On Sun, Nov 20, 2005 at 01:01:16AM -0600, Joe Wreschnig wrote: > On Sat, 2005-11-19 at 22:20 -0800, Steve Langasek wrote: > > Yes, this failure is due to a bug in python2.3, but seriously -- this isn't > > Java, and zip files aren't cool. I really can't see any good reason for > > obfuscating the python files in this manner within the Debian package.
> So read http://lists.debian.org/debian-python/2003/08/msg00133.html > where I explained what's wrong with the current Python policy. Aha, hmm. I had noticed that lack of byte-compiling was a side-effect of keeping these files in a zip, but didn't realize it was the primary *goal* of the zip file. Though it seems to me this isn't so much a bug with the python policy, as it is a bug with the python package implementation; the policy is simply written in such a way that it ensures packages don't interact badly and leave behind .pyc and .pyo files with no owner. Hmm, I guess dh_python cleans those up anyway though... so maybe we do need to look at whether bytecompiling buys us anything... (Incidentally, your zip files are only a partial solution for this, as you may realize -- I see that I do have /usr/lib/quodlibet/{exfalso,quodlibet}.py[co] files on my own system, apparently put there by the package postinst...) I agree with you that it's desirable to not have to rebuild python packages and twiddle their dependencies for every python update; shoving stuff into zip files just doesn't seem to be all that great of an alternative. :/ > I'm preparing an upload to close this now. Thanks. :) -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/
signature.asc
Description: Digital signature
