[MOAB-dev] r3036 - in MOAB/trunk: . tools/iMesh/python

[Jed Brown]

Jason Kraftcheck wrote:
> Jed Brown wrote:
>> I tried with both to no avail:
>>
>>  --with-mpi=/usr FC=/usr/bin/mpif90 F77=/usr/bin/mpif77
>>
>
> I still can't reproduce this.  I did:
>   ../src/configure --with-mpi=/usr/local \
>                    FC=/usr/local/bin/mpif90 \
>                    F77=/usr/local/bin/mpif77
> And got:
>   Fortran: /usr/local/bin/mpif90 -g -O2 -O2 -DNDEBUG
>
> Further, MOAB shouldn't even be trying to detect a Fortran compiler because
> you specified "--enable-imesh=nofortran".

Agreed.

> Which verison of MOAB are you using?  I made some changes to the Fortran
> detection stuff in early June.  Do you have those changes?

This was with r3036 (today).  With your latest (r3041) the Fortran
compiler is not printed (which makes sense because it's not being used)
but the autoconf output clearly shows that it is searching for a Fortran
compiler and finding gfortran (after looking for a bunch of others).  I
do get the correct compilers when I configure without the nofortran
argument, so maybe this is harmless?  But what if a user doesn't have
any Fortran compiler, will autoconf fail if it doesn't find one?

Jed

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 260 bytes
Desc: OpenPGP digital signature
URL: <http://lists.mcs.anl.gov/pipermail/moab-dev/attachments/20090721/9f83e432/attachment.pgp>