Python plans for Jessie?
Hi!
We're all happy for the release of Wheezy, but now it's time to think
about Jessie, and to consider the plans for Python ecosystem.
Here are some goals/points of discussion I'd like to tackle
immediately, to have enough time to squeeze them in time for Jessie.
Feel free to add more points :)
python2.6 removal
---------------------------
Unfortunately, python2.6 has been released with Wheezy as supported
version. I think we all agree it should be removed as soon as possible
to avoid maintaining a Python major version which is no longer
supported upstream. I think this could be done quite soon from now
(Release Team should have the final say, as they will probably face a
storm of new transitions in the next few days). I volunteer to drive
the removal of python2.6 from the archive.
python-central removal
---------------------------------
python-central has been deprecated in favour of dh_python2, bugs have
been filed already [0] suggesting to migrate to dh_python2, and only a
few packages are still (build-)depending on it. I think it's for our
own sake we should get rid of python-central in time for Jessie, and
we should focus on provide patches/uploads to speed up the migration.
An UDD script is in place [1] to track the packages still to be
migrated. I plan to provide some NMUs quite early for Jessie to tackle
this goal, but I'd appreciate a coordinated effort to proceed faster.
Plans for python-support
-----------------------------------
Although python-support has been deprecated by its maintainers, it's
still used by a large portion of Python packages in Debian [1], and
some maintainers still prefer it to dh_python2. I don't think there
are chances to get rid of python-support in time for Jessie, but what
we could do now is to discuss about plans to converge towards
dh_python2, smoothing out the rough edges that are found along the
way.
Multi-arch support
--------------------------
Both python and python3 (uploaded to experimental) provide Multi-Arch
support. These changes are quite invasive, and can lead to several
problems (e.g. FTBFS, as seen in Ubuntu recently). A broader testing
of the features introduced should be considered.
[0]http://bugs.debian.org/cgi-bin/pkgreport.cgi?tag=pycentral-deprecation;users=debian-python@lists.debian.org
[1] http://udd.debian.org/cgi-bin/python_helpers.cgi
Cheers,
Luca
Reply to: