Ok, in theory I have added this logic to the branch. Barry
> On Oct 29, 2017, at 7:17 PM, Jed Brown <j...@jedbrown.org> wrote: > > "Smith, Barry F." <bsm...@mcs.anl.gov> writes: > >> From below it looks like OpenMPI handles many compilers that >> don't handle type(*) by using some horrible non-portable >> pragmas? I can't say it excites me to want to add all the crap >> to PETSc configure/source. >> >> I'm inclined to have configure check if type(*) works and just >> shut down all our interfaces that require type(*) when it >> doesn't instead of pandering to out of date compilers. > > It's fine with me, particularly if this only affects PetscSF at present. > PGI is by far the most likely cause of requests from users and hopefully > they'll get their act together soon.