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

Re: Considerations for 'xmms' removal from Debian



Neil Williams wrote:
> On Sun, 08 Jul 2007 11:48:39 +0100
> Roger Leigh <rleigh@whinlatter.ukfsn.org> wrote:
> 
>> Kapil Hari Paranjape <kapil@imsc.res.in> writes:
[...]
> Such a minimal port is hardly worth doing. It is possible to migrate
> from glib1 to glib2 in such a way (see #359299) but it is much harder
> to go from gtk1 to gtk2. I've been involved in three gtk1->gtk2 ports,
> one v.large (GnuCash), one v.small with a dead upstream (quicklist) and
> one where a "minimal port" (the last act of the old upstream) combined
> with an ill-advised RCS branch has left a horrible mess of spaghetti
> code. I'm not sure if the third will ever be a sane Gtk2 application.
[...]
> $ apt-cache rdepends libgtk1.2 | grep -c -v "^lib"
> 316

Note that about 30 of these are xmms related. Also a lot of the remaining are
bindings or depending libraries. There are also some packages that are being
ported anyway.

> I'm not sure Debian needs to throw out over 300 applications before
> Lenny. True, most of those are dead upstream - AFAICT GnuCash was the
> last active upstream to make it to gtk2 - but although these packages
> use old libraries that have an undoubted *potential* for security
> problems, in the absence of actual bug reports is it really worth
> dropping so many packages?

It's worth trying to convince as many maintainers as possible to think about
migrating to newer libraries or asking for removal of their package if it's
not really usable anymore.

> Is a dead upstream sufficient cause to drop a package from Debian in
> the absence of any RC bugs? Is a dependency on libgtk1.2 going to *be*
> an RC bug for Lenny? It seems a very big step, IMHO.

A dead upstream is not sufficient cause to drop a package perse. Though it
should be sufficient to think about dropping or migrating the package...

Cheers

Luk



Reply to: