[gentoo-dev] net-zope maintenance

[Dirkjan Ochtman]

Hi there,

I just completed removing the python herd from the metadata of about
70 packages in the net-zope category (all of these packages are still
nominally maintained by the net-zope herd). This happened due to the
fact that the majority of the python team isn't interested in
maintaining Zope-related packages (there's 200+ of them in the
net-zope category), and a different herd already exists that should in
theory be able to take care of them.

The net-zope herd nominally has two members: radek and tupone. From
grepping changelogs, radek hasn't committed anything in net-zope since
2007. tupone committed one change in net-zope in 2010 and 4 in 2009.
Instead, most of the recent work has been done by arfrever, who has
recently been retired.

Given this situation, I'm not sure it makes sense for us to keep all
of those packages in the tree. Except for dependencies on
zope-interface, which is pretty accepted and will keep being
maintained by the python team in addition to the net-zope team, not
much is depending on net-zope packages, either:

app-admin/zprod-manager depends on net-zope/zope (hasn't been
committed to since 2008)

dev-python/rdflib has an optional dependency on net-zope/zodb, in
addition to many other storage backends

media-libs/FusionSound for whatever reason blocks net-zope/zodb

(There are two packages in dev-python that depend on zope-testing, but
those are actually only dependencies of net-zope packages again.)

Zope maintenance is a lot of work and delays other stuff indefinitely:

https://bugs.gentoo.org/show_bug.cgi?id=148333 (blockers for Python 2.5)
https://bugs.gentoo.org/show_bug.cgi?id=335248 (new Plone requires
80-100 new ebuilds)

And there are 3 open security bugs from 2011 with vulnerabilities in
Zope and/or Plone.

So unless someone steps up to takes a serious whack at all of this
stuff, perhaps we should just cut it from the portage tree.

Cheers,

Dirkjan