[llvm-commits] [llvm-gcc-4.2] r76781 [5/5] - in /llvm-gcc-4.2/trunk: ./ gcc/ gcc/config/ gcc/config/arm/ gcc/config/rs6000/ gcc/cp/ gcc/doc/ gcc/testsuite/g++.apple/ gcc/testsuite/g++.dg/abi/ gcc/testsuite/gcc.apple/ gcc/testsuite/gcc.target/arm/ gcc/testsuite/gcc.target/arm/neon/ gcc/testsuite/obj-c++.dg/ gcc/testsuite/objc.dg/
Bill Wendling
isanbard at gmail.com
Thu Jul 23 02:19:42 PDT 2009
On Jul 22, 2009, at 9:00 PM, Duncan Sands wrote:
> Hi Bob,
>
>> Added: llvm-gcc-4.2/trunk/gcc/testsuite/gcc.target/arm/neon/
>> vreinterprets8_s64.c
>> URL: http://llvm.org/viewvc/llvm-project/llvm-gcc-4.2/trunk/gcc/testsuite/gcc.target/arm/neon/vreinterprets8_s64.c?rev=76781&view=auto
>
> where's the commit message explaining what you're doing?
>
When the patch is huge, SVN splits it up into different email
messages. Kind of a pain, but...:-/
>> --- llvm-gcc-4.2/trunk/gcc/varasm.c (original)
>> +++ llvm-gcc-4.2/trunk/gcc/varasm.c Wed Jul 22 15:36:27 2009
>> @@ -4758,6 +4758,7 @@
>> && GET_CODE (XEXP (DECL_RTL (decl), 0)) == SYMBOL_REF)
>> SYMBOL_REF_WEAK (XEXP (DECL_RTL (decl), 0)) = 1;
>> #endif
>> + /* LLVM LOCAL end */
>
> Not sure why this change is lumped in with a bunch of new testcases.
> I've got to admit that it's harmless though :)
>
Some code didn't have good LLVM LOCAL markers on it. When I saw them,
I just added them.
>> -#define VERSUFFIX " (Based on Apple Inc. build 5646) (LLVM build "
>> LLVM_VERSION_INFO ")"
>> +#define VERSUFFIX " (Based on Apple Inc. build 5647) (LLVM build "
>> LLVM_VERSION_INFO ")"
>> #else
>> -#define VERSUFFIX " (Based on Apple Inc. build 5646) (LLVM build)"
>> +#define VERSUFFIX " (Based on Apple Inc. build 5647) (LLVM build)"
>> #endif
>> #else
>> -#define VERSUFFIX " (Based on Apple Inc. build 5646)"
>> +#define VERSUFFIX " (Based on Apple Inc. build 5647)"
>
> Likewise.
>
This is actually an important change (at least for Apple). We need to
keep the version of Apple's GCC that we're synced up with specified
here for our use.
This patch is similar to one of my "random merge" patches where I suck
over everything in Apple's GCC TOT into our tree. It just so happens
that the ARM stuff came along at this time, and it was massive.
-bw
More information about the llvm-commits
mailing list