On Fri, December 12, 2008 10:12 am, Nathaniel Echols wrote: > On Thu, Dec 11, 2008 at 3:05 AM, William Scott > <wgsc...@chemistry.ucsc.edu>wrote: > >> I need a couple of people to volunteer to test this on 10.4 and 10.5 >> intel >> machines: >> >> http://sage.ucsc.edu/xtal/coot/coot-0.5-pre1-intel_10.4_10.5.tgz >> > > This build appears to have python embedded more or less correctly, so it > now > uses the built-in interpreter when launched with "--script test.py": > > coot-0.5: > sys.executable = /Users/nat/phenix/build/bin/python > > coot-0.5-pre1: > sys.executable = /usr/local/xtal/coot/bin/coot-real > > I wish I knew what the difference in the builds was. . . I'm going to try > the fink version next. > > Unfortunately, there appear to be problems with the actual Python modules. > I'm not entirely sure what specifically is broken, but I have to manually > add /usr/local/xtal/coot/share/coot/python to sys.path, and the module > formerly known as coot_python does not appear to exist any more (but is > still referenced in several places). So it doesn't look like it's > possible > to manipulate the GUI from Python right now, although this is clearly a > step > in the right direction. [I can send more info to Paul/Bernhard off-list > if > this isn't a known problem.] > > thanks, > Nat >
I built it naively, and using fink, according to the following: http://sage.ucsc.edu/~wgscott/xtal/wiki/index.php/Building_a_stand-alone_coot_on_OS_X >From what I have seen, python within coot no longer is aware of what is in the site-packages directory. I guess sys.path is getting clobbered somehow, but I don't understand how python is implemented in coot, so I don't have anything useful to suggest. If someone smarter than me can suggest a fix, I will immediately implement it. Thanks. Bill William G. Scott Contact info: http://chemistry.ucsc.edu/~wgscott/