[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: gquilt: RC bug fix



On Sun, Feb 18, 2007 at 03:47:23PM -0800, Steve Langasek wrote:
> On Sat, Feb 17, 2007 at 10:26:23AM -0500, A. Christine Spang wrote:
> 
> > I've just recently had my sponsor upload gquilt 0.17-3 to unstable. I'm
> > requesting that it be granted a freeze exception for fixing RC bug
> > #411198 (http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=411198).
> 
> - Why is the depends: on python (>= 2.4), (<= 2.5) hardcoded now, instead of
>   being autopopulated from ${python:Depends} by way of XS-Python-Version:
>   current?

How exactly does XS-Python-Version: current work? It doesn't seem to add
the right dependencies under Depends: when I try building using
${python:Depends}. I get a Depends: line that looks like this:

 Depends: python-central (>= 0.5.8), python-gtk2 (>= 2.4), quilt

Which is why I added the hard dependency in the first place.

The only other field in the generated binary control file that I see that's
relevant is Python-Version: current, but I don't know if or how that has
an effect on dependencies.

> - The bounds of the range are wrong -- python (<= 2.5) is satisfied by a
>   version number of exactly 2.5, but that would be incompatible at the
>   bytecode level...

Huh, that's right. Note that this is stated incorrectly in the
python policy manual:
http://www.debian.org/doc/packaging-manuals/python-policy/ch-programs.html#s-current_version_progs

> - It would be nice if there were support for automatically recompiling
>   bytecode for private modules when python is upgraded; but I guess that
>   part of the proposed python policy didn't get spec'ed out in time?  Well,
>   I guess it can be done with an rtupdate script, but that it's by no means
>   required under current policy, ahwell.

While I agree that this would be nice, I'm told that there is no helper
for this yet, and I don't think it's really worth it to implement
something like that for this package in particular before etch.

Regards,
Christine



Reply to: