Re: why are new upstream versions of glib being uploaded?

[Steve Langasek <vorlon@debian.org>]

On Tue, Dec 26, 2006 at 11:11:10PM -0700, Thomas Bushnell BSG wrote:
> On Tue, 2006-12-26 at 21:59 -0800, Steve Langasek wrote:
> > You asked for the release team's input into whether to upload the new
> > upstream version, and Luk replied with an explanation that was consistent
> > with the release team's position on the matter:  new upstream versions and
> > uploads including changes that don't qualify for freeze exceptions are
> > considered risky and therefore discouraged.  But you don't actually need the
> > release team's permission to upload to unstable, and it's not anything we're
> > going to get into an argument over with maintainers that make such uploads
> > as long as there aren't ABI-breaking changes that will interfere with
> > getting necessary updates into etch for other packages.

> And yet, glib in unstable now obviously exports a slightly different
> interface than the one in testing (the changelog notwithstanding).  So,
> if an RC bug crops up in the gnucash in testing, then I am screwed in a
> royal pain.  So, there was an ABI-breaking change, and it will interfere
> with any necessary update of gnucash in etch.

Just for clarification, no, this was a change in the behavior of the
function (API), not in its binary signature (ABI).  Rebuilding gnucash
against the glib in unstable rather than the glib in testing would have had
absolutely zero consequences for being able to get gnucash fixes into
testing.

-- 
Steve Langasek                   Give me a lever long enough and a Free OS
Debian Developer                   to set it on, and I can move the world.
vorlon@debian.org                                   http://www.debian.org/