Apbs bumps from 0.5.1 to 1.0.0. A simple bump in version number of the ebuild compiles and works fine on my x86 box with gcc and gfortran. But an implementation for openmpi was introduced, enabled with --enable-openmpi while configuring, which doesn't work out of the box. Here some affored should be placed.
Hi Justin, Thanks much for the note and I am almost done with the new ebuild. I would like to still enable some of the tests that ship with the tarball; we'll see if that pans out. best, Markus
I committed apbs-1.0.0 to portage cvs. Thanks, Markus
Thank you Markus. It is working fine. Is there a reason why ifc isn't supported as compiler? On my box it is working fine. There is another thing with the pymol plugin. In the mpi mode apbs writes out its potentials with "-PE0" as an extension. This causes the plugin to fail, because it doesn't expect that. I try to report upstream, but sf is down. Perhaps we could patch it ourself. I'm not familiar with python so I cannot contribute.
Hi Justin, Regarding ifc/ifort: I don't use them so I don't typically check if packages work with them or not. If I get a chance to install the compilers on one of my dev boxes sometime I'll have a look. Thanks for pointing this out. Regarding the pymol problem. Could you email me a short testcase that would allow me to reproduce and check into this? Thanks, Markus