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

Re: RFC/RFS: PyKaraoke

On Wed, 2006-02-15 at 22:16 +0100, Miriam Ruiz wrote:
>  --- Joe Wreschnig <piman@sacredchao.net> escribió:
> > Some issues:
> > 
> > Your debian/control should not depend directly on "python", but use
> > "${python:Depends}" and call dh_python in its binary-indep target. You
> > also need to Build-Depend on Python.
> I did that in one of my packages, which I co-maintain, that is written in
> python, and it was converted to something similar to what i did this time.
> Which is the best approach and why?

Depending on just "python" is fine if you're providing a Python program
with no modules (i.e. just an executable). However, if you have .py
modules that get imported, you need to make sure the modules are
compiled for the right version of Python (not doing this is a minor
bug), and also that the compiled modules are removed when the package is
removed (not doing this is a serious bug). dh_python will use a stronger
dependency to enforce the former, and include some code in the
postinst/prerm to automatically do the latter.

In the future, using dh_python may also do more magic to help Python
transitions and dependency issues. So it's good to use it to help
transitions remain as consistent and simple as possible. Exactly when in
the future, and what kind of magic should be done, is an active debate
on debian-python right now.

> > You patch the upstream source in a number of places. The reasons seem
> > good, but you also moved the cdgBorderPreset function which makes the
> > diff unnecessarily hard to check. Also,
> > "self.FileName[len(self.FileName)-1]" is much clearer as just
> > "self.FileName[-1]". If you haven't sent the changes upstream, you
> > probably should.
> To be honest, I reported some bugs after trying the package to upstream and it
> was him who sent me the patches.

Okay, then I have no problem with it.
Joe Wreschnig <piman@debian.org>

Attachment: signature.asc
Description: This is a digitally signed message part

Reply to: