As the news show on the pymol homepage, version 1.1 is released. With this ebuild a couple of bugs are fixed. Reproducible: Always Steps to Reproduce:
Created attachment 158587 [details, diff] diff to latest in maintree
Created attachment 158589 [details] pymol-1.1-r1.ebuild
Created attachment 158591 [details] apbs-070604.patch Update to the latest version of apbs plugin
Created attachment 158593 [details] nosplash-gentoo.patch updated version fixes things mentioned in #210910
Created attachment 158595 [details] pymol-1.1-r1-data-path.patch only two files left for patching
Created attachment 158597 [details] pymol-1.1-r1-shaders.patch fixes #210910
Fixes bugs #210910 #213616 #214590 #215602 Don't know if this bug depends or block those.
Looks pretty good. Thanks for making the indentation normal! I made a couple changes related to the distutils stuff -- removed the postrm, added a call to distutils_pkg_postinst in your postinst, and added PYTHON_MODNAME="chempy pmg_tk pymol" to the top of the ebuild so distutils would handle that stuff right automatically. It's still building, so I'll let you know how it goes.
(In reply to comment #8) > PYTHON_MODNAME="chempy pmg_tk pymol" Thats a good feature! Didn't know that! There is a need for better docs of the eclasses.
Created attachment 158689 [details] pymol-1.1-r1.ebuild Looks like the ppc keyword was removed so that apbs could be pulled in? Instead, the apbs flag should be masked on that architecture. One other detail I noticed: there's a new dosym to /usr/share/data. That sounds like a bug in our data-path patch, so I removed it to aid in diagnosis. I couldn't actually open numerous PDBs, because my CPU pegged at 100% in /usr/lib64/python2.5/site-packages/pymol/_cmd.so:ObjectMoleculeConnect. I could open some small ones, but not a few larger ones (for example PDB ID 3GRS). Here's an updated ebuild, with the above changes remaining to be looked into.
(In reply to comment #10) > Looks like the ppc keyword was removed so that apbs could be pulled in? I can't do testing for ppc so I drop this always. > One other detail I noticed: > there's a new dosym to /usr/share/data. Still a relict from bug 215602. > I couldn't actually open numerous PDBs, because my CPU pegged at 100% in > /usr/lib64/python2.5/site-packages/pymol/_cmd.so:ObjectMoleculeConnect. I could > open some small ones, but not a few larger ones (for example PDB ID 3GRS). I do testing on python 2.5 on x86. Here everything is right (also for 3GRS). Upstream? > > Here's an updated ebuild, with the above changes remaining to be looked into. I did some more cosmetic changes.
Created attachment 158733 [details] diff to Donnies Version Cosmtic things which I learned from my studies in /usr/portage/eclass!
Looks mostly good, although that diff format is nearly unreadable. Please create diffs with 'diff -u' in the future. The only thing I saw is that the one you uploaded has DEPEND=$DEPEND (set equal to itself).
(In reply to comment #13) >Please create diffs with 'diff -u' in the future. Sorry, thats the way I normally do, but this time.. I don't know. > The only thing I saw is that the one you > uploaded has DEPEND=$DEPEND (set equal to itself). Think this was a typo.
My 100% CPU issue was some ricer CFLAGS, so ignore that. je_fro, do you want to commit this?
Yeah, I'll test and commit :) Thanks!
In science overlay for a week of testing, then it's off to the tree :) Thanks justin!
Every time again, je_fro!
Created attachment 160432 [details, diff] branch from upstream Hi je_fro, upstream did branch the code, which include two fixes. I updated the ESVN_URI. Perhaps you like two include it again. Thanks, justin
Please add this to the portage tree, the week of testing should be over now. Not seeing this in th tree, I went into trouble of building my own ebuild and update all related files, I have only now realized that it was wasted effort. Please add soon. This bug should not have been closed, since it never made it all the way into the official tree.
... yet another reason why a "third level" of testing sucks. (overlay, ~arch, stable)
mmkay, I'm sold... give me a couple days I'll commit it. slammed at work atm... Sorry for the hassle! je_fro
