fixed 301898 1.2.7-6 thanks Hi Michael,
I believe that the fix to bug 195509 made gfortran work as the F77 and F90 compiler, so this should work now. If not, can you please reopen? [And please don't include [EMAIL PROTECTED] in your reply -- unless you reopen the bug.] Thanks, Adam On Thu, 2005-03-31 at 13:29 -0500, Adam C Powell IV wrote: > Hello and thank you for the report, > > On Tue, 2005-03-29 at 19:49 -0500, Michael Gilbert wrote: > > Hello Dr. Powell, > > > > I believe that the required modifications are quite minor. I basically > > copied mpif90 (this turns out to be a shell script) from the mpich source > > directory. Then, i changed the "includedir" and "libdir" lines to properly > > point to the mpich include and lib dirs. With that slight change, mpich > > works flawlessly with g95 as the Fortran back end. > > > > I understand that the stable/testing package shouldn't include the Fortran > > script, but if the user builds the package from source with a Fortran 90 > > compiler, I do think that the script should be put in the correct location. > > I mean, if the user is building the package from source with a compiler > > not included in the distribution, they likely understand that they are > > moving into the unstable world. Could the build script only put mpif90 in > > the correct location if the F90 environment variable is set at build time? > > > > Lastly, I found that this should really be applicable to the > > libmpich1.0-dev package, not libmpich1.0. Thanks for your consideration. > > Would you be willing to put together a patch to the Debian packaging > (and mpich sources if necessary) which implements this while still > functioning as-is for stable/testing? > > -Adam -- GPG fingerprint: D54D 1AEE B11C CE9B A02B C5DD 526F 01E8 564E E4B6 Engineering consulting with open source tools http://www.opennovation.com/
signature.asc
Description: This is a digitally signed message part
