Hi Gerard, > Ok, I've replicated your problem and am now debugging it. I'll let > you know what I find out.
Here is what I know so far: 1. The import of the md5 module is happening from a script called that is stored in coded form in EZ_SETUP_PY in virtualenv.py 2. The failure to import md5 from either hashlib or the deprecated module is because extra system paths in /opt/csw/lib/python aren't in sys.path 3. The sys.path list is manipulated from inside the virtualenv environment based on the site.py file in $venv/lib/python/site.py and this is stored in virtualenv.py as SITE_PY. If I modify site.py after the failed installation and add both /opt/csw/lib/python and /opt/csw/lib/python/lib-dynload to sys.path (at the end) and then run the extracted bootscript with -d ./myenv/lib/python2.6/site-packages the bootscript will succeed. Without the -d argument, it expects a directory at myenv/lib/python/site-packages with isn't created. So, it seems that this is a conflict between the expectations of virtualenv and CSW's python. I'm out of time to spend on this tonight but we can hopefully tweak virtualenv so that it will play nice with CSW's python in a way that we can kick the patch upstream. HTH. Thanks -Ben _______________________________________________ users mailing list [email protected] https://lists.opencsw.org/mailman/listinfo/users
