I have a very odd issue. I have two builds of Coot, one for OS X Intel and
one for OS X PPC. On each of those platforms, I have a standalone
probe/reduce in the shell PATH. When I run either of these Coots, they
manage to find and try to use the phenix.probe installed in our linux
branch.
From the shell output:
(command-in-path? *probe-command*)
/programs/i386-linux/phenix/1.4-125/phenix-1.4-125/build/intel-linux-2.6/bin/phenix.probe
^Z
Suspended
[bene@cotterpin][/nfs/home/bene] echo $PATH | tr : '\n' | grep phenix
/programs/powermac/phenix/1.4-3/phenix-1.4-3/build/mac-ppc-osx/bin
[bene@-cotterpin][/nfs/home/bene] which probe
probe: aliased to /programs/powermac/probe/2.12.071128/probe
[bene@-cotterpin][/nfs/home/bene] unalias probe
[bene@-cotterpin][/nfs/home/bene] which probe
/programs/share/bin/probe
[bene@-cotterpin][/nfs/home/bene] /programs/share/bin/probe
ERROR: no input files
Syntax: probe input.pdb >> out.kin
or: probe [flags] "src pattern" ["target pattern"] pdbfiles... >>
out.kin
<snip>
The /programs/share/bin/probe is just a wrapper to find the correct binary
depending on your platform, and I don't think it's the culprit here.
Is Coot doing some magic when it tries to figure out where probe/reduce are
installed? Grepping through the source looking for phenix didn't turn
anything obvious up.
I will do the expedient thing and hardwire the path in group-settings.scm,
but the wrapper scripts have been working for a long time now.
Any ideas on what's going awry?
-ben
--
| Ben Eisenbraun | Software Maintainer |
| Structural Biology Grid | http://sbgrid.org |
| Harvard Medical School | http://hms.harvard.edu |
|