[petsc-dev] defaulting gnu compilers on Apple

[Aron Ahmadia]

~ ❯❯❯ man gcc
No manual entry for gcc


On Thu, Oct 31, 2013 at 2:01 PM, Satish Balay <balay at mcs.anl.gov> wrote:

> On Thu, 31 Oct 2013, Barry Smith wrote:
>
> > > As of now - the differences this wrapper might provide is not obvious
> to us.
> > >
> > > So for practical purposes '/usr/bin/gcc' is same as 'clang’.
> >
> >    How can you say that. We simply do not know.
>
> 1. /usr/bin/gcc says its clang [in verbose mode]
>
> Executing: /opt/HPC/mpich-3.0.4-gcc4.2/bin/mpicc --version
> stdout:
> Apple LLVM version 5.0 (clang-500.2.79) (based on LLVM 3.3svn)
> Target: x86_64-apple-darwin13.0.0
>
> 2. /usr/bin/gcc says its clang in error messages:
>
> > ~/s/s/w/tmp ❯❯❯ gcc -fsel-sched-pipelining file.c
> >
> > clang: warning: argument unused during compilation:
> '-fsel-sched-pipelining'
>
> 3. And it accepts clang arguments that gcc does not.
>
> > ~/s/s/w/tmp ❯❯❯ gcc --analyze file.c
> > ~/s/s/w/tmp ❯❯❯ clang --analyze file.c
>
> So - I conclude '/usr/bin/gcc' is clang - perhaps with a light-weight
> wrapper [because the binaries don't match.] until someone can show
> there is a difference in behavior in terms of 'works with clang - but
> not with /usr/bin/gcc'
>
> Satish
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20131031/53fba7c6/attachment.html>