Re: [Libmesh-users] MPI compile error in debug...

[Derek G. <fri...@gm...>]

Good points Roy... this gets my vote.

Derek

On 6/5/07, Roy Stogner <roy...@ic...> wrote:
> On Tue, 5 Jun 2007, John Peterson wrote:
>
> > Roy Stogner writes:
> > > On Tue, 5 Jun 2007, Steffen Petersen wrote:
> > >
> > > > I do not know much about that issue, but in order to
> > > > get rid of the error message I simply add -Wno-long-long to the
> > > > CXXFLAGS
> > >
> > > Sounds like a good workaround to me.  Should we add it to aclocal.m4
> > > (or Make.common.in, or whereever such flags are most appropriately
> > > set)?
> >
> >
> > I think it's only for the MPICH implementation of MPI, and only for
> > the GCC compilers.  I know we test for the latter, but I'm not sure
> > we've ever tested for which implementation of MPI you are using.
> >
> > maybe #ifdef MPICH_NAME or MPICH_VERSION?
>
> We definitely don't want to add random flags to compilers which
> might not support them, but the flags set in aclocal.m4 are defined
> for each compiler version, so we could make sure -Wno-long-long only
> got added to gcc versions that support it.
>
> I don't think we even need to check the MPI version.  If we turn off
> warnings for "long long" unnecessarily, what's the harm?  Nobody's
> relying on the compiler warnings to prevent them from putting "long
> long" variables into libMesh code, I hope; and if someone wants to use
> it in their application or depends on a library that needs to support
> it, we shouldn't make that unnecessarily difficult.
> ---
> Roy
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Libmesh-users mailing list
> Lib...@li...
> https://lists.sourceforge.net/lists/listinfo/libmesh-users
>