From bugzilla-daemon at cs.uiuc.edu  Sat Nov  1 01:31:37 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Nov 2008 01:31:37 -0500
Subject: [LLVMbugs] [Bug 3002] New: clang: crash on incomplete typedef
Message-ID: <bug-3002-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3002

           Summary: clang: crash on incomplete typedef
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Keywords: crash-on-invalid
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: richard at godbee.net
                CC: llvmbugs at cs.uiuc.edu


A .c file containing only the word "typedef" crashes clang r58543 (trunk):

(gdb) run
Starting program: /Users/rwg/llvm/inst-llvm/bin/clang -fsyntax-only
typedef-crash.c
Reading symbols for shared libraries +++. done
typedef-crash.c:1:8: error: expected identifier or '('
typedef
       ^
typedef-crash.c:1:1: error: declarator requires an identifier
typedef
^~~~~~~

Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_PROTECTION_FAILURE at address: 0x00000008
0x00528725 in clang::Decl::getKind (this=0x0) at DeclBase.h:151
151       Kind getKind() const { return DeclKind; }
(gdb) bt
#0  0x00528725 in clang::Decl::getKind (this=0x0) at DeclBase.h:151
#1  0x00528785 in clang::VarDecl::classof (D=0x0) at Decl.h:332
#2  0x0001f079 in llvm::isa_impl<clang::VarDecl, clang::Decl> (Val=@0x0) at
Casting.h:54
#3  0x00582935 in llvm::isa_impl_wrap<clang::VarDecl, clang::Decl const,
clang::Decl const>::doit (Val=@0x0) at Casting.h:71
#4  0x0001f08f in llvm::isa_impl_cl<clang::Decl>::isa<clang::VarDecl>
(Val=@0x0) at Casting.h:83
#5  0x0058294d in llvm::isa_impl_cl<clang::Decl*>::isa<clang::VarDecl>
(Val=0x0) at Casting.h:101
#6  0x0001f0a7 in llvm::isa<clang::VarDecl, clang::Decl*> (Val=@0xbfffef08) at
Casting.h:116
#7  0x0001f485 in llvm::dyn_cast<clang::VarDecl, clang::Decl*>
(Val=@0xbfffef08) at Casting.h:225
#8  0x00572182 in clang::Sema::ActOnUninitializedDecl (this=0xbffff270,
dcl=0x0) at SemaDecl.cpp:1616
#9  0x00636f5f in clang::Parser::ParseInitDeclaratorListAfterFirstDeclarator
(this=0xbffff448, D=@0xbffff030) at ParseDecl.cpp:314
#10 0x00637205 in clang::Parser::ParseSimpleDeclaration (this=0xbffff448,
Context=0) at ParseDecl.cpp:232
#11 0x00637278 in clang::Parser::ParseDeclaration (this=0xbffff448, Context=0)
at ParseDecl.cpp:209
#12 0x00651064 in clang::Parser::ParseExternalDeclaration (this=0xbffff448) at
Parser.cpp:363
#13 0x006510d7 in clang::Parser::ParseTopLevelDecl (this=0xbffff448,
Result=@0xbffff500) at Parser.cpp:289
#14 0x005690c0 in clang::ParseAST (PP=@0x1f03cb0, Consumer=0x1f04510,
PrintStats=false, FreeMemory=true) at ParseAST.cpp:55
#15 0x0003cca7 in ProcessInputFile (PP=@0x1f03cb0, PPF=@0xbffff7c8,
InFile=@0x1f025f0, PA=ParseSyntaxOnly) at clang.cpp:1336
#16 0x0003ec7e in main (argc=3, argv=0xbffff87c) at clang.cpp:1530


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov  1 19:19:22 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Nov 2008 19:19:22 -0500
Subject: [LLVMbugs] [Bug 2996] Assertion `New->getType() == getType() && "
 replaceAllUses of value with new value of different type!"' failed.
In-Reply-To: <bug-2996-206@http.llvm.org/bugs/>
Message-ID: <200811020019.mA20JMX1015681@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2996


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #4 from Dan Gohman <gohman at apple.com>  2008-11-01 19:19:21 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081027/069233.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 cs.uiuc.edu  Sat Nov  1 20:17:37 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Nov 2008 20:17:37 -0500
Subject: [LLVMbugs] [Bug 2994] assertion in indvars pass
In-Reply-To: <bug-2994-206@http.llvm.org/bugs/>
Message-ID: <200811020117.mA21Hbip017783@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2994


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #1 from Nick Lewycky <nicholas at mxc.ca>  2008-11-01 20:17:36 ---


*** This bug has been marked as a duplicate of bug 1827 ***


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov  1 21:41:58 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Nov 2008 21:41:58 -0500
Subject: [LLVMbugs] [Bug 2993] instcombine causes miscompilation with srem -1
In-Reply-To: <bug-2993-206@http.llvm.org/bugs/>
Message-ID: <200811020241.mA22fwPV020316@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2993


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #3 from Nick Lewycky <nicholas at mxc.ca>  2008-11-01 21:41:58 ---
Fixed in r58555.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov  1 22:52:24 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 1 Nov 2008 22:52:24 -0500
Subject: [LLVMbugs] [Bug 3003] New: clang: does not treat expression
 involving struct pointer as a constant in array declaration
Message-ID: <bug-3003-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3003

           Summary: clang: does not treat expression involving struct
                    pointer as a constant in array declaration
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: richard at godbee.net
                CC: llvmbugs at cs.uiuc.edu
            Blocks: 2778


This is a reduced version of another parser reject from DragonFly BSD's kernel
source (see bug 2778).

(If you think this is ugly, you should see the original code in all of its
expanded macro glory...)


struct str1 { char c; };

struct str2 {
  int intarray[ (int)&((struct str1 *)0)->c + 1];
};


% gcc-4.2 -W -Wall -pedantic -fsyntax-only -std=c99 array-length.c

% clang -fsyntax-only -std=c99 array-length.c
array-length.c:4:7: error: arrays with static storage duration must have
constant integer length
  int intarray[ (int)&((struct str1 *)0)->c + 1];
      ^~~~~~~~
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 cs.uiuc.edu  Sun Nov  2 08:35:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Nov 2008 08:35:48 -0600
Subject: [LLVMbugs] [Bug 3004] New: weak symbol is not weak
Message-ID: <bug-3004-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3004

           Summary: weak symbol is not weak
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


attribute(weak) doesn't turn the symbol into a weak symbol with llvm-gcc.

$ gcc foo.c -c -o foo.o
$ nm foo.o
0000000000000000 W init_IRQ
0000000000000000 T native_init_IRQ

^Correct

$ llvm-gcc foo.c -c -o foo.o
$ nm foo.o
0000000000000000 d EH_frame0
0000000000000000 T init_IRQ
0000000000000000 T native_init_IRQ
0000000000000018 d native_init_IRQ.eh
^Incorrect

/* testcase */

void init_IRQ(void) __attribute__((weak, alias("native_init_IRQ")));
void native_init_IRQ(void) {}


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov  2 08:38:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Nov 2008 08:38:26 -0600
Subject: [LLVMbugs] [Bug 3004] weak symbol is not weak
In-Reply-To: <bug-3004-206@http.llvm.org/bugs/>
Message-ID: <200811021438.mA2EcQ6d010051@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3004


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2008-11-02 08:38:25 ---


*** This bug has been marked as a duplicate of bug 2691 ***


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov  2 10:49:22 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Nov 2008 10:49:22 -0600
Subject: [LLVMbugs] [Bug 2691] weak attribute ignored on alias
In-Reply-To: <bug-2691-206@http.llvm.org/bugs/>
Message-ID: <200811021649.mA2GnMrY013998@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2691


Anton Korobeynikov <asl at math.spbu.ru> changed:

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




--- Comment #5 from Anton Korobeynikov <asl at math.spbu.ru>  2008-11-02 10:49:22 ---
Really 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 cs.uiuc.edu  Sun Nov  2 11:40:09 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Nov 2008 11:40:09 -0600
Subject: [LLVMbugs] [Bug 3005] New: partial specialization fails on sqlite3
Message-ID: <bug-3005-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3005

           Summary: partial specialization fails on sqlite3
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Interprocedural Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2175)
 --> (http://llvm.org/bugs/attachment.cgi?id=2175)
testcase

On this testcase, partial specialization produces invalid IR:

$ opt -partialspecialization -debug b.bc -disable-output
Both operands to ICmp instruction are not of the same type!
        %0 = icmp eq i32 (i8*, i32, i8**, i8**)* @callback, i8** null         
; <i1> [#uses=0]
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 cs.uiuc.edu  Sun Nov  2 11:45:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Nov 2008 11:45:47 -0600
Subject: [LLVMbugs] [Bug 3006] New: llc: assertion failure with inline asm
Message-ID: <bug-3006-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3006

           Summary: llc: assertion failure with inline asm
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2176)
 --> (http://llvm.org/bugs/attachment.cgi?id=2176)
bugpoint-reduced-simplified.ll

Using SVN r58569 on x86_64 linux:

$ llvm-as <bugpoint-reduced-simplified.ll |Release/bin/llc
llc: SelectionDAG.cpp:2355: llvm::SDValue llvm::SelectionDAG::getNode(unsigned
int, llvm::MVT, llvm::SDValue, llvm::SDValue): Assertion `VT.isInteger() &&
N1.getValueType() == N2.getValueType() && N1.getValueType() == VT && "Binary
operator types must match!"' failed.
Release/bin/llc[0x9f581f]
/lib/libc.so.6[0x7feb5531df60]
/lib/libc.so.6(gsignal+0x35)[0x7feb5531ded5]
/lib/libc.so.6(abort+0x183)[0x7feb5531f3f3]
/lib/libc.so.6(__assert_fail+0xe9)[0x7feb55316dc9]
Release/bin/llc(_ZN4llvm12SelectionDAG7getNodeEjNS_3MVTENS_7SDValueES2_+0xc41)[0x6d9a81]
Release/bin/llc(_ZN4llvm20SelectionDAGLowering11visitBinaryERNS_4UserEj+0xa7)[0x70f827]
Release/bin/llc(_ZN4llvm16SelectionDAGISel16SelectBasicBlockEPNS_10BasicBlockENS_14ilist_iteratorINS_11InstructionEEES5_+0x25d)[0x7223dd]
Release/bin/llc(_ZN4llvm16SelectionDAGISel20SelectAllBasicBlocksERNS_8FunctionERNS_15MachineFunctionEPNS_17MachineModuleInfoERKNS_15TargetInstrInfoE+0x2e5)[0x7227a5]
Release/bin/llc(_ZN4llvm16SelectionDAGISel13runOnFunctionERNS_8FunctionE+0x462)[0x723bd2]
Release/bin/llc(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0x15b)[0x99278b]
Release/bin/llc(_ZN4llvm23FunctionPassManagerImpl3runERNS_8FunctionE+0x86)[0x992c26]
Release/bin/llc(_ZN4llvm19FunctionPassManager3runERNS_8FunctionE+0x3e)[0x992dae]
Release/bin/llc(main+0x1051)[0x4d2d01]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7feb5530a1a6]
Release/bin/llc(_ZNSt8ios_base4InitD1Ev+0x39)[0x4d1219]
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 cs.uiuc.edu  Sun Nov  2 12:25:08 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 2 Nov 2008 12:25:08 -0600
Subject: [LLVMbugs] [Bug 3007] New: extern __attribute__((weak)) gets turned
 into undefined symbol
Message-ID: <bug-3007-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3007

           Summary: extern __attribute__((weak)) gets turned into undefined
                    symbol
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The kernel uses genksyms to generate crcs, include/linux/module.h has this
comment:
/* Mark the CRC weak since genksyms apparently decides not to
 * generate a checksums for some symbols */

/* testcase */
extern void * __attribute__((weak)) __crc_init_mm;
static const unsigned long __kcrctab_init_mm __attribute__((__used__))
__attribute__((section("__kcrctab" "_unused"), unused)) = (unsigned long)
&__crc_init_mm;
int main(void)
{
        return 0;
}

LLVM-gcc output:
$ /home/edwin/llvm-svn/install/bin/llvm-gcc foo.c
/tmp/cc2NMjnF.o:(__kcrctab_unused+0x0): undefined reference to `__crc_init_mm'
collect2: ld returned 1 exit status
$ /home/edwin/llvm-svn/install/bin/llvm-gcc foo.c -c && nm foo.o
0000000000000000 d EH_frame0
                 U __crc_init_mm
0000000000000000 r __kcrctab_init_mm
0000000000000000 T main
0000000000000018 d main.eh

GCC:
$ gcc foo.c
$ gcc -c foo.c && nm foo.o
                 w __crc_init_mm
0000000000000000 r __kcrctab_init_mm
0000000000000000 T main


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov  3 02:29:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Nov 2008 02:29:17 -0600
Subject: [LLVMbugs] [Bug 3009] New: Bogus warning on compound literals
Message-ID: <bug-3009-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3009

           Summary: Bogus warning on compound literals
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


void f(int);
void g(void) {f ((int) { 2 } ); }

produces

/tmp/foo.c:2:24: warning: braces around scalar initializer
void g(void) {f ((int) { 2 } ); }
                       ^~~~~
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 cs.uiuc.edu  Mon Nov  3 03:28:52 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Nov 2008 03:28:52 -0600
Subject: [LLVMbugs] [Bug 3001] clang: Assertion failed: ... "not sorted by
 InitIndex!", function CreateDesignation,
 include/clang/Parse/Designator.h:206
In-Reply-To: <bug-3001-206@http.llvm.org/bugs/>
Message-ID: <200811030928.mA39SqJQ021989@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3001


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-03 03:28:51 ---
Fixed, thanks!
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081103/008817.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 cs.uiuc.edu  Mon Nov  3 10:05:55 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Nov 2008 10:05:55 -0600
Subject: [LLVMbugs] [Bug 3005] partial specialization fails on sqlite3
In-Reply-To: <bug-3005-206@http.llvm.org/bugs/>
Message-ID: <200811031605.mA3G5tnc009481@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3005


Andrew Lenharth <alenhar2 at uiuc.edu> changed:

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




--- Comment #1 from Andrew Lenharth <alenhar2 at uiuc.edu>  2008-11-03 10:05:54 ---
fixed in r58615


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov  3 13:38:21 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Nov 2008 13:38:21 -0600
Subject: [LLVMbugs] [Bug 2902] -std-compile-opt produces a CFG w/ infinite
	loop
In-Reply-To: <bug-2902-206@http.llvm.org/bugs/>
Message-ID: <200811031938.mA3JcLuv023140@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2902


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #3 from Devang Patel <dpatel at apple.com>  2008-11-03 13:38:20 ---
Fixed.
Committed revision 58631.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov  3 16:56:03 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Nov 2008 16:56:03 -0600
Subject: [LLVMbugs] [Bug 2908] C binding doesn't bind VICmp, VFCmp,
 ExtractValue, InsertValue
In-Reply-To: <bug-2908-206@http.llvm.org/bugs/>
Message-ID: <200811032256.mA3Mu3k8002591@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2908


Dan Gohman <gohman at apple.com> changed:

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




--- Comment #2 from Dan Gohman <gohman at apple.com>  2008-11-03 16:56:03 ---
I applied the portion of the patch for extractvalue and insertvalue. Thanks!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081103/069321.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 cs.uiuc.edu  Mon Nov  3 21:53:58 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Nov 2008 21:53:58 -0600
Subject: [LLVMbugs] [Bug 2982] ice: This target-independent node should have
 been selected!
In-Reply-To: <bug-2982-206@http.llvm.org/bugs/>
Message-ID: <200811040353.mA43rw1B012969@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2982


Bill Wendling <isanbard at gmail.com> changed:

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




--- Comment #2 from Bill Wendling <isanbard at gmail.com>  2008-11-03 21:53:57 ---
Dan fixed this with this patch:

http://llvm.org/viewvc/llvm-project?rev=58655&view=rev


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov  3 23:45:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 3 Nov 2008 23:45:53 -0600
Subject: [LLVMbugs] [Bug 3011] New: assert fail on vector select on ppc32
Message-ID: <bug-3011-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3011

           Summary: assert fail on vector select on ppc32
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: PowerPC
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


The test for PR2504 fails on PPC/Linux:

FAIL: /home/buildbot-llvm/1/llvm-ppc/build/test/CodeGen/Generic/select-cc.ll
for PR2504
Failed with signal(SIGABRT) at line 1
while running:  llvm-as <
/home/buildbot-llvm/1/llvm-ppc/build/test/CodeGen/Generic/select-cc.ll | llc
ScalarizeVectorResult #0: 0x11080798: v1f64 = select_cc 0x1107fd80, 0x1107fc70,
0x11080688, 0x11080248, 0x110801c0
llc: LegalizeVectorTypes.cpp:41: void
llvm::DAGTypeLegalizer::ScalarizeVectorResult(llvm::SDNode*, unsigned int):
Assertion `0 && "Do not know how to scalarize the result of this operator!"'
failed.
llc[0x10e84308]
llc[0x10e844d4]
[0x100344]
[0xbf871188]
/lib/libc.so.6(abort+0x260)[0xfc0bca0]
/lib/libc.so.6(__assert_fail+0xd8)[0xfc00fd8]
llc[0x10aa8e14]
llc[0x10aa1aec]
llc(_ZN4llvm12SelectionDAG13LegalizeTypesEv+0x34)[0x10aa2444]
llc(_ZN4llvm16SelectionDAGISel17CodeGenAndEmitDAGEv+0x560)[0x10a2e7e0]
llc(_ZN4llvm16SelectionDAGISel16SelectBasicBlockEPNS_10BasicBlockENS_14ilist_iteratorINS_11InstructionEEES5_+0x324)[0x10a31228]
llc(_ZN4llvm16SelectionDAGISel20SelectAllBasicBlocksERNS_8FunctionERNS_15MachineFunctionEPNS_17MachineModuleInfoERKNS_15TargetInstrInfoE+0xa8c)[0x10a31e48]
llc(_ZN4llvm16SelectionDAGISel13runOnFunctionERNS_8FunctionE+0x3c0)[0x10a32e90]
llc[0x1068bd1c]
llc(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0x154)[0x10dd2b98]
llc(_ZN4llvm23FunctionPassManagerImpl3runERNS_8FunctionE+0x90)[0x10dd3a90]
llc(_ZN4llvm19FunctionPassManager3runERNS_8FunctionE+0xc0)[0x10dd3cc4]
llc(main+0xbb0)[0x103bc678]
/lib/libc.so.6[0xfbf1704]
/lib/libc.so.6[0xfbf18c0]


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov  4 04:28:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Nov 2008 04:28:29 -0600
Subject: [LLVMbugs] [Bug 3012] New: ActOnClassMessage assertion with some
 invalid obj-c file
Message-ID: <bug-3012-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3012

           Summary: ActOnClassMessage assertion with some invalid obj-c file
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: devlists at shadowlab.org
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2178)
 --> (http://llvm.org/bugs/attachment.cgi?id=2178)
erronous obj-c file

This is a case where clang abort instead of reporting incorrect syntax (see
attachement).


--------------------
clang rev 58688:

clang stream.m
    Assertion failed: (ClassDecl && "missing interface declaration"), function
ActOnClassMessage, file SemaExprObjC.cpp, line 226.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov  4 10:16:58 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Nov 2008 10:16:58 -0600
Subject: [LLVMbugs] [Bug 3013] New: unhelpful error message on missing
 terminator instruction
Message-ID: <bug-3013-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3013

           Summary: unhelpful error message on missing terminator
                    instruction
           Product: tools
           Version: 2.3
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-as
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: dwashington at gmx.net
                CC: llvmbugs at cs.uiuc.edu


If I try to assemble an LLVM IR file with a block missing a terminator
instruction, like the following example:

define i32 @main() {
    br i1 true, label %foo, label %bar

foo:
    add i32 1, 2
    br label %5

bar:
    ret i32 0
}

i get an error message like the following:

 llvm-as: no-terminator.ll:11,0: Undefined reference to label

This does not in any way point me to the missing terminator instruction for the
"foo" block, although this is the actual cause of the error. llvm-as should
ideally detect the missing terminator and report that instead, or at least in
addition, to the above message.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov  4 10:35:38 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Nov 2008 10:35:38 -0600
Subject: [LLVMbugs] [Bug 3013] unhelpful error message on missing terminator
	instruction
In-Reply-To: <bug-3013-206@http.llvm.org/bugs/>
Message-ID: <200811041635.mA4GZcWH015294@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3013


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-04 10:35:36 ---
Yep, unfortunately the current .ll parser has many gross problems.  The fix is
to rewrite the thing, which will provide many benefits.  However, this won't be
done until I get time to do so.  When that happens, this bug will be fixed, so
I'm just going to close this as "WONTFIX" meaning that we aren't going to fix
it in the current parser.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov  4 11:32:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Nov 2008 11:32:12 -0600
Subject: [LLVMbugs] [Bug 3011] assert fail on vector select on ppc32
In-Reply-To: <bug-3011-206@http.llvm.org/bugs/>
Message-ID: <200811041732.mA4HWCfZ017281@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3011


Duncan Sands <baldrick at free.fr> changed:

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




--- Comment #1 from Duncan Sands <baldrick at free.fr>  2008-11-04 11:32:09 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081103/069365.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 cs.uiuc.edu  Tue Nov  4 19:40:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 4 Nov 2008 19:40:41 -0600
Subject: [LLVMbugs] [Bug 2945] always-inline removes internal functions w/ o
 always inline attribute
In-Reply-To: <bug-2945-206@http.llvm.org/bugs/>
Message-ID: <200811050140.mA51efiq001963@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2945


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #4 from Devang Patel <dpatel at apple.com>  2008-11-04 19:40:40 ---
In general, never inline does not mean do not remove!.  However, always inliner
is special.
Fixed.
Committed revision 58744.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov  5 17:40:02 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Nov 2008 17:40:02 -0600
Subject: [LLVMbugs] [Bug 3015] New: Failure with ModulePass -> FunctionPass
 -> AliasAnalysis pass chain
Message-ID: <bug-3015-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3015

           Summary: Failure with ModulePass -> FunctionPass -> AliasAnalysis
                    pass chain
           Product: libraries
           Version: 2.4
          Platform: Macintosh
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: Nathan.Keynes at sun.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2181)
 --> (http://llvm.org/bugs/attachment.cgi?id=2181)
Test case using MemoryDependenceAnalysis

Crashes when adding the ModulePass to the PassManager with the following:

Assertion failed: (0 && "ERROR: Bad TargetData ctor used.  " "Tool did not
specify a TargetData to use?"), function TargetData, file
/Users/nkeynes/src/llvm-2.3/include/llvm/Target/TargetData.h, line 114.

A valid TargetData is in the PassManager, but for some reason this is not being
picked up and it attempts to create a new TargetData (which unsurprisingly
fails). The simplest way to reproduce is to require MemoryDependenceAnalysis
from a ModulePass (see attached test case)

I'm not certain if this is a bug per se or an unsupported configuration, but I
can't find anything in the pass documentation to suggest that this is not
supposed to work.

Backtrace:
#0  0x90940b9e in __kill ()
#1  0x90940b91 in kill$UNIX2003 ()
#2  0x909b7ec2 in raise ()
#3  0x909c747f in abort ()
#4  0x909b9063 in __assert_rtn ()
#5  0x00014354 in llvm::TargetData::TargetData (this=0x400a30) at
Target/TargetData.h:113
#6  0x00014365 in llvm::TargetData::TargetData (this=0x400a30) at
Target/TargetData.h:116
#7  0x00014385 in llvm::callDefaultCtor<llvm::TargetData> () at
PassSupport.h:164
#8  0x000a5551 in llvm::PassInfo::createPass (this=0x123a20) at
PassSupport.h:100
#9  0x000a270b in llvm::PMTopLevelManager::schedulePass (this=0x40098c,
P=0x400a00) at PassManager.cpp:444
#10 0x000a274f in llvm::PMTopLevelManager::schedulePass (this=0x40098c,
P=0x409630) at PassManager.cpp:449
#11 0x000aa751 in llvm::FunctionPassManagerImpl::add (this=0x400920,
P=0x409630) at PassManager.cpp:134
#12 0x000a5175 in llvm::MPPassManager::addLowerLevelRequiredPass
(this=0x400880, P=0x4097e0, RequiredPass=0x409630) at PassManager.cpp:1283
#13 0x000a47ba in llvm::PMDataManager::add (this=0x400898, P=0x4097e0,
ProcessAnalysis=true) at PassManager.cpp:742
#14 0x000a48b4 in llvm::ModulePass::assignPassManager (this=0x4097e0,
PMS=@0x4013a0, PreferredType=llvm::PMT_Unknown) at PassManager.cpp:1453
#15 0x000b06a7 in llvm::PassManagerImpl::addTopLevelPass (this=0x401330,
P=0x4097e0) at PassManager.cpp:298
#16 0x000a27a7 in llvm::PMTopLevelManager::schedulePass (this=0x40139c,
P=0x4097e0) at PassManager.cpp:456
#17 0x000aa733 in llvm::PassManagerImpl::add (this=0x401330, P=0x4097e0) at
PassManager.cpp:274
#18 0x000a27d3 in llvm::PassManager::add (this=0xbfffebfc, P=0x4097e0) at
PassManager.cpp:1347
#19 0x00002462 in main (argc=2, argv=0xbfffec38) at TestPass.cpp:41


Reproducible with 2.3 and 2.4, on Solaris and OS X, not dependent on any
particular bytecode.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov  5 19:00:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Nov 2008 19:00:39 -0600
Subject: [LLVMbugs] [Bug 2990] Requiring a ModulePass and InstructionNamerID
 causes PassManager to infloop
In-Reply-To: <bug-2990-206@http.llvm.org/bugs/>
Message-ID: <200811060100.mA610d6L024155@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2990


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #4 from Devang Patel <dpatel at apple.com>  2008-11-05 19:00:38 ---
Thx!
Fixed Committed revision 58787.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov  5 19:45:07 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 5 Nov 2008 19:45:07 -0600
Subject: [LLVMbugs] [Bug 3016] New: null ptr dereference
Message-ID: <bug-3016-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3016

           Summary: null ptr dereference
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r58787 on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp56$ valgrind --trace-children=yes llvm-gcc -O3
small.c
==2041== Memcheck, a memory error detector.
==2041== Copyright (C) 2002-2007, and GNU GPL'd, by Julian Seward et al.
==2041== Using LibVEX rev 1804, a library for dynamic binary translation.
==2041== Copyright (C) 2004-2007, and GNU GPL'd, by OpenWorks LLP.
==2041== Using valgrind-3.3.0, a dynamic binary instrumentation framework.
==2041== Copyright (C) 2000-2007, and GNU GPL'd, by Julian Seward et al.
==2041== For more details, rerun with: -v
==2041== 
==2042== Memcheck, a memory error detector.
==2042== Copyright (C) 2002-2007, and GNU GPL'd, by Julian Seward et al.
==2042== Using LibVEX rev 1804, a library for dynamic binary translation.
==2042== Copyright (C) 2004-2007, and GNU GPL'd, by OpenWorks LLP.
==2042== Using valgrind-3.3.0, a dynamic binary instrumentation framework.
==2042== Copyright (C) 2000-2007, and GNU GPL'd, by Julian Seward et al.
==2042== For more details, rerun with: -v
==2042== 
==2042== Invalid read of size 4
==2042==    at 0x8A24304: (anonymous
namespace)::LCSSA::GetValueForBlock(llvm::DomTreeNodeBase<llvm::BasicBlock>*,
llvm::Instruction*, llvm::DenseMap<llvm::DomTreeNodeBase<llvm::BasicBlock>*,
llvm::Value*, llvm::DenseMapInfo<llvm::DomTreeNodeBase<llvm::BasicBlock>*>,
llvm::DenseMapInfo<llvm::Value*> >&) (in
/home/regehr/libexec/gcc/i686-pc-linux-gnu/4.2.1/cc1)
==2042==  Address 0x0 is not stack'd, malloc'd or (recently) free'd
small.c:92: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.
==2042== 
==2042== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 21 from 1)
==2042== malloc/free: in use at exit: 461,015 bytes in 1,757 blocks.
==2042== malloc/free: 10,741 allocs, 8,984 frees, 1,435,760 bytes allocated.
==2042== For counts of detected errors, rerun with: -v
==2042== searching for pointers to 1,757 not-freed blocks.
==2042== checked 1,675,400 bytes.
==2042== 
==2042== LEAK SUMMARY:
==2042==    definitely lost: 0 bytes in 0 blocks.
==2042==      possibly lost: 351 bytes in 10 blocks.
==2042==    still reachable: 460,664 bytes in 1,747 blocks.
==2042==         suppressed: 0 bytes in 0 blocks.
==2042== Rerun with --leak-check=full to see details of leaked memory.
==2041== 
==2041== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 21 from 1)
==2041== malloc/free: in use at exit: 18,634 bytes in 73 blocks.
==2041== malloc/free: 479 allocs, 406 frees, 40,652 bytes allocated.
==2041== For counts of detected errors, rerun with: -v
==2041== searching for pointers to 73 not-freed blocks.
==2041== checked 145,304 bytes.
==2041== 
==2041== LEAK SUMMARY:
==2041==    definitely lost: 4,224 bytes in 32 blocks.
==2041==      possibly lost: 0 bytes in 0 blocks.
==2041==    still reachable: 14,410 bytes in 41 blocks.
==2041==         suppressed: 0 bytes in 0 blocks.
==2041== Rerun with --leak-check=full to see details of leaked memory.

regehr at john-home:~/volatile/tmp56$ cat small.c

int rshift_u_s (int left, int right)
{
  return left >> right;
}

int mod_rhs (int rhs)
{
}

int g_35;
int g_37;
int g_148;
int g_155;
int g_194;
int g_203;
int g_307;

int func_107 (short p_108);
int func_135 (int p_136, int p_137, int p_138,
                   short p_139, int p_140, int p_141);

int func_4 (int p_6, int p_7)
{
}

int func_89 (short p_90, int p_91, int p_92)
{
}

int func_105 (short p_106)
{
  func_107 (g_307 && 1 & p_106);
}

int func_107 (short p_108)
{
  int l_109;
  int l_111;
  int l_271;
  for (l_109 = 0; l_109 >= 0; l_109 += 0)
    {
      int l_280;
      if (rshift_u_s (p_108, 0))
        return 1;
      int l_255;
      func_129 (1 | l_255);
      for (l_111 = 1; l_111 == 0; l_111 += 1)
        {
          if (func_16 (0xA4408C10L, g_37, 0 || p_108, func_4) /
              div_rhs (lshift_u_s (l_280, 1)) | func_89(0,0,0) > rshift_s_s
(l_271,
                                                                      1))
            for (1; 1; p_108 += 1)
              {
              }
        }
    }
}

int func_123 (short p_124, int p_126)
{
}

int func_129 (int p_130)
{
  int l_133;
  int l_134;
  int l_207;
  l_133 =
    (1 %
     mod_rhs (func_135
              (l_134, 1, 1, 1, g_194,
               1 >= l_207 * p_130 * p_130 ^ 1 >= g_203) * 0xFA58L));
}

int func_135 (int p_136, int p_137, int p_138, short p_139,
          int p_140, int p_141)
{
  int l_143;
  if (g_35)
    {
      int l_168;
      for (l_143 = 0; l_143 > -29; l_143 -= 1)
        {
          if (p_140)
            g_148 = 1;
          if (func_123 (l_168, 1))
            {
              g_155 = g_148;
            }
        }
    }
}


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

From bugzilla-daemon at cs.uiuc.edu  Thu Nov  6 13:49:03 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Nov 2008 13:49:03 -0600
Subject: [LLVMbugs] [Bug 3015] Failure with ModulePass -> FunctionPass ->
 AliasAnalysis pass chain
In-Reply-To: <bug-3015-206@http.llvm.org/bugs/>
Message-ID: <200811061949.mA6Jn3Br006992@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3015


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #1 from Devang Patel <dpatel at apple.com>  2008-11-06 13:49:02 ---
This not supported at the time. I updated documentation. 
Committed revision 58813.
It'd be nice to catch this through an assertion though. But it is not straight
forward right now.


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

From bugzilla-daemon at cs.uiuc.edu  Thu Nov  6 17:55:07 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Nov 2008 17:55:07 -0600
Subject: [LLVMbugs] [Bug 3017] New: dedup (from the PARSEC benchmark)
 segfaults when compiled with LLVM 2.4 on Linux/x86-64
Message-ID: <bug-3017-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3017

           Summary: dedup (from the PARSEC benchmark) segfaults when
                    compiled with LLVM 2.4 on Linux/x86-64
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: resistor at mac.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2184)
 --> (http://llvm.org/bugs/attachment.cgi?id=2184)
Testcase

See the attached tarball for source and instructions to compile/run the
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 cs.uiuc.edu  Thu Nov  6 21:49:16 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 6 Nov 2008 21:49:16 -0600
Subject: [LLVMbugs] [Bug 3018] New: likely integer miscompilation
Message-ID: <bug-3018-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3018

           Summary: likely integer miscompilation
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r58823 on Ubuntu Hardy on x86.

Probably not the same bug as PR 2999 since here it looks wrong at -O0.

llvm-gcc valgrinds clean at -O0.

Behavior seen here is the same as 2.3 but different from 2.[0,1,2].

regehr at john-home:~/volatile/tmp57$ llvm-gcc -O0 small.c -o small
regehr at john-home:~/volatile/tmp57$ ./small
hi
regehr at john-home:~/volatile/tmp57$ llvm-gcc -O1 small.c -o small
regehr at john-home:~/volatile/tmp57$ ./small
regehr at john-home:~/volatile/tmp57$ cat small.c
#include <stdio.h>

static int
foo (int rhs)
{
  return rhs;
}

int func_17 (signed char p_21);
int func_17 (signed char p_21)
{
  return 1;
}

int func_104 (void);
int func_104 (void)
{
  return 1;
}

void func_1 (void);
void func_1 (void)
{
  int l_359 = 0;
  if ((func_17 (1 % foo (1))) -
      ((((func_104 () >= 1) > -1L) || 0) | (0x1B90L & l_359) + 1)) {
    printf ("hi\n");
  }
}

int main (void)
{
  func_1 ();
  return 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 cs.uiuc.edu  Fri Nov  7 07:52:55 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Nov 2008 07:52:55 -0600
Subject: [LLVMbugs] [Bug 3020] New: Crash during AST destruction
Message-ID: <bug-3020-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3020

           Summary: Crash during AST destruction
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: AST
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: akyrtzi at gmail.com
                CC: kremenek at apple.com, llvmbugs at cs.uiuc.edu


Compile with -fsyntax-only -x=c++

void f() {
  class A {};
  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 cs.uiuc.edu  Fri Nov  7 12:26:40 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Nov 2008 12:26:40 -0600
Subject: [LLVMbugs] [Bug 3023] New: clang: crash on union with typeof in
	debug codegen
Message-ID: <bug-3023-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3023

           Summary: clang: crash on union with typeof in debug codegen
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sanxiyn at gmail.com
                CC: llvmbugs at cs.uiuc.edu


SVN r58855.

int convert(int i) {
  union { typeof(i) in; int out; } u;
  u.in = i;
  return u.out;
}

$ clang -emit-llvm test.c
$ clang -emit-llvm -g test.c
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 cs.uiuc.edu  Fri Nov  7 16:09:15 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Nov 2008 16:09:15 -0600
Subject: [LLVMbugs] [Bug 3024] New: possible volatile bug
Message-ID: <bug-3024-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3024

           Summary: possible volatile bug
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


I apologize for the poor timing of my recent bug reports!  Unfortunately I'm at
the mercy of the random number generator here.

This is seen on r58855 on Ubuntu Hardy on x86.

It looks like a backend bug: the bitcode accesses the volatile the right number
of times at all optimization levels, but the x86 doesn't.

The code is below.  The right number of reads from g_107 is 2.  At -O1,
llvm-gcc produces x86 that reads this variable only once.

#include <stdio.h>

// VOLATILE GLOBAL g_107
volatile unsigned int g_107;
int x;

static inline unsigned int
rshift_u_u(unsigned int left, unsigned int right)
{
        if (right >= sizeof(unsigned int)*8) {
                /* Avoid undefined behavior. */
                return left;
        }
        return left >> right;
}

void func_95 (unsigned int p_97);
void func_95 (unsigned int p_97)
{
  for (; p_97 < -7; --p_97)
    if (rshift_u_u (g_107, (-9L * p_97)))
      x++;
}

int
main (void)
{
  func_95 (1);
  printf ("checksum = 0\n");
  return 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 cs.uiuc.edu  Fri Nov  7 17:01:04 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Nov 2008 17:01:04 -0600
Subject: [LLVMbugs] [Bug 642] [PowerPC] Miscompilation of some ordered
	comparisons
In-Reply-To: <bug-642-206@http.llvm.org/bugs/>
Message-ID: <200811072301.mA7N14AP005872@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=642


Dale Johannesen <dalej 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 cs.uiuc.edu  Fri Nov  7 22:52:18 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 7 Nov 2008 22:52:18 -0600
Subject: [LLVMbugs] [Bug 3023] clang: crash on union with typeof in debug
	codegen
In-Reply-To: <bug-3023-206@http.llvm.org/bugs/>
Message-ID: <200811080452.mA84qItd017096@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3023


Daniel Dunbar <daniel at zuster.org> changed:

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




--- Comment #1 from Daniel Dunbar <daniel at zuster.org>  2008-11-07 22:52:18 ---
Crash fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081103/008970.html,
although generated debug info is still wrong.

Closing as "improving debug info" is a known macroscopic 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 cs.uiuc.edu  Sat Nov  8 00:50:05 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Nov 2008 00:50:05 -0600
Subject: [LLVMbugs] [Bug 3025] New: llvmc2' s TableGen miscompiles $CALL and
 $ENV when used as the command name
Message-ID: <bug-3025-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3025

           Summary: llvmc2's TableGen miscompiles $CALL and $ENV when used
                    as the command name
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: pcwalton at cs.ucla.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2186)
 --> (http://llvm.org/bugs/attachment.cgi?id=2186)
Proposed patch.

When the $CALL and $ENV are used to form the command name of a tool in llvmc2,
an automatic conversion to const char * does not occur. The attached patch
fixes this 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 cs.uiuc.edu  Sat Nov  8 01:36:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Nov 2008 01:36:01 -0600
Subject: [LLVMbugs] [Bug 3026] New: llvmc2's TableGen doesn't escape dashes
	properly
Message-ID: <bug-3026-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3026

           Summary: llvmc2's TableGen doesn't escape dashes properly
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: pcwalton at cs.ucla.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2187)
 --> (http://llvm.org/bugs/attachment.cgi?id=2187)
Proposed patch.

llvmc2's TableGen logic for escaping special characters in argument names is
wrong, probably the result of a botched copy and paste job. The attached patch
fixes this 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 cs.uiuc.edu  Sat Nov  8 04:21:05 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Nov 2008 04:21:05 -0600
Subject: [LLVMbugs] [Bug 3026] llvmc2's TableGen doesn't escape dashes
	properly
In-Reply-To: <bug-3026-206@http.llvm.org/bugs/>
Message-ID: <200811081021.mA8AL5AH005257@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3026


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2008-11-08 04:21:05 ---
Applied here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081103/069592.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 cs.uiuc.edu  Sat Nov  8 13:31:02 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Nov 2008 13:31:02 -0600
Subject: [LLVMbugs] [Bug 2946] CellSPU Error: unknown pseudo-op: `.cstring'
In-Reply-To: <bug-2946-206@http.llvm.org/bugs/>
Message-ID: <200811081931.mA8JV2in022739@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2946


Scott Michel <scooter.phd 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 cs.uiuc.edu  Sat Nov  8 13:32:04 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Nov 2008 13:32:04 -0600
Subject: [LLVMbugs] [Bug 2616] [CellSPU] Code generation failure
In-Reply-To: <bug-2616-206@http.llvm.org/bugs/>
Message-ID: <200811081932.mA8JW4wY022785@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2616


Scott Michel <scooter.phd at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|WORKSFORME                  |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 cs.uiuc.edu  Sat Nov  8 13:46:55 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Nov 2008 13:46:55 -0600
Subject: [LLVMbugs] [Bug 3025] llvmc2' s TableGen miscompiles $CALL and $ENV
 when used as the command name
In-Reply-To: <bug-3025-206@http.llvm.org/bugs/>
Message-ID: <200811081946.mA8JktT9023246@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3025


Mikhail Glushenkov <foldr at codedgers.com> changed:

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




--- Comment #1 from Mikhail Glushenkov <foldr at codedgers.com>  2008-11-08 13:46:55 ---
Fixed in
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081103/069600.html

Thanks for your report!


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov  8 15:31:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Nov 2008 15:31:44 -0600
Subject: [LLVMbugs] [Bug 3027] New: llvm-gcc bootstrap fails: error: Link
 tests are not allowed after GCC_NO_EXECUTABLES
Message-ID: <bug-3027-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3027

           Summary: llvm-gcc bootstrap fails: error: Link tests are not
                    allowed after GCC_NO_EXECUTABLES
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


On x86_64 linux my nightly tester failed to bootstrap llvm-gcc, SVN r58921.
It bootstrapped fine yesterday.

checking for pid_t... yes
checking for library containing strerror... configure: error: Link tests are
not allowed after GCC_NO_EXECUTABLES.
make[2]: *** [configure-stage2-libiberty] Error 1
make[2]: Leaving directory `/home/edwin/llvm-bootstrap/llvm-gcc-obj'
make[1]: *** [stage2-bubble] Error 2
make[1]: Leaving directory `/home/edwin/llvm-bootstrap/llvm-gcc-obj'
make: *** [all] Error 2

My script configures like this:
../llvm-gcc4.2/configure  
--prefix=/home/edwin/llvm-bootstrap/llvm-gcc-obj/../install
--program-prefix=llvm- --enable-llvm=/home/edwin/llvm-bootstrap/llvm-obj/
--disable-multilib --enable-languages=c,c++

libiberty/config.log says:

configure:6117: checking for library containing strerror
configure:6125: error: Link tests are not allowed after GCC_NO_EXECUTABLES.

## ---------------- ##


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov  8 21:40:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Nov 2008 21:40:53 -0600
Subject: [LLVMbugs] [Bug 3028] New: Assertion `i < NumOperands &&
 "getOperand() out of range!"' failed
Message-ID: <bug-3028-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3028

           Summary: Assertion `i < NumOperands && "getOperand() out of
                    range!"' failed
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r58928 on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp58$ llvm-gcc -O2 small.c
cc1: /home/regehr/llvm/include/llvm/User.h:93: llvm::Value*
llvm::User::getOperand(unsigned int) const: Assertion `i < NumOperands &&
"getOperand() out of range!"' failed.
small.c: In function ???func_20???:
small.c:11: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.

regehr at john-home:~/volatile/tmp58$ cat small.c

int g_187;
int g_204;
int g_434;

int func_89 (void)
{
  return 1;
}

void func_20 (int p_22)
{
  if (1 & p_22 | g_204 & (1 < g_187) - func_89 ())
      g_434 = 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 cs.uiuc.edu  Sat Nov  8 22:27:14 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Nov 2008 22:27:14 -0600
Subject: [LLVMbugs] [Bug 3021] instrcombine causes miscompilation with fcmp
	and itofp
In-Reply-To: <bug-3021-206@http.llvm.org/bugs/>
Message-ID: <200811090427.mA94REXR007464@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3021


Bill Wendling <isanbard at gmail.com> changed:

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




--- Comment #2 from Bill Wendling <isanbard at gmail.com>  2008-11-08 22:27:14 ---
Fixed here:
http://llvm.org/viewvc/llvm-project?rev=58929&view=rev


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov  8 23:34:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 8 Nov 2008 23:34:01 -0600
Subject: [LLVMbugs] [Bug 3029] New: possible integer miscompilation
Message-ID: <bug-3029-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3029

           Summary: possible integer miscompilation
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Perhaps this is the same as 2999 or 3018, I cannot tell.

This is seen on r58928 on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp59$ llvm-gcc -O1 small.c -o small
regehr at john-home:~/volatile/tmp59$ ./small
1
regehr at john-home:~/volatile/tmp59$ llvm-gcc -O2 small.c -o small
regehr at john-home:~/volatile/tmp59$ ./small
0
regehr at john-home:~/volatile/tmp59$ cat small.c

#include <stdio.h>

unsigned g_102;
volatile unsigned g_138;
int g_188;
unsigned g_207;

int func_119 (void);
int func_119 (void)
{
  return g_138;
}

void func_110 (int p_111);
void func_110 (int p_111)
{
  unsigned l_192 = 0xA2471E60L;
  if (l_192 < g_188) {
    func_119 ();
  }
  else {
    g_207 = 1;
  }
}

void func_34 (void);
void func_34 (void)
{
  for (g_188 = 0; g_188 < 2; g_188++) {
    func_110 (1);
  }
}

int main (void)
{
  func_34 ();
  printf ("%d\n", g_207);
  return 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 cs.uiuc.edu  Sun Nov  9 01:30:26 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Nov 2008 01:30:26 -0600
Subject: [LLVMbugs] [Bug 3030] New: format not a string literal and no
	format arguments
Message-ID: <bug-3030-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3030

           Summary: format not a string literal and no format arguments
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


I'm seeing this warning:

  $ cat test.c
  #include <stdio.h>

  void f(const char *buf) {
    printf(buf);
  }
  $ llvm-gcc test.c -c
  test.c: In function ???f???:
  test.c:4: warning: format not a string literal and no format arguments
  $ gcc test.c -c
  $ llvm-gcc test.c -c -Wno-format
  $ 

Please disable that warning by default on non-Darwin platforms.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov  9 02:43:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Nov 2008 02:43:12 -0600
Subject: [LLVMbugs] [Bug 3030] format not a string literal and no format
	arguments
In-Reply-To: <bug-3030-206@http.llvm.org/bugs/>
Message-ID: <200811090843.mA98hC8N017691@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3030


Nick Lewycky <nicholas at mxc.ca> 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 cs.uiuc.edu  Sun Nov  9 07:36:23 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Nov 2008 07:36:23 -0600
Subject: [LLVMbugs] [Bug 3031] New: Duplicate / confusing diagnostic
Message-ID: <bug-3031-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3031

           Summary: Duplicate / confusing diagnostic
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


Given

   int (x), ;

clang gives

/tmp/foo.c:1:10: error: expected identifier or '('
int (x), ;
         ^
/tmp/foo.c:1:1: error: declarator requires an identifier
int (x), ;
^~~
2 diagnostics generated.

The second diagnostic is confusing and / or redundant.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov  9 09:12:20 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Nov 2008 09:12:20 -0600
Subject: [LLVMbugs] [Bug 3030] format not a string literal and no format
	arguments
In-Reply-To: <bug-3030-206@http.llvm.org/bugs/>
Message-ID: <200811091512.mA9FCKgM000622@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3030


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #4 from Nick Lewycky <nicholas at mxc.ca>  2008-11-09 09:12:14 ---
Apparently that change has the unwelcome side-effect of breaking 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 cs.uiuc.edu  Sun Nov  9 11:40:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Nov 2008 11:40:48 -0600
Subject: [LLVMbugs] [Bug 3032] New: Switch instruction gives an assertion
 error in SelectionDAGLowering:: handleBTSplitSwitchCase
Message-ID: <bug-3032-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3032

           Summary: Switch instruction gives an assertion error in
                    SelectionDAGLowering::handleBTSplitSwitchCase
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicolas.geoffray at lip6.fr
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2190)
 --> (http://llvm.org/bugs/attachment.cgi?id=2190)
Switch error

The following test case gives an assertion error:

$ ./Release/bin/llvm-as < switch.ll | ./Release/bin/llc -f

llc: SelectionDAGBuild.cpp:1740: bool
llvm::SelectionDAGLowering::handleBTSplitSwitchCase(llvm::SelectionDAGLowering::CaseRec&,
std::vector<llvm::SelectionDAGLowering::CaseRec,
std::allocator<llvm::SelectionDAGLowering::CaseRec> >&, llvm::Value*,
llvm::MachineBasicBlock*): Assertion `(RBegin-LEnd>=1) && "Invalid case
distance"' failed.
./Release/bin/llc[0x89fbf08]
[0xb7ef2420]
[0xb7ef2410]
/lib/tls/i686/cmov/libc.so.6(gsignal+0x55)[0xb7c79085]
/lib/tls/i686/cmov/libc.so.6(abort+0x101)[0xb7c7aa01]
/lib/tls/i686/cmov/libc.so.6(__assert_fail+0xee)[0xb7c7210e]
./Release/bin/llc(_ZN4llvm20SelectionDAGLowering23handleBTSplitSwitchCaseERNS0_7CaseRecERSt6vectorIS1_SaIS1_EEPNS_5ValueEPNS_17MachineBasicBlockE+0x90f)[0x867183f]
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 cs.uiuc.edu  Sun Nov  9 12:07:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Nov 2008 12:07:25 -0600
Subject: [LLVMbugs] [Bug 3033] New: libtasn1 1.5 test failure
Message-ID: <bug-3033-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3033

           Summary: libtasn1 1.5 test failure
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sanxiyn at gmail.com
                CC: llvmbugs at cs.uiuc.edu


libtasn1 1.5's test suite fails when compiled with ccc.

How to reproduce:

wget http://ftp.gnu.org/gnu/gnutls/libtasn1-1.5.tar.gz
tar zxf libtasn1-1.5.tar.gz
cd libtasn1-1.5
CC=gcc ./configure; make; make check &>gcc.log
make clean
CC=ccc ./configure; make; make check &>ccc.log

Cleaned up diff of gcc.log and ccc.log is 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 cs.uiuc.edu  Sun Nov  9 14:01:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Nov 2008 14:01:48 -0600
Subject: [LLVMbugs] [Bug 3034] New: Document clang C status
Message-ID: <bug-3034-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3034

           Summary: Document clang C status
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Documentation
        AssignedTo: daniel at zuster.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


There is a C++ status page, it would be nice to have a C status.

For example:
- blockers on x86_64: var_arg not implemented in LLVM


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov  9 17:19:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Nov 2008 17:19:39 -0600
Subject: [LLVMbugs] [Bug 3028] Assertion `i < NumOperands && "getOperand()
 out of range!"' failed
In-Reply-To: <bug-3028-206@http.llvm.org/bugs/>
Message-ID: <200811092319.mA9NJdtW017245@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3028


Bill Wendling <isanbard at gmail.com> changed:

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




--- Comment #5 from Bill Wendling <isanbard at gmail.com>  2008-11-09 17:19:39 ---
Fixed:
http://llvm.org/viewvc/llvm-project?rev=58946&view=rev

I believe that this is the correct approach. For this case, the code won't be
executed -- which is correct.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov  9 18:45:02 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Nov 2008 18:45:02 -0600
Subject: [LLVMbugs] [Bug 3034] Document clang C status
In-Reply-To: <bug-3034-206@http.llvm.org/bugs/>
Message-ID: <200811100045.mAA0j2EK020119@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3034


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-09 18:45:02 ---
I don't see how this is doable.  There is no specific list of "once this is
done, we are complete" and bugzilla is great for tracking specific problems.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov  9 19:29:21 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 9 Nov 2008 19:29:21 -0600
Subject: [LLVMbugs] [Bug 3034] Document clang C status
In-Reply-To: <bug-3034-206@http.llvm.org/bugs/>
Message-ID: <200811100129.mAA1TLnk021685@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3034


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|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 cs.uiuc.edu  Mon Nov 10 10:04:22 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Nov 2008 10:04:22 -0600
Subject: [LLVMbugs] [Bug 3036] New: CellSPU assert 'false && "Invalid value
	type!"'
Message-ID: <bug-3036-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3036

           Summary: CellSPU assert 'false && "Invalid value type!"'
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: bwatt at us.ibm.com
                CC: llvmbugs at cs.uiuc.edu


Given the following C-language, test_float4.c, which is compiled with clang

typedef float float4 __attribute__((ext_vector_type(4))); 

int check(float4 a) {
  return a.x == 0.0f; 
}

Or given the following IR (after optimization)

; ModuleID = 'test-float4'
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-v128:128:128"
target triple = "ppc-unknown-linux-gnu"

define i32 @check(<4 x float> %a) nounwind readnone {
entry:
        %0 = extractelement <4 x float> %a, i32 0               ; <float>
[#uses=1]
        %1 = fcmp oeq float %0, 0.000000e+00            ; <i1> [#uses=1]
        %2 = zext i1 %1 to i32          ; <i32> [#uses=1]
        ret i32 %2
}

The CellSPU backend generates an assert failure (llc -march=cellspu
test-float4)

llc: SPUISelLowering.cpp:2070: llvm::SDValue
LowerEXTRACT_VECTOR_ELT(llvm::SDValue, llvm::SelectionDAG&): Assertion `false
&& "Invalid value type!"' failed.

I brief review of the code and it appears that the LowerEXTRACT_VECTOR_ELT
function needs cases added for MVT::f32 and MVT::f64


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 10 10:58:55 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Nov 2008 10:58:55 -0600
Subject: [LLVMbugs] [Bug 3037] New: LegalizeTypes fails on i1 =
 extractelement <n x i1>, N
Message-ID: <bug-3037-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3037

           Summary: LegalizeTypes fails on i1 = extractelement <n x i1>, N
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: pgurd at rapidmind.com
                CC: llvmbugs at cs.uiuc.edu, sdt at rapidmind.com


Created an attachment (id=2193)
 --> (http://llvm.org/bugs/attachment.cgi?id=2193)
llvm ir file which reproduces the problem

The test case fails with the message

llc: LegalizeIntegerTypes.cpp:327: llvm::SDValue
llvm::DAGTypeLegalizer::PromoteIntRes_EXTRACT_VECTOR_ELT(llvm::SDNode*):
Assertion `OldVT.isSimple() && NewVT.isSimple()' failed.

Tested on x86 with -mcpu=core2


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 10 11:39:33 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Nov 2008 11:39:33 -0600
Subject: [LLVMbugs] [Bug 2667] Legalize Types - Crash in soften float
	UINT_TO_FP result.
In-Reply-To: <bug-2667-206@http.llvm.org/bugs/>
Message-ID: <200811101739.mAAHdX9q025859@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2667


Duncan Sands <baldrick at free.fr> changed:

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




--- Comment #5 from Duncan Sands <baldrick at free.fr>  2008-11-10 11:39:32 ---
Fixed in commit 58994.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 10 14:39:19 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Nov 2008 14:39:19 -0600
Subject: [LLVMbugs] [Bug 3038] New: Failure during compile on debian linux
	32 bit
Message-ID: <bug-3038-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3038

           Summary: Failure during compile on debian linux 32 bit
           Product: tools
           Version: 2.4
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: huhlig at gmail.com
                CC: llvmbugs at cs.uiuc.edu


##################### xgcc ##############################
Using built-in specs.
Target: i686-pc-linux-gnu
Configured with: ../llvm-gcc4.2-2.4.source/configure --prefix=/opt/llvm-gcc/
--program-prefix=llvm- --enable-llvm=/home/hans/projects/llvm-2.4
--enable-languages=c,c++ : (reconfigured) ../llvm-gcc4.2-2.4.source/configure
--prefix=/opt/llvm-gcc/ --program-prefix=llvm-
--enable-llvm=/home/hans/projects/llvm-2.4 --enable-languages=c,c++
Thread model: posix
gcc version 4.2.1 (Based on Apple Inc. build 5623) (LLVM build)

###################### Error ############################

make[3]: Entering directory `/home/user/projects/llvm-gcc4.2-2.4.obj/gcc'
make \
          CFLAGS="-g -W -Wall -Wwrite-strings -Wstrict-prototypes
-Wmissing-prototypes -Wold-style-definition -Wmissing-format-attribute
-fno-common " \
          CONFIG_H="tconfig.h auto-host.h
../../llvm-gcc4.2-2.4.source/gcc/../include/ansidecl.h" TM_H="tm.h     
options.h ../../llvm-gcc4.2-2.4.source/gcc/config/i386/llvm-i386-target.h
../../llvm-gcc4.2-2.4.source/gcc/config/i386/i386.h
../../llvm-gcc4.2-2.4.source/gcc/config/i386/unix.h
../../llvm-gcc4.2-2.4.source/gcc/config/i386/att.h
../../llvm-gcc4.2-2.4.source/gcc/config/dbxelf.h
../../llvm-gcc4.2-2.4.source/gcc/config/elfos.h
../../llvm-gcc4.2-2.4.source/gcc/config/svr4.h
../../llvm-gcc4.2-2.4.source/gcc/config/linux.h
../../llvm-gcc4.2-2.4.source/gcc/config/i386/linux.h
../../llvm-gcc4.2-2.4.source/gcc/defaults.h insn-constants.h insn-flags.h
options.h" \
          INCLUDES="-I. -I. -I../../llvm-gcc4.2-2.4.source/gcc
-I../../llvm-gcc4.2-2.4.source/gcc/.
-I../../llvm-gcc4.2-2.4.source/gcc/../include
-I../../llvm-gcc4.2-2.4.source/gcc/../libcpp/include 
-I../../llvm-gcc4.2-2.4.source/gcc/../libdecnumber -I../libdecnumber
-I/home/user/projects/llvm-2.4/include -I/home/user/projects/llvm-2.4/include"
\
          MAKEOVERRIDES= \
          -f libgcc.mk all
make[4]: Entering directory `/home/user/projects/llvm-gcc4.2-2.4.obj/gcc'
/home/user/projects/llvm-gcc4.2-2.4.obj/./gcc/xgcc
-B/home/user/projects/llvm-gcc4.2-2.4.obj/./gcc/
-B/opt/llvm-gcc//i686-pc-linux-gnu/bin/ -B/opt/llvm-gcc//i686-pc-linux-gnu/lib/
-isystem /opt/llvm-gcc//i686-pc-linux-gnu/include -isystem
/opt/llvm-gcc//i686-pc-linux-gnu/sys-include  -O2  -O2 -g -O2  -DIN_GCC    -W
-Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes
-Wold-style-definition  -isystem ./include  -fPIC -g -DHAVE_GTHR_DEFAULT
-DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED  -I. -I.
-I../../llvm-gcc4.2-2.4.source/gcc -I../../llvm-gcc4.2-2.4.source/gcc/.
-I../../llvm-gcc4.2-2.4.source/gcc/../include
-I../../llvm-gcc4.2-2.4.source/gcc/../libcpp/include 
-I../../llvm-gcc4.2-2.4.source/gcc/../libdecnumber -I../libdecnumber
-I/home/user/projects/llvm-2.4/include -I/home/user/projects/llvm-2.4/include
-DL_popcountsi2 -fvisibility=hidden -DHIDE_EXPORTS -c
../../llvm-gcc4.2-2.4.source/gcc/libgcc2.c -o libgcc/./_popcountsi2.o
cc1: StringMap.cpp:177: void
llvm::StringMapImpl::RemoveKey(llvm::StringMapEntryBase*): Assertion `V == V2
&& "Didn't find key?"' failed.
../../llvm-gcc4.2-2.4.source/gcc/libgcc2.c:809: internal compiler error:
Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.
make[4]: *** [libgcc/./_popcountsi2.o] Error 1
make[4]: Leaving directory `/home/user/projects/llvm-gcc4.2-2.4.obj/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 cs.uiuc.edu  Mon Nov 10 14:48:49 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Nov 2008 14:48:49 -0600
Subject: [LLVMbugs] [Bug 3038] Failure during compile on debian linux 32 bit
In-Reply-To: <bug-3038-206@http.llvm.org/bugs/>
Message-ID: <200811102048.mAAKmnip002065@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3038


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2008-11-10 14:48:48 ---
Really looks like miscompilation:
http://llvm.org/docs/GettingStarted.html#brokengcc


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 10 19:22:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Nov 2008 19:22:11 -0600
Subject: [LLVMbugs] [Bug 3039] New: llvm-gcc42's README. LLVM has the wrong
 target triple for leopard
Message-ID: <bug-3039-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3039

           Summary: llvm-gcc42's README.LLVM has the wrong target triple for
                    leopard
           Product: Documentation
           Version: 2.4
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: General docs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: idadesub at users.sourceforge.net
                CC: llvmbugs at cs.uiuc.edu


In the just-released version of llvm-gcc42, the README.LLVM file says to use
the target triple powerpc-apple-darwin8 or i686-apple-darwin8, but on leopard
it needs to be either powerpc-apple-darwin9 or i686-apple-darwin9.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 10 19:56:04 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Nov 2008 19:56:04 -0600
Subject: [LLVMbugs] [Bug 3036] CellSPU assert 'false && "Invalid value
	type!"'
In-Reply-To: <bug-3036-206@http.llvm.org/bugs/>
Message-ID: <200811110156.mAB1u4WS015877@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3036


Scott Michel <scooter.phd at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |scooter.phd at gmail.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Scott Michel <scooter.phd at gmail.com>  2008-11-10 19:56:02 ---
Fixed. Uncovered additional argument handling bug in LowerFORMAL_ARGUMENTS as
well, which was also fixed.

(In reply to comment #0)
> Given the following C-language, test_float4.c, which is compiled with clang
> 
> typedef float float4 __attribute__((ext_vector_type(4))); 
> 
> int check(float4 a) {
>   return a.x == 0.0f; 
> }
> 
> Or given the following IR (after optimization)
> 
> ; ModuleID = 'test-float4'
> 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-v128:128:128"
> target triple = "ppc-unknown-linux-gnu"
> 
> define i32 @check(<4 x float> %a) nounwind readnone {
> entry:
>         %0 = extractelement <4 x float> %a, i32 0               ; <float>
> [#uses=1]
>         %1 = fcmp oeq float %0, 0.000000e+00            ; <i1> [#uses=1]
>         %2 = zext i1 %1 to i32          ; <i32> [#uses=1]
>         ret i32 %2
> }
> 
> The CellSPU backend generates an assert failure (llc -march=cellspu
> test-float4)
> 
> llc: SPUISelLowering.cpp:2070: llvm::SDValue
> LowerEXTRACT_VECTOR_ELT(llvm::SDValue, llvm::SelectionDAG&): Assertion `false
> && "Invalid value type!"' failed.
> 
> I brief review of the code and it appears that the LowerEXTRACT_VECTOR_ELT
> function needs cases added for MVT::f32 and MVT::f64
> 


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 10 20:12:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Nov 2008 20:12:59 -0600
Subject: [LLVMbugs] [Bug 3040] New: DAGCombiner:: visitXOR transforms
 condition code without checking legality
Message-ID: <bug-3040-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3040

           Summary: DAGCombiner::visitXOR transforms condition code without
                    checking legality
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Per summary; the following code in DAGCombiner::visitXOR is wrong:

  // fold !(x cc y) -> (x !cc y)
  if (N1C && N1C->getAPIntValue() == 1 && isSetCCEquivalent(N0, LHS, RHS, CC))
{
    bool isInt = LHS.getValueType().isInteger();
    ISD::CondCode NotCC = ISD::getSetCCInverse(cast<CondCodeSDNode>(CC)->get(),
                                               isInt);
    if (N0.getOpcode() == ISD::SETCC)
      return DAG.getSetCC(VT, LHS, RHS, NotCC);
    if (N0.getOpcode() == ISD::SELECT_CC)
      return DAG.getSelectCC(LHS, RHS,
N0.getOperand(2),N0.getOperand(3),NotCC);
    assert(0 && "Unhandled SetCC Equivalent!");
    abort();
  }

This needs to check whether the inverse opcode is legal when this is running
post-legalize.  That said, this doesn't affect any existing backends as far as
I know.

It looks like combineSelectAndUse has the same issue, but a quick search
doesn't find any other places affected similarly.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 10 23:49:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 10 Nov 2008 23:49:46 -0600
Subject: [LLVMbugs] [Bug 3029] possible integer miscompilation
In-Reply-To: <bug-3029-206@http.llvm.org/bugs/>
Message-ID: <200811110549.mAB5nkug024716@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3029


Richard Godbee <richard at godbee.net> changed:

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




--- Comment #4 from Richard Godbee <richard at godbee.net>  2008-11-10 23:49:44 ---
It looks like Devang Patel fixed this bug in r58995:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081110/069690.html

I can confirm that r59024 on OS X 10.5.5 (PPC) no longer exhibits this bug,
whereas r58990 did.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 11 00:13:38 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Nov 2008 00:13:38 -0600
Subject: [LLVMbugs] [Bug 3031] Duplicate / confusing diagnostic
In-Reply-To: <bug-3031-206@http.llvm.org/bugs/>
Message-ID: <200811110613.mAB6DclB025811@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3031


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-11 00:13:37 ---
Fixed, thanks!
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081110/009021.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 cs.uiuc.edu  Tue Nov 11 02:26:13 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Nov 2008 02:26:13 -0600
Subject: [LLVMbugs] [Bug 3040] DAGCombiner:: visitXOR transforms condition
 code without checking legality
In-Reply-To: <bug-3040-206@http.llvm.org/bugs/>
Message-ID: <200811110826.mAB8QDvc030852@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3040


Bill Wendling <isanbard at gmail.com> changed:

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




--- Comment #1 from Bill Wendling <isanbard at gmail.com>  2008-11-11 02:26:11 ---
Fixed:
http://llvm.org/viewvc/llvm-project?rev=59035&view=rev


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 11 09:30:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Nov 2008 09:30:11 -0600
Subject: [LLVMbugs] [Bug 3041] New: Disabled code should only be disabled on
 certain platforms
Message-ID: <bug-3041-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3041

           Summary: Disabled code should only be disabled on certain
                    platforms
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: orion at parsed.net
                CC: llvmbugs at cs.uiuc.edu


As you can see from this snippet of code from llvm-gcc-4.2/gcc/gcc.c, the
passing of --help to as and ld is disabled on all architectures, instead of
just the ones that don't support it.


#if 0
          /* Our assembler and linkder do not support --help.  */
          /* APPLE LOCAL end verbose help 2920964 */
          add_assembler_option ("--help", 6);
          add_linker_option ("--help", 6);
          /* APPLE LOCAL verbose help 2920964 */
#endif


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 11 09:40:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Nov 2008 09:40:53 -0600
Subject: [LLVMbugs] [Bug 3042] New: Seems to miscompile qmake with -O2
Message-ID: <bug-3042-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3042

           Summary: Seems to miscompile qmake with -O2
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: parag.warudkar at gmail.com
                CC: llvmbugs at cs.uiuc.edu


So I was trying to compile Qt 4.5 TP1
(http://trolltech.com/developer/preview-qt-4.5 ) with 2.4 release of llvm and 
Svn snapshot from llvm-gcc 4.2 taken on Nov 9. 

It all works fine unless I specify the -optimized-qmake option to configure
script for Qt. This option just enables -O2 for qmake compilation. The
compilation succeeds but then qmake starts failing with weird file not found 
errors.

All of the files exist and are having right permissions and strace of optimized
qmake shows that the files are being stat()ed with return of 0 which seems all
is ok with the files.

Needless to say if I use FSF GCC both optimized and non-optimized qmake
versions work fine and so does non-optimized llvm compiled qmake. Just
switching on -O2  breaks  something when all other factors are same - this to
my naive instinct sounds like a code generation bug someplace?

I wasn't able to do more debugging as the LLVM -g compiled qmake still doesn't
produce debug info that gdb can use but that's another problem.

Any hints as to where to look will be 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 cs.uiuc.edu  Tue Nov 11 10:39:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Nov 2008 10:39:53 -0600
Subject: [LLVMbugs] [Bug 3043] New: lli fails pointer comparison test
Message-ID: <bug-3043-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3043

           Summary: lli fails pointer comparison test
           Product: libraries
           Version: trunk
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Target-Independent JIT
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: evan.cheng at apple.com
                CC: llvmbugs at cs.uiuc.edu


>From Prakash Prabhu:


#include<stdio.h>
#include<stdlib.h>

void test();
void (*funcPtr)();

int main(int argc, char **argv) {
  funcPtr = test;
  test();
}

void test() {
  if(funcPtr == test) {
    printf("OK!\n");
  } else {
    fprintf(stderr, "Bad!\n");
    exit(1);
  }
}

$ llvm-gcc -emit-llvm -o FPtrEqTest.bc -c FPtrEqTest.c
$ llc -f FPtrEqTest.bc
$ gcc -o FPtrEqTest FPtrEqTest.s
$ ./FPtrEqTest
OK!

$ lli FPtrEqTest.bc
Bad!


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 11 13:28:23 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Nov 2008 13:28:23 -0600
Subject: [LLVMbugs] [Bug 3044] New: DAGCombiner creates an invalid
	combination
Message-ID: <bug-3044-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3044

           Summary: DAGCombiner creates an invalid combination
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alenhar2 at uiuc.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2194)
 --> (http://llvm.org/bugs/attachment.cgi?id=2194)
mulvdi3.bc

(gdb) fr 3
#3  0x0000000000cf17c6 in llvm::SelectionDAG::ReplaceAllUsesWith
(this=0x304e3e0,
    From=0x304ebf8, To=0x304eff8, UpdateListener=0x7ffff3438f10)
    at SelectionDAG.cpp:4409
4409             "Cannot use this version of ReplaceAllUsesWith!");


(gdb) call From->dump()
0x304ebf8: i64,flag = add 0x304e7f8, 0x304eff8
(gdb) call To->dump()
0x304eff8: i64,ch = load 0x304e400, 0x304eef8, 0x304e8f8 <0x3042550:8>
alignment=8


llc -march=alpha mulvdi3.bc

to reproduce.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 11 15:02:09 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Nov 2008 15:02:09 -0600
Subject: [LLVMbugs] [Bug 3045] New: 'A' constraint not handled
Message-ID: <bug-3045-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3045

           Summary: 'A' constraint not handled
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: alenhar2 at uiuc.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2195)
 --> (http://llvm.org/bugs/attachment.cgi?id=2195)
inlineasm.bc

llc -f -march=x86 inlineasm.bc
Couldn't allocate output reg for constraint '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 cs.uiuc.edu  Tue Nov 11 18:10:31 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 11 Nov 2008 18:10:31 -0600
Subject: [LLVMbugs] [Bug 3046] New: __builtin__CFStringMakeConstantString
 crashes on 64-bit systems
Message-ID: <bug-3046-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3046

           Summary: __builtin__CFStringMakeConstantString crashes on 64-bit
                    systems
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sebastian.redl at getdesigned.at
                CC: llvmbugs at cs.uiuc.edu


The layout of the internal CFString struct looks something like this:
struct {
  void *ClassType;
  unsigned int Flags;
  const char *Data;
  long Length;
}

CodeGenModule::GetAddrOfConstantCFString happily assumes that this is correct
for the LLVM Type and thus builds the corresponding 4-element vector for the
initializer.

However, padding bytes are actual members on the LLVM level. On 64-bit systems,
the 32-bit Flags member is padded so that the 64-bit Data pointer is aligned.
The actual layout of the struct on the LLVM level thus looks like this:

struct {
  void *ClassType;
  unsigned int Flags;
  char padding1;
  char padding2;
  char padding3;
  char padding4;
  const char *Data;
  long Length;
}

The function then calls llvm::ConstantStruct::get with the 4-element
initializer and the 8-element structure, leading to an assertion being
triggered in the ConstantStruct constructor in llvm/lib/VMCore/Constants.cpp on
line 411.

The clang test CodeGen/cfstring.c triggers 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 cs.uiuc.edu  Wed Nov 12 02:27:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 02:27:48 -0600
Subject: [LLVMbugs] [Bug 3044] DAGCombiner creates an invalid combination
In-Reply-To: <bug-3044-206@http.llvm.org/bugs/>
Message-ID: <200811120827.mAC8Rmq7003492@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3044


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |baldrick at free.fr
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Duncan Sands <baldrick at free.fr>  2008-11-12 02:27:48 ---
Fixed in commit 59120.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 12 02:49:57 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 02:49:57 -0600
Subject: [LLVMbugs] [Bug 3047] New: llvmc2 $ENV(...) segfaults when the
 environment variable is missing
Message-ID: <bug-3047-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3047

           Summary: llvmc2 $ENV(...) segfaults when the environment variable
                    is missing
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: pcwalton at cs.ucla.edu
                CC: llvmbugs at cs.uiuc.edu


When the environment variable in question is missing, the generated code for
llvmc2 tries to create a std::string from a NULL pointer, which causes a
segfault. The attached patch fixes this 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 cs.uiuc.edu  Wed Nov 12 02:56:31 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 02:56:31 -0600
Subject: [LLVMbugs] [Bug 3048] New: Crash taking sizeof of invalid type
Message-ID: <bug-3048-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3048

           Summary: Crash taking sizeof of invalid type
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Testcase:
int x = sizeof(struct{char qq[x];});

Roughly, the issue is that it crashes trying to evaluate the sizeof.  I'm not
completely sure how to go about fixing 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 cs.uiuc.edu  Wed Nov 12 03:42:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 03:42:25 -0600
Subject: [LLVMbugs] [Bug 3049] New: Assertion failed: (0 && "Unhandled
 cast"), function VisitCastExpr, file ExprConstant.cpp, line 146.
Message-ID: <bug-3049-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3049

           Summary: Assertion failed: (0 && "Unhandled cast"), function
                    VisitCastExpr, file ExprConstant.cpp, line 146.
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: AST
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


when compiling the attached I am getting this assert, this did not happen with
clang from Oct 29.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 12 06:45:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 06:45:51 -0600
Subject: [LLVMbugs] [Bug 3047] llvmc2 $ENV(...) segfaults when the
 environment variable is missing
In-Reply-To: <bug-3047-206@http.llvm.org/bugs/>
Message-ID: <200811121245.mACCjpnX024365@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3047


Mikhail Glushenkov <foldr at codedgers.com> changed:

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




--- Comment #2 from Mikhail Glushenkov <foldr at codedgers.com>  2008-11-12 06:45:50 ---
Fixed in:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081110/069778.html

Thanks for your feedback!


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 12 08:17:08 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 08:17:08 -0600
Subject: [LLVMbugs] [Bug 3051] New: cmake fails to build llvm and clang in
 shared library mode (-DBUILD_SHARED_LIBS=ON)
Message-ID: <bug-3051-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3051

           Summary: cmake fails to build llvm and clang in shared library
                    mode (-DBUILD_SHARED_LIBS=ON)
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ljanyst at cern.ch
                CC: llvmbugs at cs.uiuc.edu


Hi,

   I'm trying to build LLVM and CLANG (the trunk of today in both cases) using
CMake on Linux in the shared library mode (with -DBUILD_SHARED_LIBS=ON
specified in a commandline). The build fails with undefined references to
dlopen and friends:

../../lib/libLLVMSystem.so: undefined reference to `dlerror'
../../lib/libLLVMSystem.so: undefined reference to `dlclose'
../../lib/libLLVMSystem.so: undefined reference to `dlopen'
../../lib/libLLVMSystem.so: undefined reference to `dlsym'

When the problem is brutally resolved by adding the following to the
add_llvm_library macro:

target_link_libraries( ${name} -ldl )

Then a new issue appears:

-- Target triple:
llvm-config: unknown component name: bitreader
CMake Error at cmake/modules/LLVMConfig.cmake:88 (message):
  llvm-config failed for executable opt
Call Stack (most recent call first):
  cmake/modules/LLVMConfig.cmake:5 (nix_llvm_config)
  cmake/modules/AddLLVM.cmake:25 (llvm_config)
  cmake/modules/AddLLVM.cmake:42 (add_llvm_executable)
  tools/opt/CMakeLists.txt:4 (add_llvm_tool)

The build process seems to complete just fine in the static library mode.

Can somebody please take a look?

Cheers,
   Lukasz


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 12 08:26:20 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 08:26:20 -0600
Subject: [LLVMbugs] [Bug 3053] New: clang doesn't check foo(int a[static
 10]) callers (C99)
Message-ID: <bug-3053-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3053

           Summary: clang doesn't check foo(int a[static 10]) callers (C99)
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nunoplopes at sapo.pt
                CC: llvmbugs at cs.uiuc.edu


clang doesn't emit any warning for the following code:

void foo(int a[static 10]);

void bar() {
  int a[2];
  foo(a);
}

Per C99 standard, section 6.7.5.3, I believe clang should produce a warning in
the call stmt. (FWIW, gcc doesn't check it as well).


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 12 08:46:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 08:46:11 -0600
Subject: [LLVMbugs] [Bug 3054] New: Struct ret promotion doesn't handle
 first-class aggregates.
Message-ID: <bug-3054-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3054

           Summary: Struct ret promotion doesn't handle first-class
                    aggregates.
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: fvbommel at wxs.nl
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2198)
 --> (http://llvm.org/bugs/attachment.cgi?id=2198)
A sample .ll with an sret parameter being used by a load.

"opt -sretpromotion" doesn't do anything if any of the sret arguments have a
load instruction as one of their uses, even though the IR should be trivial to
update in this case. Either replace all uses of the load with the return value
(if the call dominates the load) or just store the return value to the sret
argument and let -mem2reg/-scalarrepl clean it up...
(Invokes may be trickier to update, but that can be checked for if necessary)

Attaching sample .ll to test with. (The following:
$ llvm-as <first-class-sret.ll | opt -sretpromotion | llvm-dis | grep sret
should show no instances of 'sret')


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 12 13:07:02 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 13:07:02 -0600
Subject: [LLVMbugs] [Bug 3045] 'A' constraint not handled
In-Reply-To: <bug-3045-206@http.llvm.org/bugs/>
Message-ID: <200811121907.mACJ72He009290@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3045


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-11-12 13:06:45 ---
Sure, we support 64-bit values with A constraints.  The problem is that the
constraints look like this: "=r,A,r,i,0".  Because =r is tied, we preallocate
it at isel time.  Then we get to the A constraint and don't have registers left
for it.  Try removing the tied constraint and the example works.




*** This bug has been marked as a duplicate of bug 1779 ***


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 12 13:25:31 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 13:25:31 -0600
Subject: [LLVMbugs] [Bug 3058] New: String literals aren't compile-time
	constants
Message-ID: <bug-3058-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3058

           Summary: String literals aren't compile-time constants
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sebastian.redl at getdesigned.at
                CC: llvmbugs at cs.uiuc.edu


I'm sure this is well-known, but a bug report is always good to have.

At least on my platform, Linux x64, string literals are not compile-time
constants. This leads to the test cases Sema/compound-literal.c and
Sema/array-init.c failing. In both test cases, there are places where a string
literal is used in an initializer. In both test cases, these fail with,
"initializer element is not a compile-time constant".


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 12 14:41:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 14:41:53 -0600
Subject: [LLVMbugs] [Bug 3051] cmake fails to build llvm and clang in shared
 library mode ( -DBUILD_SHARED_LIBS=ON)
In-Reply-To: <bug-3051-206@http.llvm.org/bugs/>
Message-ID: <200811122041.mACKfrRw013564@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3051


??scar Fuentes <ofv at wanadoo.es> changed:

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




--- Comment #1 from ??scar Fuentes <ofv at wanadoo.es>  2008-11-12 14:41:49 ---
GenLibDeps.pl only inspects static libraries. So the problem affects the CMake
build and the traditional one as well. Patching GenLibDeps.pl for inspecting
shared libraries plus the solution provided by the reporter for the dependency
of LLVMSystem on dl fixes the 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 cs.uiuc.edu  Wed Nov 12 15:18:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 15:18:39 -0600
Subject: [LLVMbugs] [Bug 3058] String literals aren't compile-time constants
In-Reply-To: <bug-3058-206@http.llvm.org/bugs/>
Message-ID: <200811122118.mACLIdob015045@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3058


Sebastian Redl <sebastian.redl at getdesigned.at> changed:

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




--- Comment #3 from Sebastian Redl <sebastian.redl at getdesigned.at>  2008-11-12 15:18:36 ---
Yes, that did it. Fix checked in.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 12 15:27:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 15:27:25 -0600
Subject: [LLVMbugs] [Bug 3048] Crash taking sizeof of invalid type
In-Reply-To: <bug-3048-206@http.llvm.org/bugs/>
Message-ID: <200811122127.mACLRPsc015511@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3048


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-12 15:27:22 ---
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 cs.uiuc.edu  Wed Nov 12 15:37:57 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 15:37:57 -0600
Subject: [LLVMbugs] [Bug 3059] New: CBE miscompilation with global weak alias
Message-ID: <bug-3059-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3059

           Summary: CBE miscompilation with global weak alias
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: C
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: jasandov at rice.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2199)
 --> (http://llvm.org/bugs/attachment.cgi?id=2199)
CBE weak alias test case

I'm running into an error with the CBE. ??I've narrowed the problem down to a
very simple CPP program:

CBE_Bug.cpp:
--------------------------------------------------------------------------------
#include <string>

static std::string hello("Hello world!");

int main() {
????return 0;
}
--------------------------------------------------------------------------------

When I generate C code (using LLVM's C backend) with the following commands, I
get a compile error on the resulting C code:

$ llvm-g++ -emit-llvm -O2 -c -o CBE_Bug.bc CBE_Bug.cpp
$ llc -f -march=c CBE_Bug.bc
$ gcc -c CBE_Bug.cbe.c
CBE_Bug.cbe.c:238: warning: conflicting types for built-in function
???malloc???
CBE_Bug.cbe.c: In function ???main???:
CBE_Bug.cbe.c:269: warning: return type of ???main??? is not ???int???
CBE_Bug.cbe.c: In function ???__tcf_0???:
CBE_Bug.cbe.c:307: error: ???_ZL22__gthrw_pthread_cancelm??? undeclared (first
use in this function)
CBE_Bug.cbe.c:307: error: (Each undeclared identifier is reported only once
CBE_Bug.cbe.c:307: error: for each function it appears in.)


Here are the relevant lines from CBE_Bug.cbe.c. ??Notice that "pthread_cancel"
is declared as a weak external symbol, but "_ZL22__gthrw_pthread_cancelm" is
never declared (even though it is used).

$ grep -Hn pthread_cancel CBE_Bug.cbe.c
CBE_Bug.cbe.c:228:extern unsigned int pthread_cancel(unsigned long long )
__EXTERNAL_WEAK__;
CBE_Bug.cbe.c:307:  if ((((unsigned char )(bool )(((unsigned int )(bool
)(_ZL22__gthrw_pthread_cancelm != ((unsigned int  (*) (unsigned long long
))/*NULL*/0))) != 0u)) != ((unsigned char )0))) {


Here are the relevant lines from the LLVM bitcode. ??The bitcode appears
correct since "pthread_cancel" is declared as extern_weak and the alias
"_ZL22__gthrw_pthread_cancelm" is declared as a weak alias of "pthread_cancel".
??I'm just a bit confused about "pthread_cancel" -- shouldn't it be declared as
a regular external symbol?

$ llvm-dis CBE_Bug.bc
$ grep -Hn pthread_cancel CBE_Bug.ll
CBE_Bug.ll:24:@_ZL22__gthrw_pthread_cancelm = alias weak i32 (i64)*
@pthread_cancel             ; <i32 (i64)*> [#uses=1]
CBE_Bug.ll:76:  br i1 icmp ne (i8 zext (i1 icmp ne (i32 zext (i1 icmp ne (i32
(i64)* @_ZL22__gthrw_pthread_cancelm, i32 (i64)* null) to i32), i32 0) to i8),
i8 0), label %bb20, label %bb21
CBE_Bug.ll:117:declare extern_weak i32 @pthread_cancel(i64)

Finally, looking at the preprocessed CBE_Bug.cpp file (before running it
through LLVM), I can see the appropriate declarations for "pthread_cancel" and
its alias. ??This (mostly) matches the LLVM representation above.

$ llvm-g++ -emit-llvm -E -o CBE_Bug.E.cpp CBE_Bug.cpp
$ grep -Hn pthread_cancel CBE_Bug.E.cpp
CBE_Bug.E.cpp:4417:extern int pthread_cancel (pthread_t __th);
CBE_Bug.E.cpp:6061:static __typeof(pthread_cancel) __gthrw_pthread_cancel
__attribute__ ((__weakref__("pthread_cancel")));
CBE_Bug.E.cpp:6078:    = __extension__ (void *) &__gthrw_pthread_cancel;

So, it looks like the CBE code generator is not producing correct output for
the alias, since the LLVM bitcode appears correct.  The CBE correctly generates
an external weak declaration for pthread_cancel, but it produces no declaration
for the associated alias. ??This problem only appears when I turn on
optimization (any level above 0) -- the code compiles fine at "-O0".

I'm using LLVM release 2.4, but I also verified that the problem occurs with
the latest trunk revision from SVN. ??I'm on x86_64 linux (Fedora 8) with gcc
4.3.1.


The problem should be easy to reproduce with the input file I gave above; if
you want any other temp files, just let me know.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 12 18:58:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 18:58:47 -0600
Subject: [LLVMbugs] [Bug 3060] New: New IRgen of 'if' with constants is
	broken.
Message-ID: <bug-3060-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3060

           Summary: New IRgen of 'if' with constants is broken.
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: major
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


The new constant IRgen is too aggressive because it is possible to evaluate
expressions with side effects to constants.
--
ddunbar at lordcrumb:CodeGen$ cat if.c
int g0(void);

int f0(void) {
  if (g0() && 0)
    return 0;
  return 1;
}
ddunbar at lordcrumb:CodeGen$ clang -emit-llvm -o - -O3 if.c
; ModuleID = 'if.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-f80:128:128"
target triple = "i686-apple-darwin10.0.0d3"

define i32 @f0() nounwind readnone {
entry:
        ret i32 1
}
--

The same issue occurs in other places, for example with the comma operator.

Chris and I discussed and came up with a plan to have tryEvaluate return more
information about the folded constant. The idea would be to have the result be
one of
(1) not evaluatable
(2) C99 ICE, which should never have side effects or labels
(3) constant value bla, with optional bits indicating if the expression has
side effects and if the expression has labels (useful for consumers in
different contexts). One can imagine other bits like imprecise-floating-point
result being useful in time.

This bug is a placeholder for this 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 cs.uiuc.edu  Wed Nov 12 19:08:13 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 19:08:13 -0600
Subject: [LLVMbugs] [Bug 3062] New: ({}) should be rejected at file scope.
Message-ID: <bug-3062-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3062

           Summary: ({}) should be rejected at file scope.
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


We should never ever ever accept
--
int x[({10;})];
--
but we do. gcc rejects ({}) outside functions.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 12 19:16:49 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 19:16:49 -0600
Subject: [LLVMbugs] [Bug 3063] New: clang is overly permissive with
 statement expressions
Message-ID: <bug-3063-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3063

           Summary: clang is overly permissive with statement expressions
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: parser
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: daniel at zuster.org
                CC: llvmbugs at cs.uiuc.edu


clang accepts:
-- 
int f0() {
  int x = sizeof( ({ L1: g0(); 10;}) );
  goto L1;
}
--

Need I say more?

The sizeof here is overkill, gcc seems to reject all jumps into statement
expressions. A nice tether to the brink of insanity.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 12 23:24:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 12 Nov 2008 23:24:44 -0600
Subject: [LLVMbugs] [Bug 3064] New: GettingStarted guide humiliates
	linux/amd64 platform
Message-ID: <bug-3064-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3064

           Summary: GettingStarted guide humiliates linux/amd64 platform
           Product: Documentation
           Version: 2.4
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: General docs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: go0 at go.ru
                CC: llvmbugs at cs.uiuc.edu


GettingStarted guide as shipped with 2.4 software and seen online today morning
says that llvm on Linux/amd64 is unable to generate native code

In my experience this statement is completely wrong: natively built llvm-gcc is
a working replacement of regular gcc

Is there a test proving that i am wrong and documentation is right? Builtin
test nearly succeeds (2781 expected passes, 3 unexpected failures, 1 unexpected
success, 5 unexpected failures). I failed to run external test-suite and failed
to cross-compile llvm-gcc for mips target, but don't experience any problem
with llvm-gcc as a native compiler

I don't believe i'm the only one who compiled llvm-gcc on linux/amd64. Maybe
GettingStarted.html needs updating?


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

From bugzilla-daemon at cs.uiuc.edu  Thu Nov 13 02:56:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Nov 2008 02:56:44 -0600
Subject: [LLVMbugs] [Bug 3065] New: test-suite completely fails due to bad
 makefile test-suite/ SingleSource/Makefile
Message-ID: <bug-3065-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3065

           Summary: test-suite completely fails due to bad makefile test-
                    suite/SingleSource/Makefile
           Product: Test Suite
           Version: trunk
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Nightly Tester
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: go0 at go.ru
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2202)
 --> (http://llvm.org/bugs/attachment.cgi?id=2202)
configure-make-make check-... log file, test-suite Makefile's

I am completely unable to run external test suite (the big package that should
be unpacked to projects/test-suite). I try to closely follow TestingGuide, and
i even use an antique 3.79.1 version of make which is recommended by llvm
documentation as known-to-work

Symptom: 
======== log file fragment start ======================
Running
/tmp/portage/dev-libs/llvm_and_test-2.4/work/llvm-2.4/test/Transforms/StripSymbols/dg.exp
...
Running
/tmp/portage/dev-libs/llvm_and_test-2.4/work/llvm-2.4/test/Transforms/TailCallElim/dg.exp
...
Running
/tmp/portage/dev-libs/llvm_and_test-2.4/work/llvm-2.4/test/Transforms/TailDup/dg.exp
...
Running
/tmp/portage/dev-libs/llvm_and_test-2.4/work/llvm-2.4/test/Verifier/dg.exp ...

                ===  Summary ===

# of expected passes            2781
# of unexpected failures        3
# of unexpected successes       1
# of expected failures          5
antique_make-3.79.1[1]: *** [check-local] Error 1
antique_make-3.79.1[1]: Leaving directory
`/tmp/portage/dev-libs/llvm_and_test-2.4/work/llvm-2.4/test'
antique_make-3.79.1: *** [check] Error 2
antique_make-3.79.1: Entering directory
`/tmp/portage/dev-libs/llvm_and_test-2.4/work/llvm-2.4/projects/test-suite'
antique_make-3.79.1[1]: Entering directory
`/tmp/portage/dev-libs/llvm_and_test-2.4/work/llvm-2.4/projects/test-suite/SingleSource'
antique_make-3.79.1[2]: Entering directory
`/tmp/portage/dev-libs/llvm_and_test-2.4/work/llvm-2.4/projects/test-suite/SingleSource/UnitTests'
antique_make-3.79.1[3]: Entering directory
`/tmp/portage/dev-libs/llvm_and_test-2.4/work/llvm-2.4/projects/test-suite/SingleSource/UnitTests/Vector'
antique_make-3.79.1[3]: *** No rule to make target `Output/build.linked.rbc',
needed by `Output/build.linked.bc'.  Stop.
antique_make-3.79.1[3]: Leaving directory
`/tmp/portage/dev-libs/llvm_and_test-2.4/work/llvm-2.4/projects/test-suite/SingleSource/UnitTests/Vector'
antique_make-3.79.1[2]: *** [all] Error 1
antique_make-3.79.1[2]: Leaving directory
`/tmp/portage/dev-libs/llvm_and_test-2.4/work/llvm-2.4/projects/test-suite/SingleSource/UnitTests'
antique_make-3.79.1[1]: *** [UnitTests/.makeall] Error 2
antique_make-3.79.1[1]: Leaving directory
`/tmp/portage/dev-libs/llvm_and_test-2.4/work/llvm-2.4/projects/test-suite/SingleSource'

======== log file fragment start ======================

I found the rule in test-suite/SingleSource/Makefile.singlesrc:

============== Makefile.singlesrc fragment start======================
ifndef USE_PRECOMPILED_BYTECODE
Output/%.linked.rbc: Output/%.bc
        -cp -f $< $@

Output/%.LOC.txt: $(SourceDir)/%.c
        cat $< | wc -l > $@

Output/%.LOC.txt: $(SourceDir)/%.cpp
        cat $< | wc -l > $@
endif
============== Makefile.singlesrc fragment end  ======================

Looks like that this file is not included by accompaning Makefile:
======= test-suite/SingleSource/Makefile start ======
LEVEL = ..

PARALLEL_DIRS  = UnitTests Regression Benchmarks

LDFLAGS += -lm

include Makefile.singlesrc
======= test-suite/SingleSource/Makefile end   ======

LEVEL = .. is incorrect; it should be .

The wrong test-suite/SingleSource/Makefile seems to be auto-generated. But i
failed to find the script that did it. Any ideas? I want to repair it, to test
if my llvm installation is ok

I attach complete log file, some auto-generated files from test-suite
(preserving directory order); the 2 makefiles mentioned above included, of
course


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

From bugzilla-daemon at cs.uiuc.edu  Thu Nov 13 13:24:14 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Nov 2008 13:24:14 -0600
Subject: [LLVMbugs] [Bug 3064] GettingStarted guide humiliates linux/amd64
	platform
In-Reply-To: <bug-3064-206@http.llvm.org/bugs/>
Message-ID: <200811131924.mADJOEZG008799@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3064


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-11-13 13:24:09 ---
Thanks, this is fixed on mainline
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081110/069825.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 cs.uiuc.edu  Thu Nov 13 14:47:34 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Nov 2008 14:47:34 -0600
Subject: [LLVMbugs] [Bug 3063] clang is overly permissive with statement
	expressions
In-Reply-To: <bug-3063-206@http.llvm.org/bugs/>
Message-ID: <200811132047.mADKlYNR013129@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3063


Eli Friedman <sharparrow1 at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sharparrow1 at yahoo.com
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #1 from Eli Friedman <sharparrow1 at yahoo.com>  2008-11-13 14:47:32 ---


*** This bug has been marked as a duplicate of bug 2325 ***


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

From bugzilla-daemon at cs.uiuc.edu  Thu Nov 13 15:58:56 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Nov 2008 15:58:56 -0600
Subject: [LLVMbugs] [Bug 3043] lli fails pointer comparison test
In-Reply-To: <bug-3043-206@http.llvm.org/bugs/>
Message-ID: <200811132158.mADLwuWq016039@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3043


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

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




--- Comment #5 from Evan Cheng <evan.cheng at apple.com>  2008-11-13 15:58:54 ---
Right. That's an acceptable solution for now. Thanks.
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081110/069830.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 cs.uiuc.edu  Thu Nov 13 16:01:40 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Nov 2008 16:01:40 -0600
Subject: [LLVMbugs] [Bug 1779] inlineasm crash due to order of register
	assignment
In-Reply-To: <bug-1779-206@http.llvm.org/bugs/>
Message-ID: <200811132201.mADM1esO016289@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1779


Dale Johannesen <dalej at apple.com> changed:

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




--- Comment #8 from Dale Johannesen <dalej at apple.com>  2008-11-13 16:01:39 ---
Fixed here.
http://llvm.org/viewvc/llvm-project?view=rev&revision=59266


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

From bugzilla-daemon at cs.uiuc.edu  Thu Nov 13 16:25:38 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Nov 2008 16:25:38 -0600
Subject: [LLVMbugs] [Bug 3065] test-suite completely fails due to bad
 makefile test-suite/ SingleSource/Makefile
In-Reply-To: <bug-3065-206@http.llvm.org/bugs/>
Message-ID: <200811132225.mADMPc1S017282@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3065


Tanya Lattner <tonic at nondot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tonic at nondot.org
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




--- Comment #7 from Tanya Lattner <tonic at nondot.org>  2008-11-13 16:25:35 ---
I don't see a bug here. llvm-gcc was either not installed or not supplied to
configure.


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

From bugzilla-daemon at cs.uiuc.edu  Thu Nov 13 17:55:42 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 13 Nov 2008 17:55:42 -0600
Subject: [LLVMbugs] [Bug 3066] New: Assertion hit with simple store into
	4xfloat vector.
Message-ID: <bug-3066-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3066

           Summary: Assertion hit with simple store into 4xfloat vector.
           Product: libraries
           Version: 2.4
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: marejde at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I hit the following assertion:

llvm-2.4/include/llvm/CodeGen/ValueTypes.h:364: unsigned int
llvm::MVT::getSizeInBits() const: Assertion `isExtended() && "MVT has no known
size!"' failed.

when trying to jit a module with the following IR:

@a_global_variable = weak global <4 x float>            ; <<4 x float>*>
[#uses=1]

define void @main() {
entry:
        store <4 x float> < float 1.000000e+00, float 0.000000e+00, float
0.000000e+00, float 1.000000e+00 >, <4 x float>* @a_global_variable
        ret void undef
}

in the following way (error checks etc removed):

execution_engine = llvm::ExecutionEngine::create(module);
llvm::Function *f = module->getFunction("main");
main_function = (void (*) ()) execution_engine->getPointerToFunction(f);

FWIW, I didn't hit the assertion with 2.3. Instead, the program crashed. Since
2.4 got released a couple of days ago I didn't investigate it further with 2.3.


-- 
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 cs.uiuc.edu  Fri Nov 14 06:54:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Nov 2008 06:54:51 -0600
Subject: [LLVMbugs] [Bug 3067] New: minor error in gcc/Makefile.in:
	misplaced brace
Message-ID: <bug-3067-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3067

           Summary: minor error in gcc/Makefile.in: misplaced brace
           Product: tools
           Version: 2.4
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: go0 at go.ru
                CC: llvmbugs at cs.uiuc.edu


IMHO misplaced opening brace in gcc/Makefile.in found inside
llvm-gcc-4.2-2.4.source.tar.gz 



The bad line:

ifneq (($LLVMOBJDIR),$(LLVMSRCDIR))

IMHO it should be

ifneq ($(LLVMOBJDIR)),$(LLVMSRCDIR))

This error results in duped -I... when the 2 dirs coincide

I've tested this both under make 3.81 and make 3.79.1


-- 
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 cs.uiuc.edu  Fri Nov 14 08:16:34 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Nov 2008 08:16:34 -0600
Subject: [LLVMbugs] [Bug 3068] New: llvm-gcc cross-compilation for mips
 fails: unhandled type f32
Message-ID: <bug-3068-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3068

           Summary: llvm-gcc cross-compilation for mips fails: unhandled
                    type f32
           Product: tools
           Version: 2.4
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Keywords: build-problem, compile-fail
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: go0 at go.ru
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2204)
 --> (http://llvm.org/bugs/attachment.cgi?id=2204)
log file and gentoo'ish build script

Any luck cross compiling? I tried to compile for MIPS (mips-pc-linux-gnu)

xgcc built during compilation fails to compile: 

====== log fragment starts ====================
/tmp/portage/sys-devel/llvm-gcc-mips-2.4/work/obj/./gcc/xgcc
-B/tmp/portage/sys-devel/llvm-gcc-mips-2.4/work/obj/./gcc/
-B/usr/lib/llvm-gcc_mips/mips-pc-linux-gnu/bin/
-B/usr/lib/llvm-gcc_mips/mips-pc-linux-gnu/lib/ -isystem
/usr/lib/llvm-gcc_mips/mips-pc-linux-gnu/include -isystem
/usr/lib/llvm-gcc_mips/mips-pc-linux-gnu/sys-include  -O2  -O2 -O2  -DIN_GCC
-DCROSS_DIRECTORY_STRUCTURE   -W -Wall -Wwrite-strings -Wstrict-prototypes
-Wmissing-prototypes -Wold-style-definition  -isystem ./include  -fPIC -g
-DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -Dinhibit_libc -I.
-I. -I/tmp/portage/sys-devel/llvm-gcc-mips-2.4/work/llvm-gcc4.2-2.4.source/gcc
-I/tmp/portage/sys-devel/llvm-gcc-mips-2.4/work/llvm-gcc4.2-2.4.source/gcc/.
-I/tmp/portage/sys-devel/llvm-gcc-mips-2.4/work/llvm-gcc4.2-2.4.source/gcc/../include
-I/tmp/portage/sys-devel/llvm-gcc-mips-2.4/work/llvm-gcc4.2-2.4.source/gcc/../libcpp/include

-I/tmp/portage/sys-devel/llvm-gcc-mips-2.4/work/llvm-gcc4.2-2.4.source/gcc/../libdecnumber
-I../libdecnumber -I/usr/include -DL_powisf2 -c
/tmp/portage/sys-devel/llvm-gcc-mips-2.4/work/llvm-gcc4.2-2.4.source/gcc/libgcc2.c
-o libgcc/./_powisf2.o
Formal argument #0 has unhandled type f32
/tmp/portage/sys-devel/llvm-gcc-mips-2.4/work/llvm-gcc4.2-2.4.source/gcc/libgcc2.c:1765:
internal compiler error: Aborted
====== log fragment   ends ====================

Am i building it wrong? Or maybe i should try some svn version?

I enclose full log and ebuild


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

From bugzilla-daemon at cs.uiuc.edu  Fri Nov 14 12:31:23 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Nov 2008 12:31:23 -0600
Subject: [LLVMbugs] [Bug 3069] New: The return value of
 llvm::sys::DynamicLibrary:: LoadLibraryPermanently is counter-intuitive
Message-ID: <bug-3069-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3069

           Summary: The return value of
                    llvm::sys::DynamicLibrary::LoadLibraryPermanently is
                    counter-intuitive
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: System Library
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: mdevan.foobar at gmail.com
                CC: llvmbugs at cs.uiuc.edu


The (boolean) return value of llvm::sys::DynamicLibrary::LoadLibraryPermanently
is counter-intuitive. From the documentation:

"This returns false on success or returns true and fills in *ErrMsg on
failure."

Although it behaves as documented and is used in 2 places accordingly, it's a
bit.. ah.. unexpected!


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

From bugzilla-daemon at cs.uiuc.edu  Fri Nov 14 15:14:34 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 14 Nov 2008 15:14:34 -0600
Subject: [LLVMbugs] [Bug 3070] New: undefined reference to
 `__builtin_stdarg_start' with g++-4. 4.0
Message-ID: <bug-3070-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3070

           Summary: undefined reference to `__builtin_stdarg_start' with
                    g++-4.4.0
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: minor
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


This testcase works with g++ 4.3.2 as system compiler, but doesn't work with
gcc version 4.4.0 20081114 (experimental) [trunk revision 141857]:

/home/edwin/gcc_inst/bin/gcc
-I/home/edwin/llvm-svn/llvm-commit/projects/llvm-test/SingleSource/Regression/C
-I/home/edwin/llvm-svn/llvm-commit/projects/llvm-test/SingleSource/Regression/C
-I/home/edwin/llvm-svn/llvm-commit/include
-I/home/edwin/llvm-svn/llvm-commit/projects/llvm-test/include
-I../../../include -I/home/edwin/llvm-svn/llvm-commit/include -D_GNU_SOURCE
-D__STDC_LIMIT_MACROS -DNDEBUG  -O3 
/home/edwin/llvm-svn/llvm-commit/projects/llvm-test/SingleSource/Regression/C/PR640.c
-lm -o Output/PR640.native
/tmp/ccsk0D8r.o: In function `test_stdarg_builtin_va':
PR640.c:(.text+0xb1): undefined reference to `__builtin_stdarg_start'
collect2: ld returned 1 exit status
make: [Output/PR640.native] Error 1 (ignored)


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 15 01:09:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Nov 2008 01:09:12 -0600
Subject: [LLVMbugs] [Bug 3066] Assertion hit with simple store into 4xfloat
	vector.
In-Reply-To: <bug-3066-206@http.llvm.org/bugs/>
Message-ID: <200811150709.mAF79CAd032604@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3066


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #4 from Nick Lewycky <nicholas at mxc.ca>  2008-11-15 01:09:11 ---
I think it's invalid then.

It sounds like you created invalid LLVM IR, which the C++ API can't stop you
from doing, but which you can check for by running the verifier
(createVerifierPass).


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 15 09:19:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Nov 2008 09:19:41 -0600
Subject: [LLVMbugs] [Bug 3071] New: CommandLine. h assertion failed with
 linking LLVM to a shared library
Message-ID: <bug-3071-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3071

           Summary: CommandLine.h assertion failed with linking LLVM to a
                    shared library
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rushpl at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2206)
 --> (http://llvm.org/bugs/attachment.cgi?id=2206)
Test code showing the assertion fail

The following error shows up after running my program that is linked to my
custom library that is made in a fashion described in the Summary.

test1: /usr/src/packages/BUILD/llvm-2.3/include/llvm/Support/CommandLine.h:477:
void llvm::cl::parser<DataType>::addLiteralOption(const char*, const DT&, const
char*) [with DT = const llvm::TargetMachineRegistryEntry*, DataType = const
llvm::TargetMachineRegistryEntry*]: Assertion `findOption(Name) ==
Values.size() && "Option already exists!"' failed.

(Yes my LLVM packages are not the latest 2.4 but edwin on IRC confirmed with
the latest - I suppose)
I attached a sample code that will produce the error. It requires CMake for
building. Furthermore it has been suggested to also paste the building phase in
case somebody can spot something wrong with it.

rush at rushpod:~/Programowanie/LLVM_DblDef> cmake .
-- The C compiler identification is GNU
-- The CXX compiler identification is GNU
-- Check for working C compiler: /usr/bin/gcc
-- Check for working C compiler: /usr/bin/gcc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
NOTICE Libs: /usr/lib/llvm/LLVMExecutionEngine.o /usr/lib/llvm/LLVMJIT.o
/usr/lib/llvm/LLVMX86.o /usr/lib/llvm/libLLVMSelectionDAG.a
/usr/lib/llvm/libLLVMCodeGen.a /usr/lib/llvm/libLLVMScalarOpts.a
/usr/lib/llvm/libLLVMTransformUtils.a /usr/lib/llvm/libLLVMipa.a
/usr/lib/llvm/libLLVMAnalysis.a /usr/lib/llvm/libLLVMTarget.a
/usr/lib/llvm/libLLVMCore.a /usr/lib/llvm/libLLVMSupport.a
/usr/lib/llvm/libLLVMSystem.a
NOTICE linking with /usr/lib/llvm/LLVMExecutionEngine.o
NOTICE linking with /usr/lib/llvm/LLVMJIT.o
NOTICE linking with /usr/lib/llvm/LLVMX86.o
NOTICE linking with /usr/lib/llvm/libLLVMSelectionDAG.a
NOTICE linking with /usr/lib/llvm/libLLVMCodeGen.a
NOTICE linking with /usr/lib/llvm/libLLVMScalarOpts.a
NOTICE linking with /usr/lib/llvm/libLLVMTransformUtils.a
NOTICE linking with /usr/lib/llvm/libLLVMipa.a
NOTICE linking with /usr/lib/llvm/libLLVMAnalysis.a
NOTICE linking with /usr/lib/llvm/libLLVMTarget.a
NOTICE linking with /usr/lib/llvm/libLLVMCore.a
NOTICE linking with /usr/lib/llvm/libLLVMSupport.a
NOTICE linking with /usr/lib/llvm/libLLVMSystem.a
-- Configuring done
-- Generating done
-- Build files have been written to: /home/rush/Programowanie/LLVM_DblDef
rush at rushpod:~/Programowanie/LLVM_DblDef> make VERBOSE=1
/usr/bin/cmake -H/home/rush/Programowanie/LLVM_DblDef
-B/home/rush/Programowanie/LLVM_DblDef --check-build-system
CMakeFiles/Makefile.cmake 0
/usr/bin/cmake -E cmake_progress_start
/home/rush/Programowanie/LLVM_DblDef/CMakeFiles
/home/rush/Programowanie/LLVM_DblDef/CMakeFiles/progress.make
make -f CMakeFiles/Makefile2 all
make[1]: Wej??cie do katalogu `/home/rush/Programowanie/LLVM_DblDef'
make -f CMakeFiles/testlib.dir/build.make CMakeFiles/testlib.dir/depend
make[2]: Wej??cie do katalogu `/home/rush/Programowanie/LLVM_DblDef'
cd /home/rush/Programowanie/LLVM_DblDef && /usr/bin/cmake -E cmake_depends
"Unix Makefiles" /home/rush/Programowanie/LLVM_DblDef
/home/rush/Programowanie/LLVM_DblDef /home/rush/Programowanie/LLVM_DblDef
/home/rush/Programowanie/LLVM_DblDef
/home/rush/Programowanie/LLVM_DblDef/CMakeFiles/testlib.dir/DependInfo.cmake
--color=
Dependee
"/home/rush/Programowanie/LLVM_DblDef/CMakeFiles/testlib.dir/DependInfo.cmake"
is newer than depender
"/home/rush/Programowanie/LLVM_DblDef/CMakeFiles/testlib.dir/depend.internal".
Scanning dependencies of target testlib
make[2]: Opuszczenie katalogu `/home/rush/Programowanie/LLVM_DblDef'
make -f CMakeFiles/testlib.dir/build.make CMakeFiles/testlib.dir/build
make[2]: Wej??cie do katalogu `/home/rush/Programowanie/LLVM_DblDef'
/usr/bin/cmake -E cmake_progress_report
/home/rush/Programowanie/LLVM_DblDef/CMakeFiles 2
[ 50%] Building CXX object CMakeFiles/testlib.dir/testlib.c++.o
/usr/bin/c++   -D__STDC_LIMIT_MACROS -Dtestlib_EXPORTS -fPIC   -o
CMakeFiles/testlib.dir/testlib.c++.o -c
/home/rush/Programowanie/LLVM_DblDef/testlib.c++
Linking CXX shared library libtestlib.so
/usr/bin/cmake -E cmake_link_script CMakeFiles/testlib.dir/link.txt --verbose=1
/usr/bin/c++  -fPIC   -shared -Wl,-soname,libtestlib.so -o libtestlib.so
CMakeFiles/testlib.dir/testlib.c++.o /usr/lib/llvm/LLVMExecutionEngine.o
/usr/lib/llvm/LLVMJIT.o /usr/lib/llvm/LLVMX86.o
/usr/lib/llvm/libLLVMSelectionDAG.a /usr/lib/llvm/libLLVMCodeGen.a
/usr/lib/llvm/libLLVMScalarOpts.a /usr/lib/llvm/libLLVMTransformUtils.a
/usr/lib/llvm/libLLVMipa.a /usr/lib/llvm/libLLVMAnalysis.a
/usr/lib/llvm/libLLVMTarget.a /usr/lib/llvm/libLLVMCore.a
/usr/lib/llvm/libLLVMSupport.a /usr/lib/llvm/libLLVMSystem.a -lpthread -ldl
make[2]: Opuszczenie katalogu `/home/rush/Programowanie/LLVM_DblDef'
/usr/bin/cmake -E cmake_progress_report
/home/rush/Programowanie/LLVM_DblDef/CMakeFiles  2
[ 50%] Built target testlib
make -f CMakeFiles/test1.dir/build.make CMakeFiles/test1.dir/depend
make[2]: Wej??cie do katalogu `/home/rush/Programowanie/LLVM_DblDef'
cd /home/rush/Programowanie/LLVM_DblDef && /usr/bin/cmake -E cmake_depends
"Unix Makefiles" /home/rush/Programowanie/LLVM_DblDef
/home/rush/Programowanie/LLVM_DblDef /home/rush/Programowanie/LLVM_DblDef
/home/rush/Programowanie/LLVM_DblDef
/home/rush/Programowanie/LLVM_DblDef/CMakeFiles/test1.dir/DependInfo.cmake
--color=
Dependee
"/home/rush/Programowanie/LLVM_DblDef/CMakeFiles/test1.dir/DependInfo.cmake" is
newer than depender
"/home/rush/Programowanie/LLVM_DblDef/CMakeFiles/test1.dir/depend.internal".
Scanning dependencies of target test1
make[2]: Opuszczenie katalogu `/home/rush/Programowanie/LLVM_DblDef'
make -f CMakeFiles/test1.dir/build.make CMakeFiles/test1.dir/build
make[2]: Wej??cie do katalogu `/home/rush/Programowanie/LLVM_DblDef'
/usr/bin/cmake -E cmake_progress_report
/home/rush/Programowanie/LLVM_DblDef/CMakeFiles 1
[100%] Building CXX object CMakeFiles/test1.dir/test1.c++.o
/usr/bin/c++   -D__STDC_LIMIT_MACROS  -o CMakeFiles/test1.dir/test1.c++.o -c
/home/rush/Programowanie/LLVM_DblDef/test1.c++
Linking CXX executable test1
/usr/bin/cmake -E cmake_link_script CMakeFiles/test1.dir/link.txt --verbose=1
/usr/bin/c++     -fPIC CMakeFiles/test1.dir/test1.c++.o  -o test1 -rdynamic
libtestlib.so /usr/lib/llvm/LLVMExecutionEngine.o /usr/lib/llvm/LLVMJIT.o
/usr/lib/llvm/LLVMX86.o /usr/lib/llvm/libLLVMSelectionDAG.a
/usr/lib/llvm/libLLVMCodeGen.a /usr/lib/llvm/libLLVMScalarOpts.a
/usr/lib/llvm/libLLVMTransformUtils.a /usr/lib/llvm/libLLVMipa.a
/usr/lib/llvm/libLLVMAnalysis.a /usr/lib/llvm/libLLVMTarget.a
/usr/lib/llvm/libLLVMCore.a /usr/lib/llvm/libLLVMSupport.a
/usr/lib/llvm/libLLVMSystem.a -lpthread -ldl
-Wl,-rpath,/home/rush/Programowanie/LLVM_DblDef
make[2]: Opuszczenie katalogu `/home/rush/Programowanie/LLVM_DblDef'
/usr/bin/cmake -E cmake_progress_report
/home/rush/Programowanie/LLVM_DblDef/CMakeFiles  1
[100%] Built target test1
make[1]: Opuszczenie katalogu `/home/rush/Programowanie/LLVM_DblDef'
/usr/bin/cmake -E cmake_progress_start
/home/rush/Programowanie/LLVM_DblDef/CMakeFiles 0
rush at rushpod:~/Programowanie/LLVM_DblDef> ./test1
test1: /usr/src/packages/BUILD/llvm-2.3/include/llvm/Support/CommandLine.h:477:
void llvm::cl::parser<DataType>::addLiteralOption(const char*, const DT&, const
char*) [with DT = const llvm::TargetMachineRegistryEntry*, DataType = const
llvm::TargetMachineRegistryEntry*]: Assertion `findOption(Name) ==
Values.size() && "Option already exists!"' failed.
Przerwane
rush at rushpod:~/Programowanie/LLVM_DblDef>


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 15 11:00:27 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Nov 2008 11:00:27 -0600
Subject: [LLVMbugs] [Bug 3072] New: CMake build fails with LLVM in a
	subdirectory
Message-ID: <bug-3072-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3072

           Summary: CMake build fails with LLVM in a subdirectory
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: ofv at wanadoo.es
        ReportedBy: hyperquantum at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Building LLVM with CMake fails if LLVM's top-level CMake file is not the
top-level CMake file of the build.

To reproduce:

$ mkdir test
$ cd test
$ svn co http://llvm.org/svn/llvm-project/llvm/trunk llvm
$ echo "add_subdirectory(llvm)" >CMakeLists.txt
$ cd ..
$ mkdir test_build
$ cd test_build
$ cmake ../test
$ make

The error message:

[ 94%] Built target llvm-config.target
llvm/tools/opt/CMakeFiles/opt.dir/build.make:44:
tools/opt/CMakeFiles/opt.dir/depend.make: No such file or directory
llvm/tools/opt/CMakeFiles/opt.dir/build.make:47:
tools/opt/CMakeFiles/opt.dir/progress.make: No such file or directory
llvm/tools/opt/CMakeFiles/opt.dir/build.make:50:
tools/opt/CMakeFiles/opt.dir/flags.make: No such file or directory
make[2]: *** No rule to make target `tools/opt/CMakeFiles/opt.dir/flags.make'. 
Stop.
make[1]: *** [llvm/tools/opt/CMakeFiles/opt.dir/all] Error 2
make: *** [all] Error 2


Note: I didn't use "svn co" but copied the directory from another SVN checkout.
In theory this shouldn't make a difference. (The SVN revision I used was 59362)


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 15 12:57:21 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Nov 2008 12:57:21 -0600
Subject: [LLVMbugs] [Bug 3046] __builtin__CFStringMakeConstantString crashes
 on 64-bit systems
In-Reply-To: <bug-3046-206@http.llvm.org/bugs/>
Message-ID: <200811151857.mAFIvLHG003550@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3046


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Anders Carlsson <andersca at mac.com>  2008-11-15 12:57:20 ---
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081110/009271.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 cs.uiuc.edu  Sat Nov 15 13:17:34 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Nov 2008 13:17:34 -0600
Subject: [LLVMbugs] [Bug 3073] New: Sema::DiagnoseAssignmentResult has
 broken message logic
Message-ID: <bug-3073-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3073

           Summary: Sema::DiagnoseAssignmentResult has broken message logic
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sebastian.redl at getdesigned.at
                CC: llvmbugs at cs.uiuc.edu


Sema::DiagnoseAssignmentResult selects an error message and then passes three
arguments to it, in this order:
1) The destination type.
2) The source type.
3) The "kind", e.g. "initializing".

However, the selected error messages aren't all compatible with this scheme.
err_typecheck_convert_incompatible, for example, was originally written to be
passed the name of the variable being initialized as the second argument and
reads thus:
"incompatible type %2 '%1', expected '%0'"

For the following code:

struct S {};

void f()
{
  int i = *(S*)0;
}

this leads to the highly misleading message:
"error: incompatible type initializing 'struct S', expected 'int'"

As a user, I can only wonder why Clang thinks that I want to initialize 'struct
S'.

Unfortunately, the error message is used in other places, too, so it cannot
simply be adapted. The broken part here is, in my opinion, that
DiagnoseAssignmentResult and the call paths leading to it do not preserve the
InitEntity parameter that functions like CheckInitializerTypes have.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 15 22:22:19 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Nov 2008 22:22:19 -0600
Subject: [LLVMbugs] [Bug 2602] OOM in SCEV with polynomial induction variable
In-Reply-To: <bug-2602-206@http.llvm.org/bugs/>
Message-ID: <200811160422.mAG4MJp7021802@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2602


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #4 from Nick Lewycky <nicholas at mxc.ca>  2008-11-15 22:22:18 ---
Fixed in r59384, by not analyzing these.

If this sort of thing comes up in real-world code, let me know.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 15 23:05:30 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Nov 2008 23:05:30 -0600
Subject: [LLVMbugs] [Bug 3074] New: ConstantInt::get takes an "isSigned"
	parameter?
Message-ID: <bug-3074-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3074

           Summary: ConstantInt::get takes an "isSigned" parameter?
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


ConstantInt::get takes an isSigned parameter. Which is pretty amazing because
ConstantInts don't carry a sign. What it does with it, is pass it through to
the APInt constructor. Which is pretty amazing, because APInts don't carry a
sign.

The APInt constructor has two cases, one in which the entire value fits in a
int64_t. In the other case, it calls initSlowCase, which takes isSigned and
uses that to clear the memory space to all 0 bits or all 1 bits depending on
whether it's signed or unsigned.

You want to know what else is amazing? If you hard-code the isSigned parameter
to initSlowCase to false, make check fails.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 15 23:40:09 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 15 Nov 2008 23:40:09 -0600
Subject: [LLVMbugs] [Bug 3074] ConstantInt::get takes an "isSigned"
	parameter?
In-Reply-To: <bug-3074-206@http.llvm.org/bugs/>
Message-ID: <200811160540.mAG5e9vN024393@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3074


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-15 23:40:08 ---
I don't get it, why is this a problem?  Consider if you pass -1 as "unsigned"
to ConstantInt and try to make an i128.  In this case, you'll get 64-bits set. 
If you pass -1 as signed to constant int, you get 128 bits set.

The signedness flag is only useful for constantints > 64 bits (the slow 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 cs.uiuc.edu  Sun Nov 16 01:41:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Nov 2008 01:41:36 -0600
Subject: [LLVMbugs] [Bug 3075] New: Meaningless warning compiling standard
	code
Message-ID: <bug-3075-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3075

           Summary: Meaningless warning compiling standard code
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


#include <stdarg.h>
#include <stdio.h>

void foo (const char *msg, ...)
{
  va_list ap;

  va_start (ap, msg);
  vfprintf (stderr, msg, ap);
  va_end (ap);
}

gives:

$ clang /tmp/stdarg.c
/tmp/stdarg.c:8:3: warning: passing 'char const *' discards qualifiers,
expected '__builtin_va_list'
  va_start (ap, msg);
  ^~~~~~~~
1 diagnostic generated.

If it helps, the bit after the headers is preprocessed by clang to

void foo (const char *msg, ...)
{
  va_list ap;

  __builtin_stdarg_start((ap), (msg));
  vfprintf ((&__sF[2]), msg, ap);
  __builtin_va_end (ap);
}


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 16 01:52:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Nov 2008 01:52:51 -0600
Subject: [LLVMbugs] [Bug 3076] New: LLVM doesn't fold global initializers
Message-ID: <bug-3076-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3076

           Summary: LLVM doesn't fold global initializers
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Running opt -std-compile-opts on this doesn't fold the global initializer:

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-f80:128:128"
target triple = "x86_64-unknown-linux-gnu"
@a = global i32 trunc (i64 sdiv (i64 ptrtoint (i32* inttoptr (i32 100 to i32*)
to i64), i64 4) to i32), align 4             ; <i32*> [#uses=0]

This is with LLVM svn, from a few days ago.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 16 01:55:20 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Nov 2008 01:55:20 -0600
Subject: [LLVMbugs] [Bug 3077] New: -pedantic-errors unusable on NetBSD
Message-ID: <bug-3077-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3077

           Summary: -pedantic-errors unusable on NetBSD
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


-pedantic-errors fails on code that compiles cleanly with GCC, because of
constructs in system headers.  For example:

lib/cfe/semantics.c:110:22: error: extension used
COMPILE_TIME_ASSERT (offsetof (a, b) == 0);
                     ^

so something seems wrong with diagnosing offsetof; clearly not an extension.

Also

/usr/include/sys/bswap.h:20:28: error: extension used
uint16_t bswap16(uint16_t) __RENAME(__bswap16) __attribute__((__const__));
                           ^

/usr/include/sys/endian.h:203:1: warning: extension used
__GEN_ENDIAN_ENC(16, be)
^

__RENAME is used in sysheaders and expands to

  __asm ("_foo")

The other macro is defined like so:

#define __GEN_ENDIAN_DEC(bits, endian) \
static __inline __unused uint ## bits ## _t \
endian ## bits ## dec(const void *buf) \
{ \
        uint ## bits ## _t u; \
        __builtin_memcpy(&u, buf, sizeof(u)); \
        return endian ## bits ## toh (u); \
}


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 16 05:04:23 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Nov 2008 05:04:23 -0600
Subject: [LLVMbugs] [Bug 3078] New: +Checks build fails in ArgPromotion
Message-ID: <bug-3078-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3078

           Summary: +Checks build fails in ArgPromotion
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


During make check with ENABLE_EXPENSIVE_CHECKS=1:

llvm-as <
/home/edwin/llvm-bootstrap/llvm/test/CodeGen/Generic/GC/argpromotion.ll |  opt
-anders-aa -argpromotion

/usr/include/c++/4.3/debug/safe_iterator.h:180:error: attempt to
    dereference a past-the-end iterator.

Objects involved in the operation:
iterator "this" @ 0x0x7fff7e92cc10 {
type =
N11__gnu_debug14_Safe_iteratorIN9__gnu_cxx17__normal_iteratorIPKmNSt6__norm6vectorImSaImEEEEENSt7__debug6vectorI
mS7_EEEE (constant iterator);
  state = past-the-end;
  references sequence with type `NSt7__debug6vectorImSaImEEE' @
0x0x7fff7e92cc10
}

Program received signal SIGABRT, Aborted.
0x00007fb075620ed5 in raise () from /lib/libc.so.6
#1  0x00007fb0756223f3 in abort () from /lib/libc.so.6
#2  0x00007fb075e399f9 in __gnu_debug::_Error_formatter::_M_error
(this=0x7fff7e92c4e0) at ../../../../gcc/libstdc++-v3
/src/debug.cc:527
#3  0x0000000000a45750 in
__gnu_debug::_Safe_iterator<__gnu_cxx::__normal_iterator<unsigned long const*,
std::__norm::v
ector<unsigned long, std::allocator<unsigned long> > >,
std::__debug::vector<unsigned long, std::allocator<unsigned lon
g> > >::operator* (this=0x7fff7e92cc10) at
/usr/include/c++/4.3/debug/safe_iterator.h:178
#4  0x0000000000a375e9 in DoPromotion (this=0x2348b20, F=0x2349da0,
ArgsToPromote=@0x7fff7e92d390, ByValArgsToTransform
=@0x7fff7e92d330) at ArgumentPromotion.cpp:570
#5  0x0000000000a3a543 in PromoteArguments (this=0x2348b20, CGN=0x2350ba0) at
ArgumentPromotion.cpp:187
#6  0x0000000000a3a5ea in runOnSCC (this=0x2348b20, SCC=@0x7fff7e92d708) at
ArgumentPromotion.cpp:101
#7  0x0000000000c835d4 in runOnModule (this=0x2350540, M=@0x234d3c0) at
CallGraphSCCPass.cpp:102
#8  0x0000000000dc817f in llvm::MPPassManager::runOnModule (this=0x234b060,
M=@0x234d3c0) at PassManager.cpp:1393
#9  0x0000000000dc96f8 in llvm::PassManagerImpl::run (this=0x234aa70,
M=@0x234d3c0) at PassManager.cpp:1467
#10 0x0000000000dc975b in llvm::PassManager::run (this=0x7fff7e92d9e0,
M=@0x234d3c0) at PassManager.cpp:1499
#11 0x000000000098ace3 in main (argc=4, argv=0x7fff7e92dd78) at opt.cpp:590

The problem is that at ArgumentPromotion.cpp:570, SI->size() is 0, thus you are
not allowed to dereference ->begin()


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 16 11:16:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Nov 2008 11:16:53 -0600
Subject: [LLVMbugs] [Bug 3076] LLVM doesn't fold global initializers
In-Reply-To: <bug-3076-206@http.llvm.org/bugs/>
Message-ID: <200811161716.mAGHGrK3016221@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3076


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-16 11:16:52 ---
Opt shouldn't do this, the front-end which generates it should do this sort of
thing.

-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 cs.uiuc.edu  Sun Nov 16 11:21:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Nov 2008 11:21:46 -0600
Subject: [LLVMbugs] [Bug 3078] +Checks build fails in ArgPromotion
In-Reply-To: <bug-3078-206@http.llvm.org/bugs/>
Message-ID: <200811161721.mAGHLkDA016417@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3078


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

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




--- Comment #2 from T??r??k Edwin <edwintorok at gmail.com>  2008-11-16 11:21:46 ---
Fixed commited in SVN r59416.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 16 11:23:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Nov 2008 11:23:01 -0600
Subject: [LLVMbugs] [Bug 2155] Debug+Checks make check fails on x86
In-Reply-To: <bug-2155-206@http.llvm.org/bugs/>
Message-ID: <200811161723.mAGHN1Fj016483@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2155


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

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




--- Comment #4 from T??r??k Edwin <edwintorok at gmail.com>  2008-11-16 11:23:00 ---
This was fixed with SVN r59416 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 cs.uiuc.edu  Sun Nov 16 11:50:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Nov 2008 11:50:12 -0600
Subject: [LLVMbugs] [Bug 3072] CMake build fails with LLVM in a subdirectory
In-Reply-To: <bug-3072-206@http.llvm.org/bugs/>
Message-ID: <200811161750.mAGHoCRZ017262@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3072


??scar Fuentes <ofv at wanadoo.es> changed:

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




--- Comment #4 from ??scar Fuentes <ofv at wanadoo.es>  2008-11-16 11:50:12 ---
No need to apply the patch. It was committed to svn yestereday.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 16 12:36:57 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Nov 2008 12:36:57 -0600
Subject: [LLVMbugs] [Bug 3077] -pedantic-errors unusable on NetBSD
In-Reply-To: <bug-3077-206@http.llvm.org/bugs/>
Message-ID: <200811161836.mAGIavjK018628@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3077


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-16 12:36:56 ---
This should fix it, please verify:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081110/009291.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 cs.uiuc.edu  Sun Nov 16 12:44:10 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Nov 2008 12:44:10 -0600
Subject: [LLVMbugs] [Bug 3075] Meaningless warning compiling standard code
In-Reply-To: <bug-3075-206@http.llvm.org/bugs/>
Message-ID: <200811161844.mAGIiAc3018857@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3075


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-16 12:44:10 ---
Thanks Neil!
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081110/009292.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 cs.uiuc.edu  Sun Nov 16 22:48:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 16 Nov 2008 22:48:00 -0600
Subject: [LLVMbugs] [Bug 3079] New: clang doesn' t know how to generate code
 for constant remainder (%) ops
Message-ID: <bug-3079-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3079

           Summary: clang doesn't know how to generate code for constant
                    remainder (%) ops
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: pcwalton at cs.ucla.edu
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2211)
 --> (http://llvm.org/bugs/attachment.cgi?id=2211)
Proposed patch.

There's an oversight in clang's CodeGen module in which constant * and /
operations are handled, but % (remainder/mod) isn't. The attached patch fixes
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 cs.uiuc.edu  Mon Nov 17 07:49:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 07:49:17 -0600
Subject: [LLVMbugs] [Bug 3080] New: LegalizeTypes unconditionally produces
 addc/ adde nodes for shl x, 1
Message-ID: <bug-3080-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3080

           Summary: LegalizeTypes unconditionally produces addc/adde nodes
                    for shl x, 1
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: richard at xmos.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2212)
 --> (http://llvm.org/bugs/attachment.cgi?id=2212)
Proposed patch

Running llc -march=xcore on the following:

define i64 @test(i64 %a) {
        %result = shl i64 %a, 1
        ret i64 %result
}

results in the following failure:

Cannot yet select: 0x161b958: i32,flag = adde 0x161bf58, 0x161bf58, 0x161c058:1

It looks like the problem lies in LegalizeTypes: it shouldn't generate
ADDC/ADDE nodes if they are not supported by the target. A patch to fix this is
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 cs.uiuc.edu  Mon Nov 17 10:47:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 10:47:12 -0600
Subject: [LLVMbugs] [Bug 2843] llvm-gcc has -Wformat-nonliteral on by
 default, can'  t switch it off
In-Reply-To: <bug-2843-206@http.llvm.org/bugs/>
Message-ID: <200811171647.mAHGlCUY008676@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2843


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

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




--- Comment #4 from Rafael ??vila de Esp??ndola <rafael.espindola at gmail.com>  2008-11-17 10:47:11 ---
It looks like this was not fully fixed. Building llvm-gcc @56905 and doing "cc1
test.i" I get

cc1: warning: -Wformat-security ignored without -Wformat

With a newer revision (58836), I get

test.i:7: warning: format not a string literal and no format arguments

where test.i is

----------------------------------------------------------
extern int printf (__const char *__restrict __format, ...);

extern const char *foo();

int main(void) {
  const char *the_string = foo();
  printf(the_string);
}
--------------------------------------------------------


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 17 11:07:13 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 11:07:13 -0600
Subject: [LLVMbugs] [Bug 3030] format not a string literal and no format
	arguments
In-Reply-To: <bug-3030-206@http.llvm.org/bugs/>
Message-ID: <200811171707.mAHH7DIC009664@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3030


Nick Lewycky <nlewycky at google.com> changed:

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




--- Comment #5 from Nick Lewycky <nlewycky at google.com>  2008-11-17 11:07:12 ---


*** This bug has been marked as a duplicate of bug 2843 ***


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 17 11:19:35 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 11:19:35 -0600
Subject: [LLVMbugs] [Bug 3067] minor error in gcc/Makefile.in: misplaced
	brace
In-Reply-To: <bug-3067-206@http.llvm.org/bugs/>
Message-ID: <200811171719.mAHHJZ4K010117@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3067


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

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




--- Comment #1 from Evan Cheng <evan.cheng at apple.com>  2008-11-17 11:19:34 ---
Are you sure?
ifneq ($(LLVMOBJDIR)),$(LLVMSRCDIR))

That would be 3 left parentheses and 4 right 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 cs.uiuc.edu  Mon Nov 17 12:06:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 12:06:17 -0600
Subject: [LLVMbugs] [Bug 3081] New: "Whoops,
 an exception leaked out of bugpoint. This is a bug in bugpoint!"
Message-ID: <bug-3081-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3081

           Summary: "Whoops, an exception leaked out of bugpoint.  This is a
                    bug in bugpoint!"
           Product: tools
           Version: 2.4
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: bugpoint
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sliedes at cc.hut.fi
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2213)
 --> (http://llvm.org/bugs/attachment.cgi?id=2213)
Bitcode file to reproduce the bug with

Running bugpoint on the attached file gives several occurrences of the message
mentioned in summary. Run bugpoint with parameters "-mergereturn -loop-extract
-jump-threading picosat.bc". I did this on amd64, don't know if it happens on
32-bit llvm:

------------------------------------------------------------
$ bugpoint -mergereturn -loop-extract -jump-threading picosat.bc
Read input file      : 'picosat.bc'
*** All input ok
Running selected passes on program to test for crash: Whoops, an exception
leaked out of bugpoint.  This is a bug in bugpoint!
Exited with error code '1'

*** Debugging optimizer crash!
Checking to see if these passes crash: -mergereturn -loop-extract
-jump-threading: Whoops, an exception leaked out of bugpoint.  This is a bug in
bugpoint!
Exited with error code '1'
[...]
------------------------------------------------------------

Ask me if you need the entire output from bugpoint.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 17 12:23:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 12:23:12 -0600
Subject: [LLVMbugs] [Bug 3082] New: loop-extractor infinite memory usage
Message-ID: <bug-3082-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3082

           Summary: loop-extractor infinite memory usage
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2214)
 --> (http://llvm.org/bugs/attachment.cgi?id=2214)
testcase

This is a reduction of the testcase in bug 3081. This runs out of memory:

  opt -mergereturn -loop-extract -jump-threading bugpoint-reduced-simplified.bc

Here's the stack trace:

#0  0x086747cc in
llvm::StringMapEntry<llvm::Value*>::Create<llvm::MallocAllocator, void*> (
    KeyStart=0x26b21d04
"bb190.i.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exi"..., 
    KeyEnd=0x26b2e779 "", Allocator=@0x202143ac, InitVal=0x0)
    at /usr/local/src/llvm/include/llvm/ADT/StringMap.h:174
#1  0x08675490 in llvm::StringMap<llvm::Value*,
llvm::MallocAllocator>::GetOrCreateValue<void*> (this=0x20214398, 
    KeyStart=0x26b21d04
"bb190.i.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exi"..., 
    KeyEnd=0x26b2e779 "", Val=0x0)
    at /usr/local/src/llvm/include/llvm/ADT/StringMap.h:362
#2  0x0867551f in llvm::StringMap<llvm::Value*,
llvm::MallocAllocator>::GetOrCreateValue (this=0x20214398, 
    KeyStart=0x26b21d04
"bb190.i.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exi"..., 
    KeyEnd=0x26b2e779 "")
    at /usr/local/src/llvm/include/llvm/ADT/StringMap.h:379
#3  0x08674f2e in llvm::ValueSymbolTable::createValueName (this=0x20214398, 
    NameStart=0x26b21d04
"bb190.i.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exi"..., 
    NameLen=51829, V=0x20233a10) at ValueSymbolTable.cpp:97
#4  0x08673472 in llvm::Value::setName (this=0x20233a10, 
    NameStr=0x26b21d04
"bb190.i.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exitStub.exi"..., 
    NameLen=51829) at Value.cpp:225
#5  0x086736e2 in llvm::Value::setName (this=0x20233a10, name=@0xffffd0d8)
    at Value.cpp:166
#6  0x085f564e in BasicBlock (this=0x20233a10, Name=@0xffffd0d8, 
    NewParent=0x20214458, InsertBefore=0x0) at BasicBlock.cpp:92
#7  0x083afaca in llvm::BasicBlock::Create (Name=@0xffffd0d8, 
    Parent=0x20214458, InsertBefore=0x0)
    at /usr/local/src/llvm/include/llvm/BasicBlock.h:82
#8  0x084f799e in (anonymous
namespace)::CodeExtractor::emitCallAndSwitchStatement (this=0xffffd2dc,
newFunction=0x20214458, codeReplacer=0x202140b0, 
    inputs=@0xffffd22c, outputs=@0xffffd220) at CodeExtractor.cpp:456
#9  0x084f91f1 in (anonymous namespace)::CodeExtractor::ExtractCodeRegion (
    this=0xffffd2dc, code=@0x1fe6ba98) at CodeExtractor.cpp:660
#10 0x084f967f in llvm::ExtractLoop (DT=@0x87e9500, L=0x1fe6ba88, 
    AggregateArgs=false) at CodeExtractor.cpp:733
#11 0x083c74c6 in (anonymous namespace)::LoopExtractor::runOnFunction (
    this=0x87e86a0, F=@0x20213ef8) at LoopExtractor.cpp:124
#12 0x08651634 in llvm::FPPassManager::runOnFunction (this=0x87e83c0, 
    F=@0x20213ef8) at PassManager.cpp:1323
#13 0x086517de in llvm::FPPassManager::runOnModule (this=0x87e83c0, 
    M=@0x87e7088) at PassManager.cpp:1346
#14 0x08651301 in llvm::MPPassManager::runOnModule (this=0x87e71a0, 
    M=@0x87e7088) at PassManager.cpp:1393
#15 0x0865149e in llvm::PassManagerImpl::run (this=0x87e7db8, M=@0x87e7088)
    at PassManager.cpp:1467
#16 0x086514f9 in llvm::PassManager::run (this=0xffffd568, M=@0x87e7088)
    at PassManager.cpp:1499
#17 0x0837863b in main (argc=5, argv=0xffffd764) at opt.cpp:588


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 17 13:12:17 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 13:12:17 -0600
Subject: [LLVMbugs] [Bug 3067] minor error in gcc/Makefile.in: misplaced
	brace
In-Reply-To: <bug-3067-206@http.llvm.org/bugs/>
Message-ID: <200811171912.mAHJCHCX014698@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3067


Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gohman at apple.com
             Status|RESOLVED                    |REOPENED
         Resolution|INVALID                     |




--- Comment #2 from Dan Gohman <gohman at apple.com>  2008-11-17 13:12:16 ---
It looks like this is what was intended:

ifneq ($(LLVMOBJDIR),$(LLVMSRCDIR))


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 17 14:04:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 14:04:44 -0600
Subject: [LLVMbugs] [Bug 3080] LegalizeTypes unconditionally produces
 addc/adde nodes for shl x , 1
In-Reply-To: <bug-3080-206@http.llvm.org/bugs/>
Message-ID: <200811172004.mAHK4iCV016501@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3080


Duncan Sands <baldrick at free.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |baldrick at free.fr
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Duncan Sands <baldrick at free.fr>  2008-11-17 14:04:41 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081117/069960.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 cs.uiuc.edu  Mon Nov 17 15:27:08 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 15:27:08 -0600
Subject: [LLVMbugs] [Bug 3083] New: JIT: provide Garbage Collector callback
 if run out of stubs
Message-ID: <bug-3083-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3083

           Summary: JIT: provide Garbage Collector callback if run out of
                    stubs
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu, nicolas.geoffray at lip6.fr


Created an attachment (id=2215)
 --> (http://llvm.org/bugs/attachment.cgi?id=2215)
testcase to show problem

Apply the attached patch to example/HowToUseJIT, and run:

"We just constructed this LLVM module:

; ModuleID = 'test'

define i32 @add1(i32 %AnArg) {
EntryBlock:
        %addresult = add i32 1, %AnArg          ; <i32> [#uses=1]
        ret i32 %addresult
}

define i32 @foo() {
EntryBlock:
        %add1 = tail call i32 @add1(i32 10)             ; <i32> [#uses=1]
        ret i32 %add1
}


Running foo: JIT ran out of memory for function stubs!
Aborted"

The JIT could provide a callback when it runs out of memory, so that the
Garbage Collector can be run, that could possibly free up some memory.
This would especially be useful for VMKit.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 17 15:28:27 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 15:28:27 -0600
Subject: [LLVMbugs] [Bug 3084] New: vmkit: JIT function stub leak
Message-ID: <bug-3084-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3084

           Summary: vmkit: JIT function stub leak
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu, nicolas.geoffray at lip6.fr
        Depends on: 3083


Attached testcase.

~/project/llvm-svn/vmkit/Release/bin/jnjvm Invoker
 JIT ran out of memory for function stubs!
 Aborted


Why this is important: when you run an application server and redeploy a
.war/.ear multiple times, often memory is leaked.
Sun's JVM runs out of Permgen space after 36 iterations, VMKit runs out of
function stubs.

VMKit could do better and not run out of memory :)


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 17 16:33:13 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 16:33:13 -0600
Subject: [LLVMbugs] [Bug 3085] New: "Instruction does not dominate all uses!"
Message-ID: <bug-3085-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3085

           Summary: "Instruction does not dominate all uses!"
           Product: libraries
           Version: 2.4
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sliedes at cc.hut.fi
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2216)
 --> (http://llvm.org/bugs/attachment.cgi?id=2216)
Test case

Running

opt bugpoint-reduced-simplified.bc -loop-extract-single -loop-rotate
-loop-reduce -argpromotion" 

on the attached file causes the error message "Instruction does not dominate
all uses!". Platform is amd64:

$ opt bugpoint-reduced-simplified.bc -loop-extract-single -loop-rotate
-loop-reduce -argpromotion
Instruction does not dominate all uses!
        %lcollect.i.i.i.2 = phi i32 [ %lcollect.i.i.i.1, %bb60.i.i.i ]         
; <i32> [#uses=2]
        %tmp29 = sub i32 0, %lcollect.i.i.i.2           ; <i32> [#uses=1]
Instruction does not dominate all uses!
        %tmp29 = sub i32 0, %lcollect.i.i.i.2           ; <i32> [#uses=1]
        %tmp30 = add i32 %tmp29, 1              ; <i32> [#uses=1]
Instruction does not dominate all uses!
        %tmp30 = add i32 %tmp29, 1              ; <i32> [#uses=1]
        %tmp31 = lshr i32 %tmp30, 1             ; <i32> [#uses=1]
Instruction does not dominate all uses!
        %tmp31 = lshr i32 %tmp30, 1             ; <i32> [#uses=1]
        %tmp32 = add i32 %tmp31, 1              ; <i32> [#uses=1]
Broken module found, compilation aborted!
opt[0x7603af]
/lib/libc.so.6[0x7fce6afdaf60]
/lib/libc.so.6(gsignal+0x35)[0x7fce6afdaed5]
/lib/libc.so.6(abort+0x183)[0x7fce6afdc3f3]
opt[0x7232f0]
opt(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0x15b)[0x6ff8ab]
opt[0x613ef0]
opt(_ZN4llvm13MPPassManager11runOnModuleERNS_6ModuleE+0xd0)[0x6ff3d0]
opt(_ZN4llvm15PassManagerImpl3runERNS_6ModuleE+0x86)[0x6ff686]
opt(main+0xa5f)[0x49035f]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7fce6afc71a6]
opt(sinh+0x39)[0x485379]
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 cs.uiuc.edu  Mon Nov 17 18:29:57 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 18:29:57 -0600
Subject: [LLVMbugs] [Bug 3067] minor error in gcc/Makefile.in: misplaced
	brace
In-Reply-To: <bug-3067-206@http.llvm.org/bugs/>
Message-ID: <200811180029.mAI0TvVi025760@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3067


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

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




--- Comment #4 from Evan Cheng <evan.cheng at apple.com>  2008-11-17 18:29:56 ---
Fixed.
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081117/069981.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 cs.uiuc.edu  Mon Nov 17 19:32:03 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 19:32:03 -0600
Subject: [LLVMbugs] [Bug 3086] New: Segmentation fault in alias analysis
Message-ID: <bug-3086-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3086

           Summary: Segmentation fault in alias analysis
           Product: libraries
           Version: 2.4
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sliedes at cc.hut.fi
                CC: llvmbugs at cs.uiuc.edu


The attached bitcode when run with "opt -licm -loop-reduce" segfaults on amd64.

------------------------------------------------------------
$ opt bugpoint-reduced-simplified.bc -licm -loop-reduce
opt[0x7603af]
/lib/libc.so.6[0x7fe7bd569f60]
opt[0x62a09c]
opt(_ZNK4llvm8AliasSet14aliasesPointerEPKNS_5ValueEjRNS_13AliasAnalysisE+0x50)[0x623810]
opt(_ZN4llvm15AliasSetTracker22findAliasSetForPointerEPKNS_5ValueEj+0x92)[0x624402]
opt(_ZN4llvm15AliasSetTracker21getAliasSetForPointerEPNS_5ValueEjPb+0x240)[0x624fe0]
opt(_ZN4llvm15AliasSetTracker3addEPNS_8LoadInstE+0x8b)[0x6254db]
opt(_ZN4llvm15AliasSetTracker3addEPNS_11InstructionE+0x8d)[0x6255dd]
opt(_ZN4llvm15AliasSetTracker3addERNS_10BasicBlockE+0x32)[0x6256f2]
opt[0x552d55]
opt(_ZN4llvm13LPPassManager13runOnFunctionERNS_8FunctionE+0x3ab)[0x6405cb]
opt(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0x15b)[0x6ff8ab]
opt(_ZN4llvm13FPPassManager11runOnModuleERNS_6ModuleE+0xa6)[0x6ffb86]
opt(_ZN4llvm13MPPassManager11runOnModuleERNS_6ModuleE+0xd0)[0x6ff3d0]
opt(_ZN4llvm15PassManagerImpl3runERNS_6ModuleE+0x86)[0x6ff686]
opt(main+0xa5f)[0x49035f]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7fe7bd5561a6]
opt(sinh+0x39)[0x485379]
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 cs.uiuc.edu  Mon Nov 17 19:41:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 19:41:25 -0600
Subject: [LLVMbugs] [Bug 3087] New: "Value does not have a node in the
	points-to graph"
Message-ID: <bug-3087-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3087

           Summary: "Value does not have a node in the points-to graph"
           Product: libraries
           Version: 2.4
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Interprocedural Analyses
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sliedes at cc.hut.fi
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2218)
 --> (http://llvm.org/bugs/attachment.cgi?id=2218)
Test case

Running "opt -anders-aa -loopsimplify -dse" on the attached bitcode causes an
assertion failure with the message "Value does not have a node in the points-to
graph".

------------------------------------------------------------
$ opt -anders-aa -loopsimplify -dse bugpoint-reduced-simplified.bc
        %p.2.ph.ph = phi %struct.Cls** [ null, %entry ], [ %1,
%bb38.outer.loopexit ]           ; <%struct.Cls**> [#uses=1]

opt: /home/sliedes/llvm/src/llvm-2.4/lib/Analysis/IPA/Andersens.cpp:508:
unsigned int<unnamed>::Andersens::getNode(llvm::Value*): Assertion `0 && "Value
does not have a node in the points-to graph!"' failed.
opt[0x7603af]
/lib/libc.so.6[0x7f2bf0baef60]
/lib/libc.so.6(gsignal+0x35)[0x7f2bf0baeed5]
/lib/libc.so.6(abort+0x183)[0x7f2bf0bb03f3]
/lib/libc.so.6(__assert_fail+0xe9)[0x7f2bf0ba7dc9]
opt[0x5f8154]
opt[0x5f87f7]
opt(_ZN4llvm24MemoryDependenceAnalysis13getDependencyEPNS_11InstructionES2_PNS_10BasicBlockE+0x570)[0x649380]
opt[0x4e8af8]
opt[0x4e8d8a]
opt(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0x15b)[0x6ff8ab]
opt(_ZN4llvm13FPPassManager11runOnModuleERNS_6ModuleE+0xa6)[0x6ffb86]
opt(_ZN4llvm13MPPassManager11runOnModuleERNS_6ModuleE+0xd0)[0x6ff3d0]
opt(_ZN4llvm15PassManagerImpl3runERNS_6ModuleE+0x86)[0x6ff686]
opt(main+0xa5f)[0x49035f]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7f2bf0b9b1a6]
opt(sinh+0x39)[0x485379]
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 cs.uiuc.edu  Mon Nov 17 19:47:42 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 19:47:42 -0600
Subject: [LLVMbugs] [Bug 3088] New: "Pass not initialized!"
Message-ID: <bug-3088-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3088

           Summary: "Pass not initialized!"
           Product: libraries
           Version: 2.4
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sliedes at cc.hut.fi
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2219)
 --> (http://llvm.org/bugs/attachment.cgi?id=2219)
Test case

Running "opt bugpoint-reduced-simplified.bc -sretpromotion -lowerallocs" on the
attached bitcode causes an assertion failure with the message "Pass not
initialized!". Architecture is amd64.

------------------------------------------------------------
$ opt bugpoint-reduced-simplified.bc -sretpromotion -lowerallocs
opt:
/home/sliedes/llvm/src/llvm-2.4/lib/Transforms/Utils/LowerAllocations.cpp:103:
virtual bool<unnamed>::LowerAllocations::runOnBasicBlock(llvm::BasicBlock&):
Assertion `MallocFunc && FreeFunc && "Pass not initialized!"' failed.
opt[0x7603af]
/lib/libc.so.6[0x7fc1ea32ef60]
/lib/libc.so.6(gsignal+0x35)[0x7fc1ea32eed5]
/lib/libc.so.6(abort+0x183)[0x7fc1ea3303f3]
/lib/libc.so.6(__assert_fail+0xe9)[0x7fc1ea327dc9]
opt[0x5d3cb9]
opt[0x6fefb8]
opt(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0x15b)[0x6ff8ab]
opt[0x613ef0]
opt(_ZN4llvm13MPPassManager11runOnModuleERNS_6ModuleE+0xd0)[0x6ff3d0]
opt(_ZN4llvm15PassManagerImpl3runERNS_6ModuleE+0x86)[0x6ff686]
opt(main+0xa5f)[0x49035f]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7fc1ea31b1a6]
opt(sinh+0x39)[0x485379]
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 cs.uiuc.edu  Mon Nov 17 19:56:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 19:56:12 -0600
Subject: [LLVMbugs] [Bug 3089] New: Apparent miscompilation with
 -loop-index-split -loop-rotate -loop-reduce
Message-ID: <bug-3089-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3089

           Summary: Apparent miscompilation with -loop-index-split -loop-
                    rotate -loop-reduce
           Product: libraries
           Version: 2.4
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sliedes at cc.hut.fi
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2220)
 --> (http://llvm.org/bugs/attachment.cgi?id=2220)
Test case

I believe the attached bitcode is miscompiled with "opt -loop-index-split
-loop-rotate -loop-reduce bugpoint-tooptimize.bc" (bugpointed).

While I do not know for certain that the code is correct, it seems to run
without segfaulting with almost any llvm options, including all the default -O
levels. It also is production code used by lots of different people (usually
compiled with gcc). However compiling with the mentioned passes causes the code
to segfault. Architecture is amd64.

Please tell me if you need more information (source code, the whole program,
whatever).


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 17 20:07:57 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 20:07:57 -0600
Subject: [LLVMbugs] [Bug 3090] New: Apparent miscompilation with
 -loop-index-split -loop-rotate -indvars -domtree
Message-ID: <bug-3090-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3090

           Summary: Apparent miscompilation with -loop-index-split -loop-
                    rotate -indvars -domtree
           Product: libraries
           Version: 2.4
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Loop Optimizer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sliedes at cc.hut.fi
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2221)
 --> (http://llvm.org/bugs/attachment.cgi?id=2221)
Test case

Not sure if this is related to #3089, however in this case the required bitcode
is smaller and the options slightly different.

I suspect the options "-loop-index-split -loop-rotate -indvars -domtree" to opt
miscompile the attached bitcode. The same caveats as in #3089 apply: I believe
the code is correct, but don't know for sure; it works with e.g. all the -O
levels in llvm, but the program segfaults with the given options.

Architecture is amd64.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 17 20:24:38 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 20:24:38 -0600
Subject: [LLVMbugs] [Bug 3069] The return value of
 llvm::sys::DynamicLibrary:: LoadLibraryPermanently is counter-intuitive
In-Reply-To: <bug-3069-206@http.llvm.org/bugs/>
Message-ID: <200811180224.mAI2Ocat029802@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3069


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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |evan.cheng at apple.com
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




--- Comment #1 from Evan Cheng <evan.cheng at apple.com>  2008-11-17 20:24:38 ---
Sorry, we are probably not going flip the return value. It'll just as likely to
upset existing code that uses 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 cs.uiuc.edu  Mon Nov 17 22:02:03 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 17 Nov 2008 22:02:03 -0600
Subject: [LLVMbugs] [Bug 3091] New: FunctionType::get should take an
	iterator pair
Message-ID: <bug-3091-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3091

           Summary: FunctionType::get should take an iterator pair
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: minor
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: viridia at gmail.com
                CC: llvmbugs at cs.uiuc.edu


I'd like to see a version of FunctionType::get() that takes an iterator pair,
in addition to the current method that takes a std::vector. This would make it
more consistent with the other APIs such as IRBuilder. Also, it would be good
if FunctionType could use a SmallVector instead of std::vector.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 18 03:54:44 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Nov 2008 03:54:44 -0600
Subject: [LLVMbugs] [Bug 3092] New: opt segfaults with -addreadattrs -die
 -inline -globalsmodref-aa -licm
Message-ID: <bug-3092-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3092

           Summary: opt segfaults with -addreadattrs -die -inline -
                    globalsmodref-aa -licm
           Product: libraries
           Version: trunk
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sliedes at cc.hut.fi
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2222)
 --> (http://llvm.org/bugs/attachment.cgi?id=2222)
Test case

When I run "opt -addreadattrs -die -inline -globalsmodref-aa -licm" on the
attached bitcode, opt crashes with a segmentation fault. Architecture is amd64.

------------------------------------------------------------
$ opt bugpoint-reduced-simplified.bc -addreadattrs -die -inline
-globalsmodref-aa -licm
opt[0x7603af]
/lib/libc.so.6[0x7f5696d4ef60]
opt[0x618be7]
opt[0x61a2e8]
opt(_ZN4llvm13MPPassManager11runOnModuleERNS_6ModuleE+0xd0)[0x6ff3d0]
opt(_ZN4llvm15PassManagerImpl3runERNS_6ModuleE+0x86)[0x6ff686]
opt(main+0xa5f)[0x49035f]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7f5696d3b1a6]
opt(sinh+0x39)[0x485379]
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 cs.uiuc.edu  Tue Nov 18 06:43:21 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Nov 2008 06:43:21 -0600
Subject: [LLVMbugs] [Bug 3093] New: clang: preprocessor crash
Message-ID: <bug-3093-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3093

           Summary: clang: preprocessor crash
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Using SVN r59519:
$ clang -E foo.c
clang[0x8c0fd9a]
Segmentation fault

$ cat foo.c
#ifdef FOO
/* /* */

If I remove the first space, it no longer crashes:
#ifdef 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 cs.uiuc.edu  Tue Nov 18 06:53:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Nov 2008 06:53:51 -0600
Subject: [LLVMbugs] [Bug 3094] New: ccc wrong order of linker flags
Message-ID: <bug-3094-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3094

           Summary: ccc wrong order of linker flags
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Basic
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


$ CCC_ECHO=1 ccc foo.c -lcheck
clang -emit-llvm-bc foo.c -o foo.o.bc
llc -relocation-model=pic -f -o foo.o.s foo.o.bc
as -o foo.o foo.o.s
c++ -o a.out -lcheck foo.o
foo.o: In function 'main'
(.text+0x11): undefined reference to 'srunner_create'

It should be:
$ c++ -o a.out foo.o -lcheck

Which builds a.out successfully.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 18 07:53:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Nov 2008 07:53:00 -0600
Subject: [LLVMbugs] [Bug 3095] New: @synchronized crashes when passing self
Message-ID: <bug-3095-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3095

           Summary: @synchronized crashes when passing self
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: thomascl at free.fr
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2224)
 --> (http://llvm.org/bugs/attachment.cgi?id=2224)
test case

Passing self to the @synchronized statement crashes clang.
Test case and crash log attached.

r59522.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 18 11:19:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Nov 2008 11:19:01 -0600
Subject: [LLVMbugs] [Bug 3093] clang: preprocessor crash
In-Reply-To: <bug-3093-206@http.llvm.org/bugs/>
Message-ID: <200811181719.mAIHJ1ti006773@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3093


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-18 11:18:59 ---
Fixed, thanks!
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081117/009407.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 cs.uiuc.edu  Tue Nov 18 11:58:11 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Nov 2008 11:58:11 -0600
Subject: [LLVMbugs] [Bug 3094] ccc wrong order of linker flags
In-Reply-To: <bug-3094-206@http.llvm.org/bugs/>
Message-ID: <200811181758.mAIHwBWu008331@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3094


Daniel Dunbar <daniel at zuster.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel at zuster.org
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Daniel Dunbar <daniel at zuster.org>  2008-11-18 11:58:10 ---
Fixed here:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081117/009408.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 cs.uiuc.edu  Tue Nov 18 12:43:27 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Nov 2008 12:43:27 -0600
Subject: [LLVMbugs] [Bug 3088] "Pass not initialized!"
In-Reply-To: <bug-3088-206@http.llvm.org/bugs/>
Message-ID: <200811181843.mAIIhRQe010176@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3088


Devang Patel <dpatel at apple.com> changed:

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




--- Comment #2 from Devang Patel <dpatel at apple.com>  2008-11-18 12:43:26 ---
Fixed. rev 59538.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 18 12:48:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Nov 2008 12:48:59 -0600
Subject: [LLVMbugs] [Bug 3096] New: clang: crash in raw_ostream during -E
	-verify
Message-ID: <bug-3096-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3096

           Summary: clang: crash in raw_ostream  during -E -verify
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: preprocessor
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


// RUN: clang -E -verify %s
#ifdef FOO // expected-error {{unterminated conditional directive}}
/* /* */


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 18 15:21:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 18 Nov 2008 15:21:01 -0600
Subject: [LLVMbugs] [Bug 3097] New: clang refuses wchar_t foo[]
Message-ID: <bug-3097-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3097

           Summary: clang refuses wchar_t foo[]
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


clang refuses wchar_t foo[] which is legal by:

ISO/IEC 9899:1999 (E),
6.7.8 Initialization clause 15,

An array with element type compatible with wchar_t may be initialized by
a wide string literal, optionally enclosed in braces. Successive wide
characters of the wide string literal (including the terminating null
wide character if there is room or if the array is of unknown size)
initialize the elements of the array.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 19 05:57:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Nov 2008 05:57:29 -0600
Subject: [LLVMbugs] [Bug 3099] New: multiple errors in
	gcc/config/rs6000/rs6000.c
Message-ID: <bug-3099-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3099

           Summary: multiple errors in gcc/config/rs6000/rs6000.c
           Product: tools
           Version: 2.4
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: go0 at go.ru
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2227)
 --> (http://llvm.org/bugs/attachment.cgi?id=2227)
build script and full log

I'm trying to cross-compile llvm-gcc for powerpc target on amd64/linux

The compilation stopped when compiling gcc/config/rs6000/rs6000.c

================== log fragment start ==========================
c++ -c -O2 -pipe -DIN_GCC -DCROSS_DIRECTORY_STRUCTURE -W -Wall -Wwrite-strings
-pedantic -Wno-long-long -Wno-variadic-macros -Wmissing-format-attribute
-DHAVE_CONFIG_H -Wno-unused -DTARGET_NAME=\"ppc-linux-gnu\" -DNDEBUG -I. -I.
-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc
-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/.
-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/../include
-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/../libcpp/include

-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/../libdecnumber
-I../libdecnumber -I/usr/include -DENABLE_LLVM -DNDEBUG -D_GNU_SOURCE
-D__STDC_LIMIT_MACROS   -I. -I.
-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc
-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/.
-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/../include
-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/../libcpp/include

-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/../libdecnumber
-I../libdecnumber -I/usr/include
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/llvm-debug.cpp
-o llvm-debug.o
gcc -c   -O2 -pipe -DIN_GCC -DCROSS_DIRECTORY_STRUCTURE  -W -Wall
-Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -pedantic
-Wno-long-long -Wno-variadic-macros -Wno-overlength-strings
-Wold-style-definition -Wmissing-format-attribute    -DHAVE_CONFIG_H -I. -I.
-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc
-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/.
-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/../include
-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/../libcpp/include

-I/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/../libdecnumber
-I../libdecnumber -I/usr/include -DENABLE_LLVM -DNDEBUG -D_GNU_SOURCE
-D__STDC_LIMIT_MACROS   \
               
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c
-o rs6000.o
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c:
In function 'rs6000_override_options':
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c:1565:
error: 'MASK_MACHO_DYNAMIC_NO_PIC' undeclared (first use in this function)
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c:1565:
error: (Each undeclared identifier is reported only once
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c:1565:
error: for each function it appears in.)
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c:
In function 'rs6000_emit_prologue':
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c:16873:
warning: implicit declaration of function 'VECTOR_SAVE_INLINE'
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c:16874:
error: 'lr_already_set_up_for_pic' undeclared (first use in this function)
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c:
In function 'rs6000_output_mi_thunk':
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c:18333:
warning: implicit declaration of function 'add_compiler_branch_island'
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c:18333:
warning: assignment makes pointer from integer without a cast
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c:20026:8:
warning: extra tokens at end of #endif directive
In file included from
/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/llvm-gcc4.2-2.4.source/gcc/config/rs6000/rs6000.c:22322:
./gt-rs6000.h: At top level:
./gt-rs6000.h:138: error: 'branch_island_list' undeclared here (not in a
function)
make[2]: *** [rs6000.o] Error 1
make[2]: Leaving directory
`/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/build/gcc'
make[1]: *** [all-gcc] Error 2
make[1]: Leaving directory
`/tmp/portage/cross-ppc-linux-gnu/gcc-4.2.4-r2/work/build'
================== log fragment  end  ==========================

IMHO
1) MASK_MACHO_DYNAMIC_NO_PIC is not defined anywhere, though used twice. I
looked for it in the complete patch/build tree and in /usr/include

2) "#endif llvm" is present only in this file and breaks C rules and apple/llvm
coding style

3) Other errors are present

i enclose ebuild script and full log file

my aim is not to find lots of errors in llvm source code but to compile
PowerPC-capable llvm-gcc decompiler on my computer


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 19 08:56:04 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Nov 2008 08:56:04 -0600
Subject: [LLVMbugs] [Bug 3100] New: llvm-g++ is unable to devirtualize a
	function call
Message-ID: <bug-3100-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3100

           Summary: llvm-g++ is unable to devirtualize a function call
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-g++
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Testcase from the g++ testsuite
-----------------------------------------
struct S { S(); virtual void xyzzy(); };
inline void foo(S *s) { s->xyzzy(); }
void bar() { S s; foo(&s); }
----------------------------------------

With gcc we got

call    _ZN1S5xyzzyEv

with llvm-g++ we got

call    *%rax

there rax contains the pointer fetched from the vtable. The problem is that
llvm-g++ knows what the vtable looks like, but doesn't communicate that to
llvm.

A simpler testcase also fails with llvm:

----------------------------------------
struct S { S() {}; virtual void xyzzy(); };
inline void foo(S *s) { s->xyzzy(); }
void bar() { S s; foo(&s); }
----------------------------------------

Now we have the constructor body and we do a bit better:
----------------------------
call    *_ZTV1S+16
---------------------------

but llvm-g++ only says

@_ZTV1S = external constant [3 x i32 (...)*]

so there is no way for llvm to const propagate the fetch.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 19 14:13:15 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Nov 2008 14:13:15 -0600
Subject: [LLVMbugs] [Bug 3101] New: Instcombine promotes i32 mul to i64 mul
Message-ID: <bug-3101-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3101

           Summary: Instcombine promotes i32 mul to i64 mul
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: tilmann.scheller at googlemail.com
                CC: llvmbugs at cs.uiuc.edu


Translating the following snippet into LLVM IR on x86-64

long mul(long a, long b)
{
  return ((int) a) * ((int) b);
}

with llvm-gcc -O0 results in the following code:

; ModuleID = 'mul.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"
target triple = "x86_64-unknown-linux-gnu"

define i64 @mul(i64 %a, i64 %b) nounwind {
entry:
        %a_addr = alloca i64            ; <i64*> [#uses=2]
        %b_addr = alloca i64            ; <i64*> [#uses=2]
        %retval = alloca i64            ; <i64*> [#uses=2]
        %0 = alloca i64         ; <i64*> [#uses=2]
        %"alloca point" = bitcast i32 0 to i32          ; <i32> [#uses=0]
        store i64 %a, i64* %a_addr
        store i64 %b, i64* %b_addr
        %1 = load i64* %a_addr, align 8         ; <i64> [#uses=1]
        %2 = trunc i64 %1 to i32                ; <i32> [#uses=1]
        %3 = load i64* %b_addr, align 8         ; <i64> [#uses=1]
        %4 = trunc i64 %3 to i32                ; <i32> [#uses=1]
        %5 = mul i32 %2, %4             ; <i32> [#uses=1]
        %6 = sext i32 %5 to i64         ; <i64> [#uses=1]
        store i64 %6, i64* %0, align 8
        %7 = load i64* %0, align 8              ; <i64> [#uses=1]
        store i64 %7, i64* %retval, align 8
        br label %return

return:         ; preds = %entry
        %retval1 = load i64* %retval            ; <i64> [#uses=1]
        ret i64 %retval1
}

Running opt -instcombine on this results in the following code:
; ModuleID = 'mulopt.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"
target triple = "x86_64-unknown-linux-gnu"

define i64 @mul(i64 %a, i64 %b) nounwind {
entry:
        %retval = alloca i64            ; <i64*> [#uses=2]
        %0 = mul i64 %a, %b             ; <i64> [#uses=1]
        %1 = trunc i64 %0 to i32                ; <i32> [#uses=1]
        %2 = sext i32 %1 to i64         ; <i64> [#uses=1]
        store i64 %2, i64* %retval, align 8
        br label %return

return:         ; preds = %entry
        %retval1 = load i64* %retval            ; <i64> [#uses=1]
        ret i64 %retval1
}

Obviously the generated code is not incorrect, but it likely results in less
efficient code on targets without a native 64-bit multiplication instruction.

As the CellSPU does not have such a native 64-bit multiplication instruction,
i64 muls are turned into library calls to the _muldi3 function of libgcc, which
is supposed to perform 64-bit multiplications in software by using several
32-bit multiplications. However, as libgcc is compiled with -O2 by default
(which includes instcombine) the 32-bit multiplications in _muldi3 are promoted
to 64-bit multiplications, which is clearly not useful in this 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 cs.uiuc.edu  Wed Nov 19 19:39:45 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Nov 2008 19:39:45 -0600
Subject: [LLVMbugs] [Bug 3102] New: inliner: "Function not in callgraph!"
 when running -inline -inline -loop-extract-single -loop-extract-single
 -globaldce
Message-ID: <bug-3102-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3102

           Summary: inliner: "Function not in callgraph!" when running -
                    inline -inline -loop-extract-single -loop-extract-single
                    -globaldce
           Product: libraries
           Version: 2.4
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Global Analyses
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sliedes at cc.hut.fi
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2228)
 --> (http://llvm.org/bugs/attachment.cgi?id=2228)
Test case

Running

opt -inline -inline -loop-extract-single -loop-extract-single -globaldce
bugpoint-reduced-simplified.bc

on the attached bitcode causes an assertion failure with the error message
"Function not in callgraph!".

Architecture is amd64.

------------------------------------------------------------
$ opt bugpoint-reduced-simplified.bc -inline -inline -loop-extract-single
-loop-extract-single -globaldce
opt: /home/sliedes/llvm/src/llvm-2.4/include/llvm/Analysis/CallGraph.h:103:
llvm::CallGraphNode* llvm::CallGraph::operator[](const llvm::Function*):
Assertion `I != FunctionMap.end() && "Function not in callgraph!"' failed.
opt[0x7603af]
/lib/libc.so.6[0x7f668d97ff60]
/lib/libc.so.6(gsignal+0x35)[0x7f668d97fed5]
/lib/libc.so.6(abort+0x183)[0x7f668d9813f3]
/lib/libc.so.6(__assert_fail+0xe9)[0x7f668d978dc9]
opt(_ZN4llvm14InlineFunctionENS_8CallSiteEPNS_9CallGraphEPKNS_10TargetDataE+0xb61)[0x5c8a91]
opt(_ZN4llvm7Inliner8runOnSCCERKSt6vectorIPNS_13CallGraphNodeESaIS3_EE+0x965)[0x4bc375]
opt[0x613d75]
opt(_ZN4llvm13MPPassManager11runOnModuleERNS_6ModuleE+0xd0)[0x6ff3d0]
opt(_ZN4llvm15PassManagerImpl3runERNS_6ModuleE+0x86)[0x6ff686]
opt(main+0xa5f)[0x49035f]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7f668d96c1a6]
opt(sinh+0x39)[0x485379]
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 cs.uiuc.edu  Wed Nov 19 21:13:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Nov 2008 21:13:36 -0600
Subject: [LLVMbugs] [Bug 2927] Calling a function from the JIT segfaults.
In-Reply-To: <bug-2927-206@http.llvm.org/bugs/>
Message-ID: <200811200313.mAK3DavT008051@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2927


??scar Fuentes <ofv at wanadoo.es> changed:

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




--- Comment #3 from ??scar Fuentes <ofv at wanadoo.es>  2008-11-19 21:13:36 ---
This seems fixed now.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 19 21:14:50 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 19 Nov 2008 21:14:50 -0600
Subject: [LLVMbugs] [Bug 2920] JIT uses rip-relative addressing for absolute
	addresses.
In-Reply-To: <bug-2920-206@http.llvm.org/bugs/>
Message-ID: <200811200314.mAK3EofM008177@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2920


??scar Fuentes <ofv at wanadoo.es> changed:

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




--- Comment #3 from ??scar Fuentes <ofv at wanadoo.es>  2008-11-19 21:14:50 ---
Seems 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 cs.uiuc.edu  Thu Nov 20 15:26:48 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 20 Nov 2008 15:26:48 -0600
Subject: [LLVMbugs] [Bug 3103] New: instcombine doesn't produce urem
Message-ID: <bug-3103-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3103

           Summary: instcombine doesn't produce urem
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nlewycky at google.com
                CC: nicholas at mxc.ca, llvmbugs at cs.uiuc.edu


This function:

  define i8 @rem2(i8 %x, i8 %y) {
   %A = udiv i8 %x, %y
     %B = mul i8 %A, %y
    %C = sub i8 %B, %x
    ret i8 %C
  }

is actually equivalent to:

  define i8 @rem1(i8 %x, i8 %y) {
    %A = urem i8 %x, %y
    %B = sub i8 0, %A
    ret i8 %B
  }

nobody used to care because the above code is ridiculous, but now SCEV is
producing such crazy equations and we should probably simplify them whenever
possible.


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

From bugzilla-daemon at cs.uiuc.edu  Thu Nov 20 17:51:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 20 Nov 2008 17:51:59 -0600
Subject: [LLVMbugs] [Bug 3083] JIT: provide Garbage Collector callback if
 run out of stubs
In-Reply-To: <bug-3083-206@http.llvm.org/bugs/>
Message-ID: <200811202351.mAKNpxNL016092@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3083


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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |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 cs.uiuc.edu  Fri Nov 21 00:20:23 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Nov 2008 00:20:23 -0600
Subject: [LLVMbugs] [Bug 3096] clang: crash in raw_ostream during -E -verify
In-Reply-To: <bug-3096-206@http.llvm.org/bugs/>
Message-ID: <200811210620.mAL6KNrJ030164@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3096


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-21 00:20:22 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081117/009585.html
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081117/009589.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 cs.uiuc.edu  Fri Nov 21 01:34:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Nov 2008 01:34:12 -0600
Subject: [LLVMbugs] [Bug 3103] instcombine doesn't produce urem
In-Reply-To: <bug-3103-206@http.llvm.org/bugs/>
Message-ID: <200811210734.mAL7YCmM000573@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3103


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #4 from Nick Lewycky <nicholas at mxc.ca>  2008-11-21 01:34:11 ---
Fixed in r59800.


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

From bugzilla-daemon at cs.uiuc.edu  Fri Nov 21 07:59:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Nov 2008 07:59:46 -0600
Subject: [LLVMbugs] [Bug 2854] EH Lowering code confused by optimizer
In-Reply-To: <bug-2854-206@http.llvm.org/bugs/>
Message-ID: <200811211359.mALDxkdN025754@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2854


varth <nicolas.geoffray at lip6.fr> 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 cs.uiuc.edu  Fri Nov 21 10:25:22 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Nov 2008 10:25:22 -0600
Subject: [LLVMbugs] [Bug 3105] New: TwoAddressInstructionPass fails to
 correctly handle instruction with multiple results and constraints
Message-ID: <bug-3105-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3105

           Summary: TwoAddressInstructionPass fails to correctly handle
                    instruction with multiple results and constraints
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: richard at xmos.com
                CC: llvmbugs at cs.uiuc.edu


Running llc -march=xcore -mcpu=xs1b-generic on the following:

define i64 @test(i32 %a, i32 %b) {
entry:
        %0 = sext i32 %a to i64
        %1 = sext i32 %b to i64
        %2 = mul i64 %0, %1
        ret i64 %2
}

results in the following assertion:

llc:
/homelocal/richard/revb2/tools_llvm/src/lib/CodeGen/LiveIntervalAnalysis.cpp:443:
void llvm::LiveIntervals::handleVirtualRegisterDef(llvm::MachineBasicBlock*,
llvm::ilist_iterator<llvm::MachineInstr>, unsigned int, llvm::MachineOperand&,
unsigned int, llvm::LiveInterval&): Assertion `interval.containsOneValue()'
failed.

It looks like the problem is introduced by TwoAddressInstructionPass. The
output of -debug-only=twoaddrinstr is as follows:

********** REWRITING TWO-ADDR INSTRS **********
********** Function: test
        %reg1027<def> = MACCS_l4r %reg1028<def>, %reg1026<kill>, %reg1026,
%reg1024<kill>, %reg1025<kill>
                prepend:        %reg1027<def> = ADD_2rus %reg1026, 0
                rewrite to:     %reg1027<def> = MACCS_l4r %reg1028<def>,
%reg1027, %reg1027, %reg1024<kill>, %reg1025<kill>
                prepend:        %reg1028<def> = ADD_2rus %reg1027, 0
                rewrite to:     %reg1028<def> = MACCS_l4r %reg1028<def>,
%reg1028, %reg1028, %reg1024<kill>, %reg1025<kill>

It produces an instruction with more than one definition of the same register.
This is clearly incorrect. I believe it is this which causes the assertion.


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

From bugzilla-daemon at cs.uiuc.edu  Fri Nov 21 13:26:55 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Nov 2008 13:26:55 -0600
Subject: [LLVMbugs] [Bug 3095] @synchronized crashes when passing self
In-Reply-To: <bug-3095-206@http.llvm.org/bugs/>
Message-ID: <200811211926.mALJQtK4006620@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3095


Fariborz Jahanian <fjahanian at apple.com> changed:

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




--- Comment #2 from Fariborz Jahanian <fjahanian at apple.com>  2008-11-21 13:26:55 ---
This is now fixed in clang TOT.


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

From bugzilla-daemon at cs.uiuc.edu  Fri Nov 21 15:58:12 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Nov 2008 15:58:12 -0600
Subject: [LLVMbugs] [Bug 3106] New: Load/Store through pointer implies it
	isn't null
Message-ID: <bug-3106-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3106

           Summary: Load/Store through pointer implies it isn't null
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: code-quality
          Severity: normal
          Priority: P2
         Component: Scalar Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


LLVM should optimize this (gcc already gets it right):

void foo(int *P) {
  *P = 4;
  if (P == 0) bar();
}

This sort of things occurs in a variety of cases, one of which I care about is
in Clang diagnostic code.

-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 cs.uiuc.edu  Fri Nov 21 18:45:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Nov 2008 18:45:25 -0600
Subject: [LLVMbugs] [Bug 2387] setcc gets promoted from i1 to i32,
 but select remains i1 and triggers assert
In-Reply-To: <bug-2387-206@http.llvm.org/bugs/>
Message-ID: <200811220045.mAM0jPvl018181@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2387


Tilmann Scheller <tilmann.scheller at googlemail.com> changed:

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




--- Comment #3 from Tilmann Scheller <tilmann.scheller at googlemail.com>  2008-11-21 18:45:24 ---
Bug has been fixed by the new type legalization infrastructure.


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

From bugzilla-daemon at cs.uiuc.edu  Fri Nov 21 18:56:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 21 Nov 2008 18:56:32 -0600
Subject: [LLVMbugs] [Bug 3107] No patterns for certain truncate operations.
In-Reply-To: <bug-3107-206@http.llvm.org/bugs/>
Message-ID: <200811220056.mAM0uWIp018510@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3107


Tilmann Scheller <tilmann.scheller at googlemail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |llvmbugs at cs.uiuc.edu




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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 22 00:43:13 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 00:43:13 -0600
Subject: [LLVMbugs] [Bug 2992] Assertion failed: (Ty && "GEP indices
 invalid!"), function getGetElementPtr, file Constants.cpp, line 2147.
In-Reply-To: <bug-2992-206@http.llvm.org/bugs/>
Message-ID: <200811220643.mAM6hDr5029996@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2992


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #4 from Anders Carlsson <andersca at mac.com>  2008-11-22 00:43:12 ---
Now that we use Expr::Evaluate for CodeGen of constants this should not longer
be a problem (and I can't reproduce it)

I added the reduced test case to test/CodeGen/const-init.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 cs.uiuc.edu  Sat Nov 22 03:44:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 03:44:41 -0600
Subject: [LLVMbugs] [Bug 3108] New: Clang too quiet
Message-ID: <bug-3108-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3108

           Summary: Clang too quiet
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: NetBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: neil at daikokuya.co.uk
                CC: llvmbugs at cs.uiuc.edu


Clang complains about none of the following with a simple compile:

x;
const;
;
inline;

and even with -pedantic-errors it only catches two of them.  This is more
permissive than even 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 cs.uiuc.edu  Sat Nov 22 04:56:28 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 04:56:28 -0600
Subject: [LLVMbugs] [Bug 3109] New: "Doxygen" content missing
Message-ID: <bug-3109-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3109

           Summary: "Doxygen" content missing
           Product: Website
           Version: unspecified
          Platform: All
               URL: http://llvm.org/doxygen/
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Documentation
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rich at rd.gen.nz
                CC: llvmbugs at cs.uiuc.edu


There is a link to http://llvm.org/doxygen/ from http://llvm.org/docs/ (text of
link is "Doxygen"). Unfortunately the content at the URL seems to be missing.
Going there, I just get an empty directory listing.

The "classes" and "tarball" links do go somewhere, so maybe only part of the
Doxygen content is missing.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 22 05:08:07 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 05:08:07 -0600
Subject: [LLVMbugs] [Bug 3110] New: runFunction and fastcc
Message-ID: <bug-3110-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3110

           Summary: runFunction and fastcc
           Product: libraries
           Version: 2.4
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Target-Independent JIT
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rich at rd.gen.nz
                CC: llvmbugs at cs.uiuc.edu


When I upgraded from LLVM 2.3 to 2.4 recently I ran into a problem with calling
into fastcc functions using the JIT. I'm raising it here because I didn't have
this problem with 2.3. It could be a regression, or LLVM 2.4 could just be
better at letting me know that I'm using it wrong. :-)

I've attached the smallest testcase I could come up with. The attached code
will run fine while "func->setCallingConv(CallingConv::Fast)" is commented out.
Uncommenting it will cause a segfault, which didn't happen in 2.3.

I'm running Mac OS 10.4.11, compiling with GCC 4.0.1.

Cheers
Rich


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 22 05:27:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 05:27:25 -0600
Subject: [LLVMbugs] [Bug 3109] "Doxygen" content missing
In-Reply-To: <bug-3109-206@http.llvm.org/bugs/>
Message-ID: <200811221127.mAMBRPsp016872@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3109


Anton Korobeynikov <asl at math.spbu.ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |asl at math.spbu.ru
             Status|NEW                         |RESOLVED
         Resolution|                            |WORKSFORME




--- Comment #1 from Anton Korobeynikov <asl at math.spbu.ru>  2008-11-22 05:27:25 ---
There is a content. doxygen stuff is updating daily. It can be, that you hit
the page during such update :)


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 22 10:18:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 10:18:25 -0600
Subject: [LLVMbugs] [Bug 3027] llvm-gcc bootstrap fails: error: Link tests
 are not allowed after GCC_NO_EXECUTABLES
In-Reply-To: <bug-3027-206@http.llvm.org/bugs/>
Message-ID: <200811221618.mAMGIPGU027735@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3027


Anton Korobeynikov <asl at math.spbu.ru> changed:

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




--- Comment #12 from Anton Korobeynikov <asl at math.spbu.ru>  2008-11-22 10:18:20 ---
(In reply to comment #11)
> (In reply to comment #10)
> > (In reply to comment #5)
> > > Created an attachment (id=2189)
 --> (http://llvm.org/bugs/attachment.cgi?id=2189) [details] [details] [details]
> > > I believe this patch fixes the problem. It fixes the given testcase and some
> > > other simple test cases.
> > Now bootstrap was ok. Dan, please apply!
> 
> Applied:
> http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081110/069719.html
Thanks! Reapplied patch in
http://llvm.org/viewvc/llvm-project?rev=59872&view=rev


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 22 13:18:30 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 13:18:30 -0600
Subject: [LLVMbugs] [Bug 3079] clang doesn't know how to generate code for
 constant remainder ( %) ops
In-Reply-To: <bug-3079-206@http.llvm.org/bugs/>
Message-ID: <200811221918.mAMJIU2O001292@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3079


Anders Carlsson <andersca at mac.com> changed:

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




--- Comment #2 from Anders Carlsson <andersca at mac.com>  2008-11-22 13:18:30 ---
Now that we use the evaluator for constant exprs, this already works! Closing
as 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 cs.uiuc.edu  Sat Nov 22 13:45:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 13:45:51 -0600
Subject: [LLVMbugs] [Bug 2019] Crash in codegen with clang + inline asm
 using "+S" and "+D" operands
In-Reply-To: <bug-2019-206@http.llvm.org/bugs/>
Message-ID: <200811221945.mAMJjpSB002359@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2019


Anders Carlsson <andersca at mac.com> changed:

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




--- Comment #4 from Anders Carlsson <andersca at mac.com>  2008-11-22 13:45:51 ---
Both "+S" and "+D" have been implemented now and do the right thing, closing
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 cs.uiuc.edu  Sat Nov 22 13:48:58 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 13:48:58 -0600
Subject: [LLVMbugs] [Bug 3003] clang: does not treat expression involving
 struct pointer as a constant in array declaration
In-Reply-To: <bug-3003-206@http.llvm.org/bugs/>
Message-ID: <200811221948.mAMJmwql002587@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3003


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #2 from Anders Carlsson <andersca at mac.com>  2008-11-22 13:48:57 ---
This works fine now!


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 22 14:41:06 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 14:41:06 -0600
Subject: [LLVMbugs] [Bug 3111] New: inlining recursive function instead of
 tail call eliminating it
Message-ID: <bug-3111-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3111

           Summary: inlining recursive function instead of tail call
                    eliminating it
           Product: libraries
           Version: 1.0
          Platform: PC
        OS/Version: All
            Status: NEW
          Keywords: code-quality
          Severity: normal
          Priority: P2
         Component: Interprocedural Optimizations
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: clattner at apple.com
                CC: llvmbugs at cs.uiuc.edu


We unroll this function a bunch of times instead of tail call eliminating it,
this causes significant code bloat and reduced performance:

/* compute sum of 1..n */
unsigned sum(unsigned n)
{
 if (n == 0)
   return 0;
 else
   return n + sum(n-1);
}

This is also a regression from previous llvm releases.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 22 15:09:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 15:09:29 -0600
Subject: [LLVMbugs] [Bug 2525] clang doesn't accept non-ICE expressions in
 case statements ( gcc extension)
In-Reply-To: <bug-2525-206@http.llvm.org/bugs/>
Message-ID: <200811222109.mAML9TFb006210@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2525


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Anders Carlsson <andersca at mac.com>  2008-11-22 15:09:28 ---
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081117/009658.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 cs.uiuc.edu  Sat Nov 22 15:11:29 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 15:11:29 -0600
Subject: [LLVMbugs] [Bug 1957] initialization of a __complex__ number
 triggers an assert()
In-Reply-To: <bug-1957-206@http.llvm.org/bugs/>
Message-ID: <200811222111.mAMLBTHp006304@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1957


Anders Carlsson <andersca at mac.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 cs.uiuc.edu  Sat Nov 22 16:04:28 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 16:04:28 -0600
Subject: [LLVMbugs] [Bug 3112] New: bootstrapped llvm-g++ fails to find files
Message-ID: <bug-3112-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3112

           Summary: bootstrapped llvm-g++ fails to find files
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


On x86_64 linux, SVN r59876 fails 8 tests in FrontendC++:
FAIL:
/home/edwin/llvm-bootstrap/llvm/test/FrontendC++/2003-08-28-ForwardType.cpp
FAIL:
/home/edwin/llvm-bootstrap/llvm/test/FrontendC++/2003-11-04-CatchLabelName.cpp
FAIL:
/home/edwin/llvm-bootstrap/llvm/test/FrontendC++/2003-11-25-ReturningOpaqueByValue.cpp
FAIL:
/home/edwin/llvm-bootstrap/llvm/test/FrontendC++/2004-09-27-CompilerCrash.cpp
FAIL:
/home/edwin/llvm-bootstrap/llvm/test/FrontendC++/2005-02-27-PlacementArrayNewCrash.cpp
FAIL: /home/edwin/llvm-bootstrap/llvm/test/FrontendC++/2006-09-08-powi.cpp
FAIL: /home/edwin/llvm-bootstrap/llvm/test/FrontendC++/2007-04-31-TryCatch.cpp
FAIL:
/home/edwin/llvm-bootstrap/llvm/test/FrontendC++/2008-10-29-WrongOffset.cpp for
PR2917

$ cat foo.cpp:
/* testcase: llvm-g++ -c foo.cpp */
#include <new>

$ ~/llvm-bootstrap/install/bin/llvm-g++ foo.cpp
foo.cpp:1:15: error: new: No such file or directory

Using llvm-g++ from 2.4 works, and so did the llvm-g++ I bootstrapped
yesterday:
$ llvm-g++ -c foo.cpp 

The file is there:
$ find  ~/llvm-bootstrap/install/ -name new
/home/edwin/llvm-bootstrap/install/include/c++/4.2.1/new


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 22 16:31:21 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 16:31:21 -0600
Subject: [LLVMbugs] [Bug 3049] Assertion failed: (0 && "Unhandled cast"),
 function VisitCastExpr, file ExprConstant.cpp, line 146.
In-Reply-To: <bug-3049-206@http.llvm.org/bugs/>
Message-ID: <200811222231.mAMMVL3t009003@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3049


Anders Carlsson <andersca at mac.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 cs.uiuc.edu  Sat Nov 22 17:49:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 17:49:51 -0600
Subject: [LLVMbugs] [Bug 3113] New: "this->replaceAllUsesWith(this) is NOT
	valid!"
Message-ID: <bug-3113-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3113

           Summary: "this->replaceAllUsesWith(this) is NOT valid!"
           Product: libraries
           Version: 2.4
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Core LLVM classes
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sliedes at cc.hut.fi
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2235)
 --> (http://llvm.org/bugs/attachment.cgi?id=2235)
Test case

Running "opt bugpoint-reduced-simplified.bc -codegenprepare" on the attached
bitcode file produces an assertion failure in Value::replaceAllUsesWith() with
the message "this->replaceAllUsesWith(this) is NOT valid!":

------------------------------------------------------------
$ opt bugpoint-reduced-simplified.bc -codegenprepare
opt: /home/sliedes/llvm/src/llvm-2.4/lib/VMCore/Value.cpp:317: void
llvm::Value::replaceAllUsesWith(llvm::Value*): Assertion `New != this &&
"this->replaceAllUsesWith(this) is NOT valid!"' failed.
opt[0x7603af]
/lib/libc.so.6[0x7f3a05afcf60]
/lib/libc.so.6(gsignal+0x35)[0x7f3a05afced5]
/lib/libc.so.6(abort+0x183)[0x7f3a05afe3f3]
/lib/libc.so.6(__assert_fail+0xe9)[0x7f3a05af5dc9]
opt(_ZN4llvm5Value18replaceAllUsesWithEPS0_+0x16f)[0x7132df]
opt[0x4ded38]
opt[0x4df3f1]
opt[0x4dfab0]
opt(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0x15b)[0x6ff8ab]
opt(_ZN4llvm13FPPassManager11runOnModuleERNS_6ModuleE+0xa6)[0x6ffb86]
opt(_ZN4llvm13MPPassManager11runOnModuleERNS_6ModuleE+0xd0)[0x6ff3d0]
opt(_ZN4llvm15PassManagerImpl3runERNS_6ModuleE+0x86)[0x6ff686]
opt(main+0xa5f)[0x49035f]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7f3a05ae91a6]
opt(sinh+0x39)[0x485379]
Aborted
------------------------------------------------------------

Architecture is amd64.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 22 20:36:46 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 20:36:46 -0600
Subject: [LLVMbugs] [Bug 3114] New: emit-html messes up some tr entries
Message-ID: <bug-3114-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3114

           Summary: emit-html messes up some tr entries
           Product: clang
           Version: unspecified
          Platform: Macintosh
        OS/Version: MacOS X
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Basic
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: philip.aker at shaw.ca
                CC: llvmbugs at cs.uiuc.edu


Mac OS X 10.5.5/Intel

###

clang -version
Low Level Virtual Machine (http://llvm.org/):
  llvm version 2.5svn
  DEBUG build with assertions.

###

clang -emit-html -o CFBase.html
/System/Library/Frameworks/CoreFoundation.framework/Headers/CFBase.h

###

The output from the above call is messed up in that sporadic tr entries are
missing the opening &lt; and indented by 1 tab.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 22 20:51:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 20:51:39 -0600
Subject: [LLVMbugs] [Bug 3042] Seems to miscompile qmake with -O2
In-Reply-To: <bug-3042-206@http.llvm.org/bugs/>
Message-ID: <200811230251.mAN2pdTt017885@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3042


Parag Warudkar <parag.warudkar at gmail.com> changed:

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




--- Comment #3 from Parag Warudkar <parag.warudkar at gmail.com>  2008-11-22 20:51:38 ---
Not reproducible with latest svn llvm and llvm-gcc -  I changed the build to
use proper linux-g++ spec and hacked the spec file to use llvm-gcc and llvm-g++
this time though - so not clear if the fix was in LLVM/LLVM-GCC tree or is a Qt
Build system / code bug that is exposed with linux-llvm spec but not linux-g++
spec hacked to use llvm.

So 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 cs.uiuc.edu  Sat Nov 22 23:11:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 22 Nov 2008 23:11:00 -0600
Subject: [LLVMbugs] [Bug 3115] New: clang integer miscompilation
Message-ID: <bug-3115-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3115

           Summary: clang integer miscompilation
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Windows XP
            Status: NEW
          Severity: normal
          Priority: P2
         Component: AST
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r59898 on Ubuntu Hardy on x86.

The program below should print 1 instead of 0.

regehr at john-home:~/volatile/tmp61$ clang small.c -I/usr/include --emit-llvm -o
- | llvm-as | llc | as -o small.o ; gcc small.o -o small
regehr at john-home:~/volatile/tmp61$ ./small
0
regehr at john-home:~/volatile/tmp61$ cat small.c
#include <stdio.h>

unsigned int g_122;

int func_1 (void);
int func_1 (void)
{
  unsigned int l_19 = 1;
  if (1 ^ l_19 && 1) return 0;
  return 1;
}

int
main (void)
{
  printf ("%d\n", func_1());
  return 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 cs.uiuc.edu  Sun Nov 23 02:00:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Nov 2008 02:00:32 -0600
Subject: [LLVMbugs] [Bug 3110] runFunction and fastcc
In-Reply-To: <bug-3110-206@http.llvm.org/bugs/>
Message-ID: <200811230800.mAN80Wac027643@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3110


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #4 from Chris Lattner <clattner at apple.com>  2008-11-23 02:00:30 ---
This should fix it:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081117/070262.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 cs.uiuc.edu  Sun Nov 23 06:29:20 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Nov 2008 06:29:20 -0600
Subject: [LLVMbugs] [Bug 3116] New: SelectionDAG assertion fails: Binary
 operator types must match
Message-ID: <bug-3116-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3116

           Summary: SelectionDAG assertion fails: Binary operator types must
                    match
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


Using SVN r59908 on x86_64 linux:

$ ~/llvm-svn/llvm/Release/bin/llc bugpoint-reduced-simplified.bc
llc: SelectionDAG.cpp:2380: llvm::SDValue llvm::SelectionDAG::getNode(unsigned
int, llvm::MVT, llvm::SDValue, llvm::SDValue): Assertion `VT.isInteger() &&
N1.getValueType() == N2.getValueType() && N1.getValueType() == VT && "Binary
operator types must match!"' failed.
/home/edwin/llvm-svn/llvm/Release/bin/llc[0xc1770f]
/lib/libc.so.6[0x7f7b32457f60]
/lib/libc.so.6(gsignal+0x35)[0x7f7b32457ed5]
/lib/libc.so.6(abort+0x183)[0x7f7b324593f3]
/lib/libc.so.6(__assert_fail+0xe9)[0x7f7b32450dc9]
/home/edwin/llvm-svn/llvm/Release/bin/llc(_ZN4llvm12SelectionDAG7getNodeEjNS_3MVTENS_7SDValueES2_+0x44d)[0x90f42d]
/home/edwin/llvm-svn/llvm/Release/bin/llc(_ZN4llvm20SelectionDAGLowering11visitBinaryERNS_4UserEj+0xa8)[0x940918]
/home/edwin/llvm-svn/llvm/Release/bin/llc(_ZN4llvm16SelectionDAGISel16SelectBasicBlockEPNS_10BasicBlockENS_14ilist_iteratorINS_11InstructionEEES5_+0x22d)[0x95361d]
/home/edwin/llvm-svn/llvm/Release/bin/llc(_ZN4llvm16SelectionDAGISel20SelectAllBasicBlocksERNS_8FunctionERNS_15MachineFunctionEPNS_17MachineModuleInfoERKNS_15TargetInstrInfoE+0x2f2)[0x9539b2]
/home/edwin/llvm-svn/llvm/Release/bin/llc(_ZN4llvm16SelectionDAGISel13runOnFunctionERNS_8FunctionE+0x482)[0x954f82]
/home/edwin/llvm-svn/llvm/Release/bin/llc(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0x15b)[0xbb3d4b]
/home/edwin/llvm-svn/llvm/Release/bin/llc(_ZN4llvm23FunctionPassManagerImpl3runERNS_8FunctionE+0x86)[0xbb4146]
/home/edwin/llvm-svn/llvm/Release/bin/llc(_ZN4llvm19FunctionPassManager3runERNS_8FunctionE+0x3e)[0xbb42ce]
/home/edwin/llvm-svn/llvm/Release/bin/llc(main+0x1061)[0x4fdec1]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7f7b324441a6]
/home/edwin/llvm-svn/llvm/Release/bin/llc(_ZNSt8ios_base4InitD1Ev+0x39)[0x4fc3a9]
Aborted

The original .bc file was created by llvm-gcc, I can reduce a C testcase too if
needed.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 23 10:55:06 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Nov 2008 10:55:06 -0600
Subject: [LLVMbugs] [Bug 3117] New: ice: Result type 0 illegal!
Message-ID: <bug-3117-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3117

           Summary: ice: Result type 0 illegal!
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r59898 on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp62$ llvm-gcc -O3 small.c
Result type 0 illegal!
0x8e7baf4: i1 = setcc 0x8e7bf34, 0x8e7bfbc, 0x8e5c73c
small.c:54: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.

regehr at john-home:~/volatile/tmp62$ cat small.c

__extension__ typedef long long int int64_t;

int64_t safe_add_int64_t_s_s (int64_t si1, int si2)
{
  return si1 + si2;
}

int safe_sub_char_u_u (char ui1, int ui2)
{
  return ui1 - ui2;
}

int g_7;
char g_118;
int g_147;

int func_99 (char p_101);

int func_25 (void)
{
  int l_34 = 0xD2C298C2L;
  return l_34;
}

int func_73 (int p_74)
{
  int l_195;
  func_96 (1);
  func_76 (p_74, l_195, 1);
}

int func_76 (int p_78, int p_79, int p_81)
{
  func_86 (g_147, func_86 (g_7, 1), 1, 1);
}

int func_86 (int p_88, int p_90)
{
  func_96 (1);
}

int func_96 (int p_98)
{
  func_99 ((1, func_99 (1)));
}

int func_99 (char p_101)
{
  int l_117 = 0x628AL;
  g_118 =
    safe_sub_char_u_u (g_118 <=
                       safe_add_int64_t_s_s ((0, g_7) != p_101,
                                             func_25 ()) | 0, l_117);
}


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 23 14:28:45 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Nov 2008 14:28:45 -0600
Subject: [LLVMbugs] [Bug 3119] New: maybe a loop miscompilation
Message-ID: <bug-3119-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3119

           Summary: maybe a loop miscompilation
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


This bug is in 2.4 but not 2.3 and is still in r59914.  This is seen on Ubuntu
Hardy on x86.

regehr at john-home:~/volatile/tmp63$ llvm-gcc -O0 small.c -o small
regehr at john-home:~/volatile/tmp63$ ./small
-1
regehr at john-home:~/volatile/tmp63$ llvm-gcc -O1 small.c -o small
regehr at john-home:~/volatile/tmp63$ ./small
0
regehr at john-home:~/volatile/tmp63$ cat small.c
#include <stdint.h>
#include <stdio.h>

static inline uint64_t
safe_mod_uint64_t_u_u (uint64_t ui1, uint64_t ui2)
{
        if (ui2 == 0) return ui1;
        return ui1 % ui2;
}

static inline int8_t
safe_sub_int8_t_s_s (int8_t si1, int8_t si2)
{
        if (((si1^si2)
                 & (((si1 ^ ((si1^si2)
                 & (1 << (sizeof(int8_t)*8-1))))-si2)^si2)) < 0) {
                return si1;
        }
        return si1 - si2;
}

uint32_t g_61;
uint32_t g_123;
uint32_t g_147 = 1;

int32_t func_19 (void);
int32_t func_37 (uint8_t p_39);
int32_t func_64 (uint32_t p_65, uint32_t p_66);
int32_t func_90 (uint8_t p_92, int16_t p_93, uint8_t p_95);
int32_t func_105 (int32_t p_107, int32_t p_108);

int32_t
func_19 (void)
{
  func_37 (1);
  return g_147;
}

int32_t
func_37 (uint8_t p_39)
{
  int32_t l_243 = -1L;
  if (func_90 (1, (safe_mod_uint64_t_u_u (l_243, -1L)), 1))
    {
      func_64 (1, 1);
    }
  return 1;
}

int32_t
func_64 (uint32_t p_65, uint32_t p_66)
{
  for (g_147 = 0; 0; 1)
    return 1;
  return 1;
}

int32_t
func_90 (uint8_t p_92, int16_t p_93, uint8_t p_95)
{
  return p_93;
}

int32_t
func_105 (int32_t p_107, int32_t p_108)
{
  return 1;
}

int
main (void)
{
  printf ("%d\n", safe_sub_int8_t_s_s ((func_19 ()), 1));
  return 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 cs.uiuc.edu  Sun Nov 23 19:54:04 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Nov 2008 19:54:04 -0600
Subject: [LLVMbugs] [Bug 3120] New: Small interpreter-like testcase where
 LLVM fails to shine
Message-ID: <bug-3120-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3120

           Summary: Small interpreter-like testcase where LLVM fails to
                    shine
           Product: new-bugs
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: minor
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: danchr at gmail.com
                CC: llvmbugs at cs.uiuc.edu, danchr at gmail.com


Created an attachment (id=2238)
 --> (http://llvm.org/bugs/attachment.cgi?id=2238)
Testcase.tar

Attached below is a small program testing various techniques for writing the
core loop of an interpreter, and a shell script to execute it.

Example output from running on RHEL 5.1 on a 2.66GHz Core 2 Duo:

GCC 4.2.1 (LLVM): O4
Switched interpreter           602
Threaded interpreter           417
Recursive interpreter          610
Closure-based interpreter      868
GCC 4.2.1 (LLVM): O3
Switched interpreter           603
Threaded interpreter           416
Recursive interpreter          609
Closure-based interpreter      867
GCC 4.1.2: O3
Switched interpreter           335
Threaded interpreter           172
Recursive interpreter          444
GCC 3.4.6: O3
Switched interpreter           330
Threaded interpreter           411
Recursive interpreter          439
GCC 4.2.1 (LLVM): O2
Switched interpreter           604
Threaded interpreter           416
Recursive interpreter          609
Closure-based interpreter      869
GCC 4.1.2: O2
Switched interpreter           336
Threaded interpreter           169
Recursive interpreter          454
GCC 3.4.6: O2
Switched interpreter           327
Threaded interpreter           259
Recursive interpreter          446
GCC 4.2.1 (LLVM): Os
Switched interpreter           275
Threaded interpreter           365
Recursive interpreter          619
Closure-based interpreter      879
GCC 4.1.2: Os
Switched interpreter           343
Threaded interpreter           412
Recursive interpreter          475
GCC 3.4.6: Os
Switched interpreter           328
Threaded interpreter           259
Recursive interpreter          445

The timings for, for instance, Pentium 4 Xeon and PowerPC G4 are also quite
odd. Also, blocks are slow ;)


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 23 20:09:35 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Nov 2008 20:09:35 -0600
Subject: [LLVMbugs] [Bug 3121] New: __cxa_atexit multiply defined on Mac OS
	X 10.4
Message-ID: <bug-3121-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3121

           Summary: __cxa_atexit multiply defined on Mac OS X 10.4
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: scooter.phd at gmail.com
                CC: llvmbugs at cs.uiuc.edu


a) svn up to current trunk llvm
b) svn up to current trunk llvm-gcc-4.2
c) bootstrap compile llvm with platform gcc (4.0.1, for Tiger)
   [rm -rf <object directory>; configure; make]
d) bootstrap compile llvm-gcc-4.2 with llvm produced in previous step
   [rm -rf <object directory>; configure, make]
e) compile llvm with newly created llvm-gcc

The newly created llvm-gcc's runtime evidently has its own defintion of
__cxa_atexit. The generates a linker warning with no harmful side effects
(earliest you encounter this is when tblgen links.)

However, this does create problems running "make check" because the linker
warning creates many unexpected failures.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 23 21:30:33 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Nov 2008 21:30:33 -0600
Subject: [LLVMbugs] [Bug 3122] New: PPC vector operations failing in "make
 check" using llvm-gcc
Message-ID: <bug-3122-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3122

           Summary: PPC vector operations failing in "make check" using
                    llvm-gcc
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: scooter.phd at gmail.com
                CC: llvmbugs at cs.uiuc.edu


a) svn up to trunk llvm
b) bootstrap build by removing all object files, running configure and make
c) svn up to trunk llvm-gcc-4.2
d) remove all object files, run configure, and make with llvm libs built in (b)
e) Clean recompile llvm in a separate object directory using llvm-gcc built in
(d)

During "make check", the following are produced (looks like PPC codegen needs
help):

Running /Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/dg.exp ...
FAIL:
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/2006-04-19-vmaddfp-crash.ll
Failed with signal(SIGABRT) at line 1
while running:  llvm-as <
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/2006-04-19-vmaddfp-crash.ll
|  llc -march=ppc32 -mcpu=g5
/Users/scottm/play/llvm-trunk/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:4194:
failed assertion `Result.getValueType() == Op.getValueType() && "Bad
legalization!"'

FAIL:
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/2006-08-11-RetVector.ll
Failed with exit(1) at line 1
while running:  llvm-as <
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/2006-08-11-RetVector.ll | 
llc -march=ppc32 -mcpu=g5 | /usr/bin/grep vsldoi
/Users/scottm/play/llvm-trunk/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:4194:
failed assertion `Result.getValueType() == Op.getValueType() && "Bad
legalization!"'

FAIL:
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/2006-11-29-AltivecFPSplat.ll
Failed with signal(SIGABRT) at line 1
while running:  llvm-as <
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/2006-11-29-AltivecFPSplat.ll
|  llc -march=ppc32 -mcpu=g5
/Users/scottm/play/llvm-trunk/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:4194:
failed assertion `Result.getValueType() == Op.getValueType() && "Bad
legalization!"'

FAIL:
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/2007-03-30-SpillerCrash.ll
Failed with signal(SIGABRT) at line 1
while running:  llvm-as <
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/2007-03-30-SpillerCrash.ll |
 llc -march=ppc32 -mcpu=g5
/Users/scottm/play/llvm-trunk/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:4194:
failed assertion `Result.getValueType() == Op.getValueType() && "Bad
legalization!"'

FAIL:
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/2007-11-19-VectorSplitting.ll
for PR1811
Failed with signal(SIGABRT) at line 3
while running:  llvm-as <
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/2007-11-19-VectorSplitting.ll
|  llc -march=ppc32 -mcpu=g5
/Users/scottm/play/llvm-trunk/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:4194:
failed assertion `Result.getValueType() == Op.getValueType() && "Bad
legalization!"'

FAIL: /Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/vec_perf_shuffle.ll
Failed with signal(SIGABRT) at line 1
while running:  llvm-as <
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/vec_perf_shuffle.ll |  llc
-march=ppc32 -mcpu=g5 | not /usr/bin/grep vperm
/Users/scottm/play/llvm-trunk/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:4194:
failed assertion `Result.getValueType() == Op.getValueType() && "Bad
legalization!"'

FAIL: /Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/vec_shuffle.ll
Failed with signal(SIGABRT) at line 1
while running:  llvm-as <
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/vec_shuffle.ll |  opt
-instcombine |   llc -march=ppc32 -mcpu=g5 | not /usr/bin/grep vperm
/Users/scottm/play/llvm-trunk/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:4194:
failed assertion `Result.getValueType() == Op.getValueType() && "Bad
legalization!"'

FAIL: /Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/vec_splat.ll
Failed with signal(SIGABRT) at line 2
while running:  llvm-as <
/Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/vec_splat.ll |  llc
-march=ppc32 -mcpu=g5 -o vec_splat.ll.tmp -f -disable-legalize-types
/Users/scottm/play/llvm-trunk/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp:4194:
failed assertion `Result.getValueType() == Op.getValueType() && "Bad
legalization!"'

FAIL: /Users/scottm/play/llvm-trunk/test/CodeGen/PowerPC/vector.ll


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 23 22:23:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Nov 2008 22:23:25 -0600
Subject: [LLVMbugs] [Bug 3115] clang integer miscompilation
In-Reply-To: <bug-3115-206@http.llvm.org/bugs/>
Message-ID: <200811240423.mAO4NPwu005879@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3115


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Anders Carlsson <andersca at mac.com>  2008-11-23 22:23:24 ---
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20081117/009699.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 cs.uiuc.edu  Sun Nov 23 23:11:31 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Nov 2008 23:11:31 -0600
Subject: [LLVMbugs] [Bug 2413] void* address cast error
In-Reply-To: <bug-2413-206@http.llvm.org/bugs/>
Message-ID: <200811240511.mAO5BV07007459@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2413


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Anders Carlsson <andersca at mac.com>  2008-11-23 23:11:30 ---
Adding         test/CodeGen/PR2413-void-address-cast-error.c
Transmitting file data .
Committed revision 59941.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 23 23:20:16 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Nov 2008 23:20:16 -0600
Subject: [LLVMbugs] [Bug 2971] &((char*)0)[0] not regarded as static
	initializer
In-Reply-To: <bug-2971-206@http.llvm.org/bugs/>
Message-ID: <200811240520.mAO5KGVB007806@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2971


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #5 from Anders Carlsson <andersca at mac.com>  2008-11-23 23:20:16 ---
This is a duplicate of 2832

*** This bug has been marked as a duplicate of bug 2832 ***


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 23 23:24:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 23 Nov 2008 23:24:36 -0600
Subject: [LLVMbugs] [Bug 2832] ICE doesn't consider addrof as constant
In-Reply-To: <bug-2832-206@http.llvm.org/bugs/>
Message-ID: <200811240524.mAO5Oajv008050@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2832


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #3 from Anders Carlsson <andersca at mac.com>  2008-11-23 23:24:36 ---
Sending        lib/AST/Expr.cpp
Sending        test/CodeGen/staticinit.c
Transmitting file data ..
Committed revision 59946.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 24 00:24:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Nov 2008 00:24:01 -0600
Subject: [LLVMbugs] [Bug 3123] New: clang miscompiling comma operator
Message-ID: <bug-3123-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3123

           Summary: clang miscompiling comma operator
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


Seen using r59934 on Ubuntu Hardy on x86.

Obviously the right answer is 1.

regehr at john-home:~/volatile/tmp67$ clang small.c -I/usr/include --emit-llvm -o 
- | llvm-as | llc | as -o small.o ; gcc small.o -o small
regehr at john-home:~/volatile/tmp67$ ./small
0
regehr at john-home:~/volatile/tmp67$ cat small.c
#include <stdio.h>
#include <stdint.h>

uint32_t g_60;

int32_t func_111 (void);
int32_t func_111 (void)
{
  g_60 = 1;
  return 1;
}

void func_1 (void);
void func_1 (void)
{
  if (func_111 (), 0) { }
}

int main (void)
{
  func_1 ();
  printf ("%d\n", g_60);
  return 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 cs.uiuc.edu  Mon Nov 24 04:02:23 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Nov 2008 04:02:23 -0600
Subject: [LLVMbugs] [Bug 3033] libtasn1 1.5 test failure
In-Reply-To: <bug-3033-206@http.llvm.org/bugs/>
Message-ID: <200811241002.mAOA2NJ3026923@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3033


Seo Sanghyeon <sanxiyn at gmail.com> changed:

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




--- Comment #4 from Seo Sanghyeon <sanxiyn at gmail.com>  2008-11-24 04:02:20 ---
Verified.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 24 05:40:21 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Nov 2008 05:40:21 -0600
Subject: [LLVMbugs] [Bug 3124] New: VirtRegMap: Assertion `KillRegs[0] ==
 Dst || TRI-> isSubRegister(KillRegs[0], Dst)' failed
Message-ID: <bug-3124-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3124

           Summary: VirtRegMap: Assertion `KillRegs[0] == Dst || TRI-
                    >isSubRegister(KillRegs[0], Dst)' failed
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: edwintorok at gmail.com
                CC: llvmbugs at cs.uiuc.edu


With SVN r59914 on x86_64 linux:


$ ~/llvm-bootstrap/llvm-obj/Release/bin/llc bugpoint-reduced-simplified.bc
llc: /home/edwin/llvm-bootstrap/llvm/lib/CodeGen/VirtRegMap.cpp:1764:
void<unnamed>::LocalSpiller::RewriteMBB(llvm::MachineBasicBlock&,
llvm::VirtRegMap&): Assertion `KillRegs[0] == Dst ||
TRI->isSubRegister(KillRegs[0], Dst)' failed.
/home/edwin/llvm-bootstrap/llvm-obj/Release/bin/llc[0xc1693f]
/lib/libc.so.6[0x7f9471a8ef60]
/lib/libc.so.6(gsignal+0x35)[0x7f9471a8eed5]
/lib/libc.so.6(abort+0x183)[0x7f9471a903f3]
/lib/libc.so.6(__assert_fail+0xe9)[0x7f9471a87dc9]
/home/edwin/llvm-bootstrap/llvm-obj/Release/bin/llc[0xa97f8b]
/home/edwin/llvm-bootstrap/llvm-obj/Release/bin/llc[0xa98c49]
/home/edwin/llvm-bootstrap/llvm-obj/Release/bin/llc[0xa4bc09]
/home/edwin/llvm-bootstrap/llvm-obj/Release/bin/llc(_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE+0x15b)[0xbb252b]
/home/edwin/llvm-bootstrap/llvm-obj/Release/bin/llc(_ZN4llvm23FunctionPassManagerImpl3runERNS_8FunctionE+0x86)[0xbb2926]
/home/edwin/llvm-bootstrap/llvm-obj/Release/bin/llc(_ZN4llvm19FunctionPassManager3runERNS_8FunctionE+0x3e)[0xbb2aae]
/home/edwin/llvm-bootstrap/llvm-obj/Release/bin/llc(main+0x1061)[0x4fd6d1]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7f9471a7b1a6]
/home/edwin/llvm-bootstrap/llvm-obj/Release/bin/llc(_ZNSt8ios_base4InitD1Ev+0x39)[0x4fbbb9]
Aborted


Testcase:
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"
target triple = "x86_64-unknown-linux-gnu"
        %struct.cpuinfo_x86 = type { i8, i8, i8, i8, i32, i8, i8, i8, i32, i32,
[9 x i32], [16 x i8], [64 x i8], i32, i32, i32, i64, %struct.cpumask_t, i16,
i16, i16, i16, i16, i16, i16, i16, i32 }
        %struct.cpumask_t = type { [1 x i64] }
@.str10 = external constant [70 x i8]           ; <[70 x i8]*> [#uses=1]

declare i32 @printk(i8*, ...)

define void @display_cacheinfo(%struct.cpuinfo_x86* %c) nounwind section
".cpuinit.text" {
entry:
        %asmtmp = tail call { i32, i32, i32, i32 } asm "cpuid",
"={ax},={bx},={cx},={dx},0,2,~{dirflag},~{fpsr},~{flags}"(i32 -2147483643, i32
0) nounwind          ; <{ i32, i32, i32, i32 }> [#uses=0]
        %0 = tail call i32 (i8*, ...)* @printk(i8* getelementptr ([70 x i8]*
@.str10, i32 0, i64 0), i32 0, i32 0, i32 0, i32 0) nounwind           ; <i32>
[#uses=0]
        unreachable
}


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 24 12:50:42 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Nov 2008 12:50:42 -0600
Subject: [LLVMbugs] [Bug 3123] clang miscompiling comma operator
In-Reply-To: <bug-3123-206@http.llvm.org/bugs/>
Message-ID: <200811241850.mAOIogUu018918@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3123


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-24 12:50:42 ---
This is almost certainly the same problem as 3060 (clang constant folds the
expr to 0 but doesn't evaluate the side effects).  I hope to have this fixed in
the next couple weeks.

*** This bug has been marked as a duplicate of bug 3060 ***


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 24 13:26:34 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Nov 2008 13:26:34 -0600
Subject: [LLVMbugs] [Bug 3113] "this->replaceAllUsesWith(this) is NOT valid!"
In-Reply-To: <bug-3113-206@http.llvm.org/bugs/>
Message-ID: <200811241926.mAOJQYGJ020417@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3113


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-24 13:26:30 ---
Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081124/070304.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 cs.uiuc.edu  Mon Nov 24 13:52:27 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Nov 2008 13:52:27 -0600
Subject: [LLVMbugs] [Bug 2807] CellSPU cannot select fdiv
In-Reply-To: <bug-2807-206@http.llvm.org/bugs/>
Message-ID: <200811241952.mAOJqRDB021428@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2807


Brian Watt <bwatt at us.ibm.com> changed:

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




--- Comment #6 from Brian Watt <bwatt at us.ibm.com>  2008-11-24 13:52:26 ---
Retested current llvm with IR defined below, and still getting same error, that
is:

Cannot yet select: 0x9e24824: f64 = fdiv 0x9e24a44, 0x9e24b54


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 24 18:29:27 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Nov 2008 18:29:27 -0600
Subject: [LLVMbugs] [Bug 2807] CellSPU cannot select fdiv
In-Reply-To: <bug-2807-206@http.llvm.org/bugs/>
Message-ID: <200811250029.mAP0TRPg000365@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2807


Scott Michel <scooter.phd at gmail.com> changed:

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




--- Comment #7 from Scott Michel <scooter.phd at gmail.com>  2008-11-24 18:29:27 ---
Defer this bug until llvm-gcc-4.2 actually compiles. This is a libcall.


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 24 22:21:25 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Nov 2008 22:21:25 -0600
Subject: [LLVMbugs] [Bug 3128] New: assertion with bitfields
Message-ID: <bug-3128-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3128

           Summary: assertion with bitfields
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


This comes from a randomly generated test produced by the 'quest' compiler
testing system:

  void foo(int x);
  struct hgt0
  {
    short int hg0;
    short int hg1[2];
    signed hg2:9;
  };
  void callee_hg0f (struct hgt0 hgp6)
  {
    foo(hgp6.hg2);
  }

$ llvm-gcc tescase.c
cc1: ../../src/gcc/llvm-convert.cpp:2330: llvm::Value*
TreeToLLVM::EmitLoadOfLValue(tree_node*, const MemRef*): Assertion
`ValSizeInBits > LV.BitStart && "Bad bitfield lvalue!"' failed.
23878.c: In function ???callee_hg0f???:
23878.c:10: internal compiler error: Aborted
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.

Here's the backtrace:

#0  0xffffe425 in __kernel_vsyscall ()
#1  0xf7c9a640 in raise () from /lib/i686/cmov/libc.so.6
#2  0xf7c9c018 in abort () from /lib/i686/cmov/libc.so.6
#3  0xf7c935be in __assert_fail () from /lib/i686/cmov/libc.so.6
#4  0x084cb441 in TreeToLLVM::EmitLoadOfLValue (this=0xff86f498, 
    exp=0xf79ff02c, DestLoc=0x0) at ../../src/gcc/llvm-convert.cpp:2330
#5  0x084bb21d in TreeToLLVM::Emit (this=0xff86f498, exp=0xf79ff02c, 
    DestLoc=0x0) at ../../src/gcc/llvm-convert.cpp:973
#6  0x084cf2d8 in TreeToLLVM::EmitMODIFY_EXPR (this=0xff86f498, 
    exp=0xf7977190, DestLoc=0x0) at ../../src/gcc/llvm-convert.cpp:2944
#7  0x084bb0e4 in TreeToLLVM::Emit (this=0xff86f498, exp=0xf7977190, 
    DestLoc=0x0) at ../../src/gcc/llvm-convert.cpp:979
#8  0x084d9d45 in TreeToLLVM::EmitFunction (this=0xff86f498)
    at ../../src/gcc/llvm-convert.cpp:909
#9  0x084a0da5 in llvm_emit_code_for_current_function (fndecl=0xf79f5a00)
    at ../../src/gcc/llvm-backend.cpp:734
#10 0x0810dc24 in tree_rest_of_compilation ()
(gdb) up 4
#4  0x084cb441 in TreeToLLVM::EmitLoadOfLValue (this=0xff86f498, 
    exp=0xf79ff02c, DestLoc=0x0) at ../../src/gcc/llvm-convert.cpp:2330
2330        assert(ValSizeInBits > LV.BitStart && "Bad bitfield lvalue!");
(gdb) p ValSizeInBits
$1 = 32
(gdb) p LV
$2 = {Ptr = 0x8f18194, BitStart = 0 '\0', BitSize = 0 '\0'}
(gdb) up
#5  0x084bb21d in TreeToLLVM::Emit (this=0xff86f498, exp=0xf79ff02c, 
    DestLoc=0x0) at ../../src/gcc/llvm-convert.cpp:973
973         Result = EmitLoadOfLValue(exp, DestLoc);
(gdb) p exp
$4 = (tree) 0xf79ff02c
(gdb) p DestLoc
$5 = (const MemRef *) 0x0


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

From bugzilla-daemon at cs.uiuc.edu  Mon Nov 24 22:48:10 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Mon, 24 Nov 2008 22:48:10 -0600
Subject: [LLVMbugs] [Bug 3129] New: llvm-gcc infinite loop memory muncher
Message-ID: <bug-3129-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3129

           Summary: llvm-gcc infinite loop memory muncher
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


This illegal testcase causes llvm-gcc to chew up all the available memory:

  $ cat testcase.c
  struct sqt13 {};

  struct sqt13 
  {
    struct sqt13 sq64; 
    double sp11[2];
  } sp13 = {{46891.65, 45096},
            {26324.53, 89651.64}};
  $ ulimit -v $((512*1024))
  $ llvm-gcc testcase.c -emit-llvm -c -o x.bc
  x.c:4: error: redefinition of ???struct sqt13???

  cc1: out of memory allocating 68 bytes after a total of 463675392 bytes

It appears to happen in the frontend, before getting to any LLVM passes.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 25 10:28:01 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Nov 2008 10:28:01 -0600
Subject: [LLVMbugs] [Bug 3130] New: "int a = 2.0 ? 1 : 2;" crashes clang
Message-ID: <bug-3130-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3130

           Summary: "int a = 2.0 ? 1 : 2;" crashes clang
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: nunoplopes at sapo.pt
                CC: llvmbugs at cs.uiuc.edu


the following crashes 'clang -emit-llvm':
int a = 2.0 ? 1 : 2;

it crashes with -std=gnu99 and -std=c99, which are two different bugs (because
clang has different code for strict and GNU modes)


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 25 12:41:37 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Nov 2008 12:41:37 -0600
Subject: [LLVMbugs] [Bug 3131] New: assignment of a complex float crashes
	codegen
Message-ID: <bug-3131-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3131

           Summary: assignment of a complex float crashes codegen
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: nunoplopes at sapo.pt
                CC: llvmbugs at cs.uiuc.edu


the following code crashes 'clang -emit-llvm':

float _Complex bar();

void foo() {
        float _Complex x = bar();
}


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 25 16:23:02 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Nov 2008 16:23:02 -0600
Subject: [LLVMbugs] [Bug 3131] assignment of a complex float crashes codegen
In-Reply-To: <bug-3131-206@http.llvm.org/bugs/>
Message-ID: <200811252223.mAPMN2is026751@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3131


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Anders Carlsson <andersca at mac.com>  2008-11-25 16:23:01 ---
Sending        lib/CodeGen/CGCall.cpp
Sending        test/CodeGen/complex.c
Transmitting file data ..
Committed revision 60058.


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

From bugzilla-daemon at cs.uiuc.edu  Tue Nov 25 23:19:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Tue, 25 Nov 2008 23:19:00 -0600
Subject: [LLVMbugs] [Bug 3002] clang: crash on incomplete typedef
In-Reply-To: <bug-3002-206@http.llvm.org/bugs/>
Message-ID: <200811260519.mAQ5J0F8008238@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3002


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-25 23:18:59 ---
I now get:

t.c:1:8: error: expected identifier or '('
typedef
       ^
t.c:1:8: error: parse error
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 cs.uiuc.edu  Wed Nov 26 00:11:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Nov 2008 00:11:00 -0600
Subject: [LLVMbugs] [Bug 3132] New: llvm-gcc build broken on
	x86_64-redhat-linux
Message-ID: <bug-3132-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3132

           Summary: llvm-gcc build broken on x86_64-redhat-linux
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llvm-gcc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: lhames at gmail.com
                CC: llvmbugs at cs.uiuc.edu


During llvm-gcc build, compilation of insn-preds.c fails with the following
error:

../../src/gcc/config/i386/predicates.md: In function
'local_symbolic_operand_1':
../../src/gcc/config/i386/predicates.md:470: error: 'TARGET_FIX_AND_CONTINUE'
undeclared (first use in this function)
../../src/gcc/config/i386/predicates.md:470: error: (Each undeclared identifier
is reported only once
../../src/gcc/config/i386/predicates.md:470: error: for each function it
appears in.)
../../src/gcc/config/i386/predicates.md:471: warning: implicit declaration of
function 'machopic_data_defined_p'

target = x86_64-redhat-linux
build = x86_64-redhat-linux
host = x86_64-redhat-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 cs.uiuc.edu  Wed Nov 26 01:17:15 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Nov 2008 01:17:15 -0600
Subject: [LLVMbugs] [Bug 3132] llvm-gcc build broken on x86_64-redhat-linux
In-Reply-To: <bug-3132-206@http.llvm.org/bugs/>
Message-ID: <200811260717.mAQ7HFKx012641@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3132


Bill Wendling <isanbard at gmail.com> changed:

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




--- Comment #2 from Bill Wendling <isanbard at gmail.com>  2008-11-26 01:17:14 ---
"Fixed:"

http://llvm.org/viewvc/llvm-project?rev=60097&view=rev

If TARGET_FIX_AND_CONTINUE isn't #defined, then the behavior's now reverted
back to its original behavior. This may or may not be correct. But
-mfix-and-continue isn't something people should use anyway. . .


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 26 07:00:28 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Nov 2008 07:00:28 -0600
Subject: [LLVMbugs] [Bug 3134] New: debug info emitter crashes with 'char
	xpto[]; ' decl
Message-ID: <bug-3134-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3134

           Summary: debug info emitter crashes with 'char xpto[];' decl
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: LLVM Codegen
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: nunoplopes at sapo.pt
                CC: llvmbugs at cs.uiuc.edu


clang -emit-llvm -g crashes with the following code:
char xpto[];


clang: ASTContext.cpp:235: std::pair<long long unsigned int, unsigned int>
clang::ASTContext::getTypeInfo(const clang::Type*): Assertion `0 && "Incomplete
types have no size!"' 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 cs.uiuc.edu  Wed Nov 26 11:41:00 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Nov 2008 11:41:00 -0600
Subject: [LLVMbugs] [Bug 3134] debug info emitter crashes with 'char xpto[];
	' decl
In-Reply-To: <bug-3134-206@http.llvm.org/bugs/>
Message-ID: <200811261741.mAQHf04D012955@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3134


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Anders Carlsson <andersca at mac.com>  2008-11-26 11:40:58 ---
Sending        lib/CodeGen/CGDebugInfo.cpp
Sending        test/CodeGen/debug-info.c
Transmitting file data ..
Committed revision 60109.


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 26 13:56:43 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Nov 2008 13:56:43 -0600
Subject: [LLVMbugs] [Bug 3000] clang: used type 'union u' where arithmetic
 or pointer type is required
In-Reply-To: <bug-3000-206@http.llvm.org/bugs/>
Message-ID: <200811261956.mAQJuhTc017660@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3000


Nuno Lopes <nunoplopes at sapo.pt> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |nunoplopes at sapo.pt
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE




--- Comment #4 from Nuno Lopes <nunoplopes at sapo.pt>  2008-11-26 13:56:42 ---


*** This bug has been marked as a duplicate of bug 2746 ***


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 26 15:44:41 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Nov 2008 15:44:41 -0600
Subject: [LLVMbugs] [Bug 3136] New: clang crashes
Message-ID: <bug-3136-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3136

           Summary: clang crashes
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rdivacky at freebsd.org
                CC: llvmbugs at cs.uiuc.edu


clang crashes with this:

#0  0x08796206 in clang::ArrayType::getElementType (this=0x0) at Type.h:690
690       QualType getElementType() const { return ElementType; }
[New Thread 9001140 (LWP 100114)]
(gdb) bt
#0  0x08796206 in clang::ArrayType::getElementType (this=0x0) at Type.h:690
#1  0x0889d70b in clang::Sema::IsStringLiteralInit (this=0xbfbfe470,
Init=0x905e3c0, DeclType={ThePtr = 151073504}) at SemaDecl.cpp:670
#2  0x0889daa2 in clang::Sema::CheckInitializerTypes (this=0xbfbfe470,
Init=@0xbfbfd2b0, DeclType=@0xbfbfd2ac, InitLoc={ID = 2147510311}, InitEntity=
      {Ptr = 151623808}) at SemaDecl.cpp:695
#3  0x0889e0ab in clang::Sema::AddInitializerToDecl (this=0xbfbfe470,
dcl=0x90a1d30, init=0x905e3c0) at SemaDecl.cpp:1760
#4  0x0898343b in clang::Parser::ParseInitDeclaratorListAfterFirstDeclarator
(this=0xbfbfe648, D=@0xbfbfd420) at ParseDecl.cpp:295
#5  0x089838b3 in clang::Parser::ParseSimpleDeclaration (this=0xbfbfe648,
Context=5) at ParseDecl.cpp:236
#6  0x08983934 in clang::Parser::ParseDeclaration (this=0xbfbfe648, Context=5)
at ParseDecl.cpp:213
#7  0x0899892e in clang::Parser::ParseStatementOrDeclaration (this=0xbfbfe648,
OnlyStatement=false) at ParseStmt.cpp:101
#8  0x0899b100 in clang::Parser::ParseCompoundStatementBody (this=0xbfbfe648,
isStmtExpr=true) at ParseStmt.cpp:363
#9  0x0899b605 in clang::Parser::ParseCompoundStatement (this=0xbfbfe648,
isStmtExpr=true) at ParseStmt.cpp:341
#10 0x089890c2 in clang::Parser::ParseParenExpression (this=0xbfbfe648,
ExprType=@0xbfbfd930, CastTy=@0xbfbfd92c, RParenLoc=@0xbfbfd924)
    at ParseExpr.cpp:1046
#11 0x0898a832 in clang::Parser::ParseCastExpression (this=0xbfbfe648,
isUnaryExpression=false) at ParseExpr.cpp:453
#12 0x0898af29 in clang::Parser::ParseCastExpression (this=0xbfbfe648,
isUnaryExpression=false) at ParseExpr.cpp:557
#13 0x08988036 in clang::Parser::ParseExpression (this=0xbfbfe648) at
ParseExpr.cpp:176
#14 0x089893c7 in clang::Parser::ParseParenExpression (this=0xbfbfe648,
ExprType=@0xbfbfddb0, CastTy=@0xbfbfddac, RParenLoc=@0xbfbfdda4)
    at ParseExpr.cpp:1082
#15 0x0898a832 in clang::Parser::ParseCastExpression (this=0xbfbfe648,
isUnaryExpression=false) at ParseExpr.cpp:453
#16 0x08988036 in clang::Parser::ParseExpression (this=0xbfbfe648) at
ParseExpr.cpp:176
#17 0x08998480 in clang::Parser::ParseReturnStatement (this=0xbfbfe648) at
ParseStmt.cpp:1013
#18 0x08998ca5 in clang::Parser::ParseStatementOrDeclaration (this=0xbfbfe648,
OnlyStatement=false) at ParseStmt.cpp:159
#19 0x0899b100 in clang::Parser::ParseCompoundStatementBody (this=0xbfbfe648,
isStmtExpr=false) at ParseStmt.cpp:363
#20 0x0899b4ea in clang::Parser::ParseFunctionStatementBody (this=0xbfbfe648,
Decl=0x9006e80, L={ID = 1446241}, R={ID = 1446241}) at ParseStmt.cpp:1240
#21 0x089797be in clang::Parser::ParseFunctionDefinition (this=0xbfbfe648,
D=@0xbfbfe230) at Parser.cpp:527
#22 0x08979dc5 in clang::Parser::ParseDeclarationOrFunctionDefinition
(this=0xbfbfe648) at Parser.cpp:451
#23 0x0897a0fb in clang::Parser::ParseExternalDeclaration (this=0xbfbfe648) at
Parser.cpp:355
#24 0x0897a165 in clang::Parser::ParseTopLevelDecl (this=0xbfbfe648,
Result=@0xbfbfe700) at Parser.cpp:278
#25 0x08894df6 in clang::ParseAST (PP=@0x9001280, Consumer=0x9006340,
PrintStats=false, FreeMemory=true) at ParseAST.cpp:55
#26 0x085b7362 in ProcessInputFile (PP=@0x9001280, PPF=@0xbfbfea3c,
InFile=@0x90070a8, PA=EmitBC) at clang.cpp:1347
#27 0x085b91a2 in main (argc=16, argv=0xbfbfeb10) at clang.cpp:1545
(gdb) 

on the attached example


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 26 16:20:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Nov 2008 16:20:36 -0600
Subject: [LLVMbugs] [Bug 3136] clang crashes
In-Reply-To: <bug-3136-206@http.llvm.org/bugs/>
Message-ID: <200811262220.mAQMKaiZ022845@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3136


Roman Divacky <rdivacky at freebsd.org> changed:

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




--- Comment #2 from Roman Divacky <rdivacky at freebsd.org>  2008-11-26 16:20:36 ---
this is a problem on my side.. sorry for the noise


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

From bugzilla-daemon at cs.uiuc.edu  Wed Nov 26 17:10:20 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Wed, 26 Nov 2008 17:10:20 -0600
Subject: [LLVMbugs] [Bug 3137] New: strange decl rejected by clang but not
	by gcc
Message-ID: <bug-3137-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3137

           Summary: strange decl rejected by clang but not by gcc
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Semantic Analyzer
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: nunoplopes at sapo.pt
                CC: llvmbugs at cs.uiuc.edu


The code below doesn't seem much valid to me, but gcc accepts it.
if function var is removed, it works.

# gcc -fsyntax-only a.c && clang -fsyntax-only a.c
a.c:8:2: error: indirection requires pointer operand ('int' invalid)
        *__errno_location();
        ^~~~~~~~~~~~~~~~~~~



extern int *__errno_location(void);

void bar() {
        int *__errno_location();
}

void foo() {
        *__errno_location();
}


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

From bugzilla-daemon at cs.uiuc.edu  Thu Nov 27 08:38:59 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 27 Nov 2008 08:38:59 -0600
Subject: [LLVMbugs] [Bug 3138] New: jump threading horribly broken
Message-ID: <bug-3138-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3138

           Summary: jump threading horribly broken
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


Compiling this:

define i32 @jt() {
entry:
        br i1 true, label %bb3, label %bb

bb:             ; preds = %entry
        unreachable

bb3:            ; preds = %entry
        ret i32 0
}

with opt -jump-threading gives:

define i32 @jt() {
bb:
        unreachable

bb3:            ; No predecessors!
        ret i32 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 cs.uiuc.edu  Thu Nov 27 13:25:43 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Thu, 27 Nov 2008 13:25:43 -0600
Subject: [LLVMbugs] [Bug 3138] jump threading horribly broken
In-Reply-To: <bug-3138-206@http.llvm.org/bugs/>
Message-ID: <200811271925.mARJPhKL012771@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3138


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #1 from Chris Lattner <clattner at apple.com>  2008-11-27 13:25:43 ---
Fixed: 
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081124/070452.html

Thank you for the awesome 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 cs.uiuc.edu  Fri Nov 28 08:02:43 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Nov 2008 08:02:43 -0600
Subject: [LLVMbugs] [Bug 3140] New: .file directive is not emitted
Message-ID: <bug-3140-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3140

           Summary: .file directive is not emitted
           Product: tools
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: llc
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: rafael.espindola at gmail.com
                CC: llvmbugs at cs.uiuc.edu, nlewycky at google.com


Unlike gcc, llvm won't print .file directives. This is an easy way to get very
minimal debugging support at -g0.

If a file t.c is compiled with gcc the very first line in the assembly is

        .file   "t.c"

If it is compiled with llvm-gcc, the bitcode has

; ModuleID = 't.c'

but that is not propagated to the assembler. Is it safe to assume that we can
always print the ModelueID as a .file directive?


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

From bugzilla-daemon at cs.uiuc.edu  Fri Nov 28 14:20:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Nov 2008 14:20:47 -0600
Subject: [LLVMbugs] [Bug 3141] New: DeadStoreElimination crash
Message-ID: <bug-3141-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3141

           Summary: DeadStoreElimination crash
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2245)
 --> (http://llvm.org/bugs/attachment.cgi?id=2245)
Testcase .ll

$ opt -dse dse.bc -disable-output
opt[0x86114b4]
... (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 cs.uiuc.edu  Fri Nov 28 16:51:30 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Fri, 28 Nov 2008 16:51:30 -0600
Subject: [LLVMbugs] [Bug 3141] DeadStoreElimination crash
In-Reply-To: <bug-3141-206@http.llvm.org/bugs/>
Message-ID: <200811282251.mASMpUwI012121@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3141


Chris Lattner <clattner at apple.com> changed:

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




--- Comment #2 from Chris Lattner <clattner at apple.com>  2008-11-28 16:51:30 ---
Fixed, thank you for the nice testcase!
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20081124/070488.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 cs.uiuc.edu  Sat Nov 29 01:47:32 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Nov 2008 01:47:32 -0600
Subject: [LLVMbugs] [Bug 3142] New: Invalid instcombine of (-X)/-3
Message-ID: <bug-3142-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3142

           Summary: Invalid instcombine of (-X)/-3
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Windows NT
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: sharparrow1 at yahoo.com
                CC: llvmbugs at cs.uiuc.edu


Originally reported in bug 1931, but I apparently missed the following issue:

define i32 @a(i32 %X) nounwind readnone {
entry:
        %0 = sub i32 0, %X
        %1 = sdiv i32 %0, -3
        ret i32 %1
}

define i32 @b(i32 %X) nounwind readnone {
entry:
        %0 = call i32 @a(i32 -2147483648)
        ret i32 %0
}

define i32 @c(i32 %X) nounwind readnone {
entry:
        %0 = sub i32 0, -2147483648
        %1 = sdiv i32 %0, -3
        ret i32 %1
}

Run through opt -instcombine -inline, functions b and c return different
results.  This is the result of illegally combining (-X)/-3 to X/3: the
transformation is only valid when the negation doesn't overflow.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 29 10:55:39 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Nov 2008 10:55:39 -0600
Subject: [LLVMbugs] [Bug 3143] New: JIT doesn' t try to lookup function
 definiton in other Modules
Message-ID: <bug-3143-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3143

           Summary: JIT doesn't try to lookup function definiton in other
                    Modules
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Target-Independent JIT
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2246)
 --> (http://llvm.org/bugs/attachment.cgi?id=2246)
proposed fix to JIT

Consider an ExecutionEngine with two ModuleProviders. In one module, function
foo() is defined, in the other it's declared, and function bar() calls foo().

If you call JIT->getPointerToFunction on bar(), you'll get "Program use
external function 'foo' which could not be resolved!" because it's not even
trying to look in the other Modules for functions of the same 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 cs.uiuc.edu  Sat Nov 29 12:32:38 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Nov 2008 12:32:38 -0600
Subject: [LLVMbugs] [Bug 3143] JIT doesn't try to lookup function definiton
 in other Modules
In-Reply-To: <bug-3143-206@http.llvm.org/bugs/>
Message-ID: <200811291832.mATIWc7U024128@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3143


Nick Lewycky <nicholas at mxc.ca> changed:

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




--- Comment #1 from Nick Lewycky <nicholas at mxc.ca>  2008-11-29 12:32:37 ---


*** This bug has been marked as a duplicate of bug 2606 ***


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 29 20:12:51 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Nov 2008 20:12:51 -0600
Subject: [LLVMbugs] [Bug 2969] clang turns const arrays inside functions
 into allocas/ stores instead of "internal constant" arrays
In-Reply-To: <bug-2969-206@http.llvm.org/bugs/>
Message-ID: <200811300212.mAU2CpMJ005722@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2969


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #5 from Eli Friedman <sharparrow1 at yahoo.com>  2008-11-29 20:12:50 ---
As far as clang is concerned, the basic issue is fixed in r60270; if there are
other issues anyone wants to pursue, please file a new bug or email the
relevant mailing list.


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

From bugzilla-daemon at cs.uiuc.edu  Sat Nov 29 21:42:36 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sat, 29 Nov 2008 21:42:36 -0600
Subject: [LLVMbugs] [Bug 3142] Invalid instcombine of (-X)/-3
In-Reply-To: <bug-3142-206@http.llvm.org/bugs/>
Message-ID: <200811300342.mAU3gaIp009343@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3142


Bill Wendling <isanbard at gmail.com> changed:

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




--- Comment #3 from Bill Wendling <isanbard at gmail.com>  2008-11-29 21:42:35 ---
Fixed here:
http://llvm.org/viewvc/llvm-project?rev=60275&view=rev


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 30 00:02:57 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Nov 2008 00:02:57 -0600
Subject: [LLVMbugs] [Bug 2164] Codegen doesn't optimize 16-bit divide into
	multiply
In-Reply-To: <bug-2164-206@http.llvm.org/bugs/>
Message-ID: <200811300602.mAU62vl4013420@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2164


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #8 from Eli Friedman <sharparrow1 at yahoo.com>  2008-11-30 00:02:57 ---
Fix committed in r60283; sorry it took so long to take care 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 cs.uiuc.edu  Sun Nov 30 01:40:37 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Nov 2008 01:40:37 -0600
Subject: [LLVMbugs] [Bug 2330] Misc missed instcombines
In-Reply-To: <bug-2330-206@http.llvm.org/bugs/>
Message-ID: <200811300740.mAU7ebTf016018@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=2330


Eli Friedman <sharparrow1 at yahoo.com> changed:

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




--- Comment #17 from Eli Friedman <sharparrow1 at yahoo.com>  2008-11-30 01:40:37 ---
I moved the remaining issues into lib/Target/README.txt in r60285.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 30 03:47:49 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Nov 2008 03:47:49 -0600
Subject: [LLVMbugs] [Bug 3144] New: instcombine broken for large constant
	divisions
Message-ID: <bug-3144-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3144

           Summary: instcombine broken for large constant divisions
           Product: new-bugs
           Version: unspecified
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: baldrick at free.fr
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=2247)
 --> (http://llvm.org/bugs/attachment.cgi?id=2247)
testcase.ll

Bill's recent instcombine patches broke the Fortran build
because they use getSExtValue/getZExtValue.  These are only
valid for constants that fit in 64 bits.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 30 06:38:53 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Nov 2008 06:38:53 -0600
Subject: [LLVMbugs] [Bug 3144] instcombine broken for large constant
	divisions
In-Reply-To: <bug-3144-206@http.llvm.org/bugs/>
Message-ID: <200811301238.mAUCcri5026772@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=3144


Bill Wendling <isanbard at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |isanbard at gmail.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #1 from Bill Wendling <isanbard at gmail.com>  2008-11-30 06:38:52 ---
Fixed:

http://llvm.org/viewvc/llvm-project?rev=60288&view=rev

Very sorry for the breakage.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 30 11:19:42 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Nov 2008 11:19:42 -0600
Subject: [LLVMbugs] [Bug 3145] New: bogus "braces around scalar initializer"
	warning
Message-ID: <bug-3145-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3145

           Summary: bogus "braces around scalar initializer" warning
           Product: clang
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: minor
          Priority: P2
         Component: parser
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: nunoplopes at sapo.pt
                CC: llvmbugs at cs.uiuc.edu


The following code triggers the "braces around scalar initializer" warning,
which is bogus in this case.

struct a {
        int y;
};

struct b {
        int y;
        struct a x;
};

struct b z = {
        .x = {
                .y = 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 cs.uiuc.edu  Sun Nov 30 14:00:38 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Nov 2008 14:00:38 -0600
Subject: [LLVMbugs] [Bug 3146] New: intrinsics declarations,
 when passed through llc -march=cpp  , fail verification
Message-ID: <bug-3146-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3146

           Summary: intrinsics declarations, when passed through llc -
                    march=cpp, fail verification
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: flld0 at greynode.net
                CC: llvmbugs at cs.uiuc.edu


When an intrinsic like llvm.atomic.load.sub is declared in a module, and that
module is passed through the C++ backend and the resulting code run to
re-generate the module, the resulting declaration fails verification. In
particular, the assert at Verifier.cpp:1536 fails, as the function attributes
on the re-generated declaration aren't right.

The problem seems to be at CPPBackend.cpp:440, where the index for the function
attributes is truncated to 16 bits - the original data being an unsigned as
declared at Attributes.h:94.  The case in question being function attributes,
the index before serialization is ~0U, in serialized form is 65535, in
re-constituted form is 65535, and is then compared to attributes built
internally using ~0U again causing the failure. I think the declaration at
CPPBackend.cpp:440 should be an unsigned.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 30 15:26:52 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Nov 2008 15:26:52 -0600
Subject: [LLVMbugs] [Bug 3147] New: Can no longer compile clang with gcc-4.2
 on OpenBSD 4. 4 after r59835
Message-ID: <bug-3147-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3147

           Summary: Can no longer compile clang with gcc-4.2 on OpenBSD 4.4
                    after r59835
           Product: clang
           Version: unspecified
          Platform: All
               URL: http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-
                    Mon-20081117/009618.html
        OS/Version: OpenBSD
            Status: NEW
          Keywords: build-problem, regression
          Severity: blocker
          Priority: P2
         Component: AST
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: cnst+llvm at bugmail.mojo.ru
                CC: llvmbugs at cs.uiuc.edu, sebastian.redl at getdesigned.at


A week or two ago, the compilation of clang got broken on OpenBSD.  After I
backed out r59835 last week, I could again compile clang without a problem. 
(Today, after updating the tree, a clean backout of r59835 is no longer
possible due to conflicts.)

Below is the error messages that I get after ensuring that no part of r59835 is
backed out (e.g. the tree is totally as per svn, except for some small changes
in lib/{Lex/Lexer.cpp,Driver/InitHeaderSearch.cpp}):

phen:clang {4532} env CC=egcc CPP=ecpp CXX=eg++ gmake
gmake[1]: Entering directory `/home/constant/llvm/tools/clang/lib'
gmake[2]: Entering directory `/home/constant/llvm/tools/clang/lib/Headers'
gmake[2]: Nothing to be done for `all'.
gmake[2]: Leaving directory `/home/constant/llvm/tools/clang/lib/Headers'
gmake[2]: Entering directory `/home/constant/llvm/tools/clang/lib/Basic'
gmake[2]: Nothing to be done for `all'.
gmake[2]: Leaving directory `/home/constant/llvm/tools/clang/lib/Basic'
gmake[2]: Entering directory `/home/constant/llvm/tools/clang/lib/Lex'
gmake[2]: Nothing to be done for `all'.
gmake[2]: Leaving directory `/home/constant/llvm/tools/clang/lib/Lex'
gmake[2]: Entering directory `/home/constant/llvm/tools/clang/lib/Parse'
gmake[2]: Nothing to be done for `all'.
gmake[2]: Leaving directory `/home/constant/llvm/tools/clang/lib/Parse'
gmake[2]: Entering directory `/home/constant/llvm/tools/clang/lib/AST'
llvm[2]: Compiling StmtSerialization.cpp for Release build
/home/constant/llvm/include/llvm/Bitcode/Serialization.h: In static member
function 'static void llvm::SerializeTrait<T>::Emit(llvm::Serializer&, const
T&) [with T = short unsigned int:14]':
/home/constant/llvm/include/llvm/Bitcode/Serialize.h:42:   instantiated from
'void llvm::Serializer::Emit(const T&) [with T = short unsigned int:14]'
StmtSerialization.cpp:1448:   instantiated from here
/home/constant/llvm/include/llvm/Bitcode/Serialization.h:38: error: request for
member 'Emit' in 'X', which is of non-class type 'const short unsigned int:14'
/home/constant/llvm/include/llvm/Bitcode/Serialization.h: In static member
function 'static void llvm::SerializeTrait<T>::Emit(llvm::Serializer&, const
T&) [with T = short unsigned int:15]':
/home/constant/llvm/include/llvm/Bitcode/Serialize.h:42:   instantiated from
'void llvm::Serializer::Emit(const T&) [with T = short unsigned int:15]'
StmtSerialization.cpp:1449:   instantiated from here
/home/constant/llvm/include/llvm/Bitcode/Serialization.h:38: error: request for
member 'Emit' in 'X', which is of non-class type 'const short unsigned int:15'
gmake[2]: ***
[/home/constant/llvm/tools/clang/lib/AST/Release/StmtSerialization.o] Error 1
gmake[2]: Leaving directory `/home/constant/llvm/tools/clang/lib/AST'
gmake[1]: *** [AST/.makeall] Error 2
gmake[1]: Leaving directory `/home/constant/llvm/tools/clang/lib'
gmake: *** [all] Error 1
1.030u 0.200s 0:01.23 100.0%    0+0k 0+51io 0pf+0w
phen:clang {4533} uname -rms
OpenBSD 4.4 amd64
phen:clang {4534} egcc --version
egcc (GCC) 4.2.0 20070307 (prerelease)
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.

phen:clang {4535}


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 30 20:21:09 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Nov 2008 20:21:09 -0600
Subject: [LLVMbugs] [Bug 1955] tgmath.h is broken on Linux
In-Reply-To: <bug-1955-206@http.llvm.org/bugs/>
Message-ID: <200812010221.mB12L9XQ002542@zion.cs.uiuc.edu>

http://llvm.org/bugs/show_bug.cgi?id=1955


Anders Carlsson <andersca at mac.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andersca at mac.com
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED




--- Comment #5 from Anders Carlsson <andersca at mac.com>  2008-11-30 20:21:09 ---
With the changes I've made to integer constant expression handling, this no
longer generates an error, 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 cs.uiuc.edu  Sun Nov 30 22:28:15 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Nov 2008 22:28:15 -0600
Subject: [LLVMbugs] [Bug 3148] New: unnecessary GOT lookup when calling
 visibility hidden or protected function
Message-ID: <bug-3148-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3148

           Summary: unnecessary GOT lookup when calling visibility hidden or
                    protected function
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Common Code Generator Code
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nicholas at mxc.ca
                CC: llvmbugs at cs.uiuc.edu


GCC elides the GOT lookup when calling a hidden or protected function. Here's a
quick, silly example:

  class __attribute__((visibility("hidden"))) Foo {
  public:
    void f(void) __attribute__((noinline)) {
      for (int i = 0; i < 1000000; i += 2)
        if (i < 123) f();
      return;
    }
  };

  void g(void) {
    Foo foo;
    foo.f();
  }

Firstly, if f() is empty, GCC inlines it into g() despite the noinline tag. I
consider that naughty, and we shouldn't follow its lead there.

Secondly, GCC generates this code for g() with -O2 -fPIC:
_Z1gv:
        pushl   %ebp
        movl    %esp, %ebp
        subl    $24, %esp
        leal    -1(%ebp), %eax
        movl    %eax, (%esp)
        call    _ZN3Foo1fEv
        leave
        ret

while LLVM produces this:

_Z1gv:
        pushl   %ebp
        movl    %esp, %ebp
        pushl   %ebx
        subl    $12, %esp
        call    .Lllvm$1.$piclabel
.Lllvm$1.$piclabel:
        popl    %ebx
        addl    $_GLOBAL_OFFSET_TABLE_ + [.-.Lllvm$1.$piclabel], %ebx
        leal    -8(%ebp), %eax
        movl    %eax, (%esp)
        call    _ZN3Foo1fEv
        addl    $12, %esp
        popl    %ebx
        popl    %ebp
        ret

Because it's hidden, the GOT lookup is unnecessary. Please eliminate it. Same
with protected.


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

From bugzilla-daemon at cs.uiuc.edu  Sun Nov 30 22:50:47 2008
From: bugzilla-daemon at cs.uiuc.edu (bugzilla-daemon at cs.uiuc.edu)
Date: Sun, 30 Nov 2008 22:50:47 -0600
Subject: [LLVMbugs] [Bug 3149] New: possible integer miscompilation
Message-ID: <bug-3149-206@http.llvm.org/bugs/>

http://llvm.org/bugs/show_bug.cgi?id=3149

           Summary: possible integer miscompilation
           Product: new-bugs
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: regehr at cs.utah.edu
                CC: llvmbugs at cs.uiuc.edu


This one is interesting because llvm-gcc's output is wrong at all optimization
levels (-O0 through -O3).

This is a regression since 2.3.

Seen on r60325 on Ubuntu Hardy on x86.

regehr at john-home:~/volatile/tmp71$ llvm-gcc -O small.c -o small ; ./small
-1
regehr at john-home:~/volatile/tmp71$ llvm-gcc-2.4 -O small.c -o small ; ./small
-1
regehr at john-home:~/volatile/tmp71$ llvm-gcc-2.3 -O small.c -o small ; ./small
-10347
regehr at john-home:~/volatile/tmp71$ llvm-gcc-2.2 -O small.c -o small ; ./small
-10347
regehr at john-home:~/volatile/tmp71$ icc -O small.c -o small ; ./small
-10347
regehr at john-home:~/volatile/tmp71$ gcc -O small.c -o small ; ./small
-10347
regehr at john-home:~/volatile/tmp71$ cat small.c
#include <stdint.h>
#include <stdio.h>


static inline int16_t
safe_mul_int16_t_s_s (int16_t si1, int16_t si2) 
{                                               
        if (si1 > 0){  /* si1 is positive */    
                if (si2 > 0) {  /* si1 and si2 are positive */
                        if (si1 > (INT16_MAX / si2)) {          
                                return si1;                     
                        }                                       
                } /* end if si1 and si2 are positive */         
                else { /* si1 positive, si2 non-positive */     
                        if (si2 < (INT16_MIN / si1)) {          
                                return si1;                     
                        }                                       
                } /* si1 positive, si2 non-positive */          
        } /* end if si1 is positive */                          
        else { /* si1 is non-positive */                        
                if (si2 > 0) { /* si1 is non-positive, si2 is positive */       
                        if (si1 < (INT16_MIN / si2)) {                  
                                return si1;                             
                        }                                               
                } /* end if si1 is non-positive, si2 is positive */     
                else { /* si1 and si2 are non-positive */               
                        if ( (si1 != 0) && (si2 < (INT16_MAX / si1))) { 
                                return si1;                     
                        }                                               
                } /* end if si1 and si2 are non-positive */
        } /* end if si1 is non-positive */                      
        return si1 * si2;                                       
}                                                               

static inline uint64_t                                                  
safe_mod_uint64_t_u_u (uint64_t ui1, uint64_t ui2)                      
{                                                                       
        if (ui2 == 0) return ui1;                                       
        return ui1 % ui2;                                               
}                                                                       

int32_t g_355 = 0xC5E0286AL;
int32_t g_379 = -1L;
uint32_t g_450;

int32_t func_3 (uint32_t p_4);
int32_t func_3 (uint32_t p_4)
{
  g_450 = safe_mul_int16_t_s_s (p_4, 1);
  return 1;
}

int32_t func_1 (void);
int32_t func_1 (void)
{
  func_3 (safe_mod_uint64_t_u_u (g_379, g_355));
  return 1;
}

int main (void)
{
  func_1 ();
  printf ("%d\n", g_450);
  return 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.