Re: [OMPI devel] pfile_read_at_f.c:99: error: invalid lvalue in unary '&'

[Jeff Squyres]

Sorry for the delay in replying -- this has been a crazy week in trying to catch up from SC06 and the US holiday...

Ok, I think I've got all the fixes in now; I simulated a full build in your environment and I think I fixed up 2 more typos (including the one you noted). Attached is a patch to apply on top of the one you already applied. The fixes are in the OMPI development trunk and will be moved over to the release branches shortly. They will be included in all future OMPI releases. Let me know if this works for you; we have not tested the sizeof (int) != sizeof(INTEGER) code well. Out of curiosity: is there a reason you're using g95 instead of gfortran?

f77-2.patch
Description: Binary data

On Nov 28, 2006, at 8:14 AM, Anglade Pierre-Matthieu wrote:

Hi,

Now it's ptestsome_f.c which gives the following error:

ptestsome_f.c: In function 'mpi_testsome_f':
ptestsome_f.c:98: error: invalid lvalue in decrement
ptestsome_f.c:105: error: invalid lvalue in increment
make[4]: *** [ptestsome_f.lo] Error 1



I'll be grateful if you could again send me a patch for this one.

Best regards,

PMA


These are typos in the code path for where the size of a C integer is
not the same as the size for a Fortran INTEGER.  I have fixed the
problems on our development head; they will be included in the next
release of Open MPI.

For the moment, the attached patch should fix the problem for you.





On Nov 28, 2006, at 4:23 AM, Anglade Pierre-Matthieu wrote:

> Hi all,
>
> Trying to compile open-mpi 1.1.2 I've encounter the error message
> used as a title. Please find attached a tar.file cointaining:
> -the configure output and log
> -the make output
>
> I've try to correct the error and indeed changing line 98-99 in
> pfile_read_at_f.c
> from
> 98   if (MPI_SUCCESS == OMPI_FINT_2_INT(*ierr) &&&
> 99       MPI_STATUS_IGNORE != c_status) {
> to
> 98   if (MPI_SUCCESS == OMPI_FINT_2_INT(*ierr) &&
> 99       MPI_STATUS_IGNORE != c_status) {
> allow continuing the compilation up to the next file containing
> precisely the same 'structure'. So I'm wondering what went wrong
> with my build. Can someone help me with that ?
>
> Thanks in advance.
>
> --
> Pierre-Matthieu Anglade
> <build-problem.tar.bz2>
> _______________________________________________
> devel mailing list
> de...@open-mpi.org
> http://www.open-mpi.org/mailman/listinfo.cgi/devel


--
Jeff Squyres
Server Virtualization Business Unit
Cisco Systems




_______________________________________________
devel mailing list
de...@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/devel





--
Pierre-Matthieu Anglade
_______________________________________________
devel mailing list
de...@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/devel

The patch works. No suprprise. Thanks very much. Yet it seems that my unusual machine setup keeps making troubles. gcc -DHAVE_CONFIG_H -I. -I../../../../../ompi/mpi/f77/profile - I../../../../opal/include -I../../../../orte/include -I../../../../ ompi/include -I../../../../ompi/include -DOMPI_PROFILE_LAYER=1 - DOMPI_COMPILING_F77_WRAPPERS=1 -I../../../../.. -I../../../.. - I../../../../../opal/include -I../../../../../orte/include - I../../../../../ompi/include -O3 -DNDEBUG -fno-strict-aliasing - pthread -MT ptestsome_f.lo -MD -MP -MF .deps/ptestsome_f.Tpo -c ptestsome_f.c -fPIC -DPIC -o .libs/ptestsome_f.o make[4]: Leaving directory `/home/anglade/mpi/openmpi- 1.1.2/build2/ ompi/mpi/f77/profile' I guess that you don't need again the different logs (but I can obviously send them if required). On 11/28/06, Jeff Squyres <jsquy...@cisco.com> wrote: Many thanks for the bug report!

--
Jeff Squyres
Server Virtualization Business Unit
Cisco Systems