[llvm-commits] [llvm] r71986 - /llvm/trunk/test/FrontendC/2009-05-17-AlwaysInline.c

[Bill Wendling]

On Mon, May 18, 2009 at 7:05 AM, Duncan Sands <baldrick at free.fr> wrote:
> Hi Bill,
>
>> It's very strange, but on a self-hosted release build, I'm getting
>> this error:
>>
>> Running /Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvmCore/test/
>> FrontendC/dg.exp ... FAIL:
>> /Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvmCore/test/FrontendC/
>> 2009-05-17-AlwaysInline.c Failed with exit(1) at line 1 while running:
>> /Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvmgcc42.roots/
>> llvmgcc42~dst/Developer/usr/bin/llvm-gcc-4.2 -m32 -fstrict-aliasing -
>> Wstrict-aliasing -I/usr/include/c++/4.0.0/i686-apple-darwin9 -I/usr/
>> include/c++/4.0.0 -emit-llvm -w -S
>> /Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvmCore/test/FrontendC/
>> 2009-05-17-AlwaysInline.c -O0 -o - -mllvm --disable-llvm-optzns | /usr/
>> bin/grep bar cc1: Unknown command line argument '-fdisable-llvm-
>> optzns'. Try: 'cc1 --help'
>> I'm not sure where the '-fdisable-llvm-optzns' command is coming from...
>
> this is similar to what you get when you are compiling with the system
> gcc and not llvm-gcc, and you pass it -mllvm --disable-llvm-optzns.  So
> is it possible that it's not llvm-gcc being used here, but standard gcc?
>
I doubt it. The log file shows that it's using "llvm-gcc". This must
be a release-mode deal, as it's failing in release mode for me (and
not just self-hosting). Here's an option. Use "-disable-llvm-optzns"
instead of "--disable-llvm-optzns". It seems that '--' is converted to
'-f' in gcc. I have no clue why:

$ /Developer/usr/bin/llvm-gcc -c t.c -o /dev/null -mllvm --relocation-model=pic
cc1: for the -relocation-model=pic option: : may only occur zero or one times!

$ /Developer/usr/bin/llvm-gcc -c t.c -o /dev/null -mllvm --relocation-model=pic
cc1: Unknown command line argument '-frelocation-model=pic'.  Try: 'cc1 --help'

-bw