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

Re: Summary of python transition problems

On Tue, 2003-09-30 at 06:50, Matthias Klose wrote:
> Colin Watson writes:
> > On Mon, Sep 29, 2003 at 11:29:12PM +1000, Donovan Baarda wrote:
> > > what about python2.3 having "Depends: python (>= 2.3)"
> > > 
> > > This is IMHO a policy violation. python (2.3) should depend on python2.3,
> > > python2.3 should not depend on python.
> > > 
> > > It's playing havoc with my mixed testing/unstable systems, as some
> > > unstable packages want python2.3 (>=2.3), which then wants to pull in
> > > every single python package from unstable because it forces the python
> > > (2.3) transition which I really don't want to do until it is ready. It
> > > doesn't have to be this way.
> > 
> > While this would probably help users, it won't make the transition
> > easier as far as testing is concerned, because python comes from the
> > python2.3 source package and python2.3 is only at 2.2.106 in testing.
> > Perhaps it could be left until after the testing transition is complete
> > so that we have something more or less releaseable first?
> yes, exactly my plan: remove the dependency after the migration to
> testing.

I'm not exactly sure how packages transition to testing; is it "by
binary package", or "by source package"?

If it is "by binary package", then with correct dependencies python2.3
could transition into testing before python (2.3). If it is "by source
package", then wouldn't a python2.3 with fixed dependencies be held back
anyway because the python2.3 source also generates python (2.3) which
can't transition until all the python-<foo> packages are OK?

In either case, I don't see how having python2.3 with busted
dependencies in unstable helps. I'd prefer to have it fixed _before_ it
transitions to testing.

But then I'm not the one doing it, so take my comments with a grain of
salt :-)

Donovan Baarda <abo@minkirri.apana.org.au>

Reply to: