You would probably be better off using the CentOS5-supplied python, tck,
tk, blt, and so forth (and if they aren't supplied, then consider a linux
distro that provides these things).
Also, if you compile your own ccp4, you can include the latest patches and
direct it to use whatever versions of these dependencies suit you, in
whatever location.
I may be mistaken, but I think phaser and cctbx are the only things that
require python. In my experience, python-2.3.x - 2.5.x all work, but if
you have something precompiled, you presumably will need the same major
version.
[log in to unmask] wrote:
> Hello,
> I have been running into issues with the tcl/tk and python bundles
> currently distributed with ccp4-6.0.2 (x86 system running Centos5).
>
> 1. The tcl/tk/blt package for Linux now extracts into a ./tcltkblt
> subdirectory. Apart from making upgrades slightly easier, this does not
> seem to make very much sense to me. The ccp4.setup file created by the
> install script still points to the standard location (/usr/local/bin) for
> the binaries, so I suspect the extract path was not intentionally changed.
>
> 2. On Centos5 (which comes with python 2.4.3), the python 2.4.2 package
> provided with ccp4 leads to a dilemma: As soon as /usr/local/bin/python is
> present, it is also used by standard desktop applications, which leads to
> undesired effects. For instance, gedit (the standard gnome text editor)
> complains about missing modules. If /usr/local/bin/python is made
> unavailable (by renaming, for example), sourcing ccp4.setup leads to an
> "incomplete libtbx environment" error.
>
> a) Is there a possibility to define the python binary used by ccp4
> (similar to the tcltk path in ccp4.setup)? This way you could extract the
> python package in a different (non-standard) location and avoid
> interference with other applications.
> b) Is the ccp4 python package customized in any way, or is it perfectly ok
> to use the version which comes with the OS?
>
> Any suggestions?
> Oliver
>
>
>
> -------------------------------
> Dr. Oliver H. Weiergraeber
> Institute of Neurosciences and Biophysics
> Molecular Biophysics
> Research Centre Juelich
> D-52425 Juelich
> Germany
> -------------------------------
>
|