an example (distilled
from Boost.Spirit) where we fail to link because CodeGen tries to call the
const copy-assignment operator when it should call the non-const one.

struct X {
  X &operator=(const X&) {
    T* a = 1;
    return *this;
  }

  X &operator=(X &) { return *this; }
};

struct Y {
  Y &operator=(Y &) { return *this; }
};

struct Z : X<int>, Y { };

void test_Z(Z z1, Z z2) {
  z1 = z2;
}
terfin:clang dgregor$ clang++ t.cpp 
Undefined symbols:
  "_main", referenced from:
      start in crt1.10.6.o
  "__ZN1XIiEaSERKS0_", referenced from:
      __ZN1ZaSERS_ in cc-kOxAc3.o

The plan is to implement the generation of copy-assignment operators entirely
in Sema, synthesizing a body that CodeGen can translate like any other function
body.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 20:19:32 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 20:19:32 -0500 (CDT)
Subject: [LLVMbugs] [Bug 6724] call to devirtualized thunk not being inlined
In-Reply-To: <bug-6724-206@http.llvm.org/bugs/>
References: <bug-6724-206@http.llvm.org/bugs/>
Message-ID: <20100501011932.082083128018@llvm.org>

http://llvm.org/bugs/show_bug.cgi?id=6724

Chris Lattner <clattner at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #9 from Chris Lattner <clattner at apple.com> 2010-04-30 20:19:31 CDT ---
Implemented in a series of patches leading up to r102823.  The testcase is now
compiled into "return 4" with a lot of dead vtable gunk due to Bug 6996. 
llvm-gcc is generating some thunks as weak instead of weak_odr, so it isn't
getting the inlining.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 20:28:10 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 20:28:10 -0500 (CDT)
Subject: [LLVMbugs] [Bug 7000] New: Diagnostic for Invalid Cast should give
 Actual Type Name in addition to Typedef Type Name
Message-ID: <bug-7000-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=7000

           Summary: Diagnostic for Invalid Cast should give Actual Type
                    Name in addition to Typedef Type Name
           Product: clang
           Version: 2.7
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: -New Bugs
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: me22.ca at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4785)
 --> (http://llvm.org/bugs/attachment.cgi?id=4785)
Source file (as pasted in description)

Compiling this bit of code:

    typedef unsigned char value_type;
    template <typename T>
    void update_n(T *p_) {
        value_type const *p = static_cast<value_type const *>(p_);
    }
    int main() {
        char *s;
        update_n(s);
    }

I get the following error from clang++:

    clang.cc:4:27: error: static_cast from 'char *' to 'value_type const *' is
not
          allowed
        value_type const *p = static_cast<value_type const *>(p_);
                              ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    clang.cc:8:5: note: in instantiation of function template specialization
          'update_n<char>' requested here
        update_n(s);
        ^

While it's nice to see the clean name sometimes, in this case knowing about
"value_type" is useless in figuring out why the cast is invalid -- particularly
since it just reiterates what's clear in the caret diagnostic.

In the actual situation from which I took this test case, the value_type
typedef is something derived through a metafunction from template parameters,
so I can't just look at its definition to tell what the type is, so it'd be
much more helpful to know the expanded type.  I'm sure there are plenty of
others where the typedef would be more helpful, and even this time it's nice to
know the typedef name, so it'd be great to have both.  In fact, probably the
nicest I could hope for would be something like this: "static_cast from 'char*'
(T*) to 'unsigned char const*' (value_type const *) is not allowed".  For
implicit casts it might be better to give the typedef names as primary and the
expanded names as parenthetical, though.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

From bugzilla-daemon at llvm.org  Fri Apr 30 22:02:04 2010
From: bugzilla-daemon at llvm.org (bugzilla-daemon at llvm.org)
Date: Fri, 30 Apr 2010 22:02:04 -0500 (CDT)
Subject: [LLVMbugs] [Bug 7001] New: MSP430 codegen: "Invalid truncate node"
 on a simple loop
Message-ID: <bug-7001-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=7001

           Summary: MSP430 codegen: "Invalid truncate node" on a simple
                    loop
           Product: new-bugs
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Keywords: compile-fail
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ransford at cs.umass.edu
                CC: asl at math.spbu.ru, llvmbugs at cs.uiuc.edu


Created an attachment (id=4786)
 --> (http://llvm.org/bugs/attachment.cgi?id=4786)
simplified test case (generated as described in ticket)

% clang -v
clang version 2.0 (trunk 102826)
Target: x86_64-apple-darwin10
Thread model: posix

% cat gt.c                                              
int main (void) {
    int l = 10, x = 100, r = 1;
    while (r < 10 && l <= x)
        ++r;
}

% clang -ccc-host-triple msp430-elf -emit-llvm -c -o gt.bc gt.c

% llc -march=msp430 gt.bc
Assertion failed:
(Operand.getValueType().getScalarType().bitsGT(VT.getScalarType()) && "Invalid
truncate node, src < dst!"), function getNode, file
/Users/ransford/vc/llvm-svn/lib/CodeGen/SelectionDAG/SelectionDAG.cpp, line
2474.
0  llc               0x00000001008b2cea PrintStackTrace(void*) + 38
1  llc               0x00000001008b31ca SignalHandler(int) + 312
2  libSystem.B.dylib 0x00007fff850c580a _sigtramp + 26
3  libSystem.B.dylib 0x00007fff8506a50a tiny_malloc_from_free_list + 1196
4  libSystem.B.dylib 0x00007fff85140ef0 __pthread_markcancel + 0
5  llc               0x000000010042af93 llvm::SelectionDAG::getNode(unsigned
int, llvm::DebugLoc, llvm::EVT, llvm::SDValue) + 7311
6  llc               0x00000001004a8d54
llvm::TargetLowering::SimplifySetCC(llvm::EVT, llvm::SDValue, llvm::SDValue,
llvm::ISD::CondCode, bool, llvm::TargetLowering::DAGCombinerInfo&,
llvm::DebugLoc) const + 8140
7  llc               0x00000001003521b5 (anonymous
namespace)::DAGCombiner::SimplifySetCC(llvm::EVT, llvm::SDValue, llvm::SDValue,
llvm::ISD::CondCode, llvm::DebugLoc, bool) + 187
8  llc               0x00000001003522a1 (anonymous
namespace)::DAGCombiner::visitSETCC(llvm::SDNode*) + 203
9  llc               0x00000001003816c1 (anonymous
namespace)::DAGCombiner::visit(llvm::SDNode*) + 1815
10 llc               0x0000000100381bdc (anonymous
namespace)::DAGCombiner::combine(llvm::SDNode*) + 44
11 llc               0x0000000100383b68 (anonymous
namespace)::DAGCombiner::Run(llvm::CombineLevel) + 498
12 llc               0x0000000100383f50
llvm::SelectionDAG::Combine(llvm::CombineLevel, llvm::AliasAnalysis&,
llvm::CodeGenOpt::Level) + 54
13 llc               0x000000010048ba2e
llvm::SelectionDAGISel::CodeGenAndEmitDAG(llvm::MachineBasicBlock*) + 1612
14 llc               0x000000010048dae8
llvm::SelectionDAGISel::SelectBasicBlock(llvm::MachineBasicBlock*,
llvm::BasicBlock const*, llvm::ilist_iterator<llvm::Instruction const>,
llvm::ilist_iterator<llvm::Instruction const>, bool&) + 242
15 llc               0x000000010048e0fe
llvm::SelectionDAGISel::SelectAllBasicBlocks(llvm::Function const&) + 1556
16 llc               0x000000010048e40b
llvm::SelectionDAGISel::runOnMachineFunction(llvm::MachineFunction&) + 671
17 llc               0x0000000100560f7b
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) + 85
18 llc               0x000000010082fe8c
llvm::FPPassManager::runOnFunction(llvm::Function&) + 348
19 llc               0x00000001008313fb
llvm::FunctionPassManagerImpl::run(llvm::Function&) + 79
20 llc               0x00000001008315c2
llvm::FunctionPassManager::run(llvm::Function&) + 136
21 llc               0x000000010002edff main + 2913
22 llc               0x000000010002db0c start + 52
23 llc               0x0000000000000003 start + 4294780203
Stack dump:
0.    Program arguments: llc -march=msp430 gt.bc 
1.    Running pass 'MSP430 DAG->DAG Pattern Instruction Selection' on function
'@main'
[1]    27390 abort      llc -march=msp430 gt.bc

% bugpoint -run-llc gt.bc --tool-args -march=msp430

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.