On Fri, 3 Jul 2026, Kirill A. Korinsky wrote:
> I never understand why not just leave python2 in ports alone?
>
> It builds fast, some software still works which is writteon in this and I
> doubt that it will be rewritten to python3 anyway soon.
>
Stuart shared that this software was indeed rewritten to Python 3 if
that's the way folks want to go:
https://github.com/jaredmauch/mailman2-python3
On your broader question.
As there is no development on Py 2.7, things like ssl or compiler updates
continue to be risks to break python2 itself.
Many of the ports below that use Python 2 are pretty much not evolving,
most things have moved to Python 3 so updating away from Python 2 was
required in order to update the port itself (e.g. games/vegastrike which I
updated recently) so getting newer updates from software that is still
active required moving to Python 3 anyway.
What's left is mostly dead software.
I believe a few years ago we looked at Tauthon as a continuation of
Python 2, but that hasn't been the direction we've taken as you can see
in ports history.
In any case, here's everything that's left. And these are all solvable
except for cvs2svn so I don't see any good reason to not finish the
cleanup. What's your use case for keeping Python 2?
devel/cvs2svn (no solution yet; but on my radar)
mail/archivemail (remove or port to py3; dead upstream;
last release in 2011)
textproc/svndumptool (remove or port to py3; dead upstream;
last release in 2009)
mail/mailman (current discussion)
net/py2-dnspython (solved by mailman)
www/moinmoin (I am waiting for any ok's to import new deps;
sent to ports@)
graphics/gimp/stable (move to 3.x series, has been ready for years)
devel/py-gobject (solved by gimp)
graphics/py2-cairo (solved by gimp)
x11/py-gtk2 (solved by gimp)
devel/py2-pip (last cleanup along with py2)
devel/py2-setuptools (last cleanup along with py2)
(And some tests in security/john-jumbo, probably this port could be
updated to HEAD which converted things to Python 3 from what I understand)
