Nicolas Mailhot <nicolas.mailhot <at> laposte.net> writes:
> Just shows compat packages need an explicit EOL date or we'll carry
> them forever (some upstreams won't even look at the new version while
> the compat package is available)
On the other hand, removing compat packages which are still needed by important
packages just because some arbitrary EOL date set much in advance has been
reached is not a good idea either. Sure, I don't like compat packages either
and I'm much in favor of porting apps to the new library versions whenever
possible, even in Fedora-specific patches. But on the other hand, some compat
packages really _need_ to be there for a very long time ("forever"), such as
gtk+ (v1) (hopefully that will be obsolete soon, but with imlib (v1) depending
on it for some strange reason and parts of KDE in turn depending on imlib, I
don't see that happening right now) or kdelibs3 (KDE 4 has major changes from
KDE 3, probably even more than KDE 2 and GTK+/GNOME 2 had from their respective
v1, and definitely a lot more than from KDE 2 to 3). It really all depends on:
* how many changes the new version has (i.e. how easy it is to port
applications (or other libraries) written for the old version),
* whether there is someone willing to maintain the compat library,
* how many applications use the old library, and how important they are (also
to some extent whether someone is maintaining the old applications in Fedora),
* how many of these have been ported at the time the compat library is supposed
to be dropped (which also depends on the previous factors to some extent),
* how easy it is to get the old and new versions of the library to coexist
(something which was the determining factor against a compat-python),
so there is no easy answer there.
Kevin Kofler
--
fedora-devel-list mailing list
fedora-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-devel-list
