APBS is a poisson-boltzmann solver for calculating electrostatic surfaces. i've updated the ebuild for it to include the most recent version available. this program is used within pymol so it might be reasonable to add it as a dependency for pymol. the old ebuild (apbs-0.4.0) was marked ~x86 so perhaps this one should be as well. i haven't done extensive testing but i see nothing that this breaks on the system, and since it's an independent program unrelated to normal OS operation i can't see how it would break something critical. Reproducible: Always Steps to Reproduce: 1.new ebuild submission - no reproduction.. ebuild here: http://extrospective.net/apbs-0.5.0.ebuild digest here: http://extrospective.net/digest-apbs-0.5.0
Created attachment 110881 [details] the updated ebuild for apbs 0.5.0 this is the new ebuild file
Created attachment 110883 [details] this is the digest this is the new digest for abps-0.5.0
I've been working on this...it'll be finished pretty soon.
Cool, thanks to both of you! Cheers, Markus
is there any compelling reason to not use the attached ebuild? it installs apbs to the proper directory, the tools, examples, and documentation. if there's something else i've missed then it is appropriate to use a different ebuild but this one seems to work. i am admittedly a novice to apbs though so it is perhaps best left in your hands jeff. thanks much -snf
(In reply to comment #5) > is there any compelling reason to not use the attached ebuild? it installs > apbs to the proper directory, the tools, examples, and documentation. if > there's something else i've missed then it is appropriate to use a different > ebuild but this one seems to work. i am admittedly a novice to apbs though so > it is perhaps best left in your hands jeff. > > thanks much > -snf > Hi Stephen, First off, thanks for the good work, it is very much appreciated! We devs always strive to check packages thoroughly before committing new ebuilds in order to make sure they work for (ideally) all users. And this simply takes time. E.g., your ebuild omits the mpi checks for maloc; does this mean that it works now as intended or will users compiling in mpi support have similar problems than they did for v0.4. Also, you omitted the patch for the examples; does this mean that they now work out of the box? These and many more things we have to check out, ideally on all of x86, amd64 and ppc. Hence, please bear with us for a little and keep up the good work. cheers, Markus
sounds great thanks guys. -snf
Commited...Thanks again Stephen!
cool :) my only worry is that the things that markus brought up haven't been addressed, for example i was building apbs without maloc on this computer. i guess we can always wait to see other bugs reported relating to apbs and solve them as they arise but more testing on this may be in order. if i have some time this weekend i'll see what i can do. -snf
maloc is included with this release, and the examples scripts are ready to use.
(In reply to comment #10) > maloc is included with this release, and the examples scripts are ready to use. > Thanks a lot for bumping apbs, Jeffrey! One cosmetic thing I noticed in the ebuild is that you can actually remove src_unpack() since it only implements the default behavior anyway. Have a good weekend, Markus
