Changes in directory llvm/lib/AsmParser:
llvmAsmParser.y updated: 1.274 - 1.275
---
Log message:
Initialize some pointers to quiet the compiler (when doing build_gcc
builds).
---
Diffs of the changes: (+2 -2)
llvmAsmParser.y |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
Second Update:
I ran findmisopt one more time and this time it found something:
-mem2reg -instcombine. I'm currently tracking down the misoptimization.
It seems to be a little too aggressive in eliminating setcc:
136,139c129,131
%tmp71 = load ubyte* %tmp70 ; ubyte [#uses=1]
Changes in directory llvm-test/MultiSource/Applications/hbd:
d6-arith.cpp updated: 1.2 - 1.3
decomp.cpp updated: 1.2 - 1.3
---
Log message:
Fix array delete instances. This shuts up about 1300 errors from valgrind.
---
Diffs of the changes: (+5 -5)
d6-arith.cpp |2 +-
decomp.cpp |
Changes in directory llvm/lib/Target:
README.txt updated: 1.47 - 1.48
---
Log message:
Cute example from Chris Lattner.
---
Diffs of the changes: (+40 -0)
README.txt | 40
1 files changed, 40 insertions(+)
Index: llvm/lib/Target/README.txt
diff
Attached is the llvm-gcc4 patch to correspond with the new CAST
instructions in LLVM.
For Review Only. PLEASE DO NOT COMMIT THIS!
Reid.
Index: gcc/llvm-convert.cpp
===
--- gcc/llvm-convert.cpp (revision 204)
+++
On Nov 12, 2006, at 6:48 PM, Reid Spencer wrote:Attached is the llvm-gcc4 patch to correspond with the new CASTinstructions in LLVM.For Review Only. PLEASE DO NOT COMMIT THIS!This patch looks fine: the uses of getCast are safe, because the CFE never has types that disagree with what the FE thinks