|
I can speak to Joerg's issue this, although all of my own work has be validated on py39.
The "best practice" for py27 -> py3k migration since the EoL of py27 had been py27 -> py36. Since we (and upstream) dropped py36, the next best option is py27 -> py37. My guess is by dropping py37, Joerg is going to have to update the migration to support validation
of py27 -> py38 first.
(FWIW, I have code that currently has locked into a dependency on tornado 6.1 because of asycnio changes in recent python driving deprecations of methods I depend on in older versions of tornado, so I understand Joerg's pain).
From: pkgsrc-users-owner%NetBSD.org@localhost on behalf of Greg Troxel Sent: Sunday, October 30, 2022 7:51 PM To: Joerg Sonnenberger Cc: pkgsrc-users%netbsd.org@localhost Subject: [External] Re: python 3.7 removal proposal (before upstream EOL!!) Joerg Sonnenberger <joerg%bec.de@localhost> writes: > Am Sun, Oct 30, 2022 at 06:40:21PM -0400 schrieb Greg Troxel: >> >> Joerg Sonnenberger <joerg%bec.de@localhost> writes: >> >> > Am Sun, Oct 30, 2022 at 08:11:13AM -0400 schrieb Greg Troxel: >> >> Does anyone actually use python37 from pkgsrc (who is running >> >> pkgsrc-current, or who is running 2022Q3 and who regularly updates to >> >> new quarterly releases)*? Please speak up and explain why. (Or send >> >> me text to post anonymously if you don't want to admit it public.) >> > >> > Yes, I do. It is one of the upgrade paths for complex Python 2.7 >> > programs... >> >> You mean that you port things from 2.7 to 3.7, and then from 3.7 to >> 3.11? >> >> To do that, do you need >> >> python3.7 in wip >> >> python3.7 in pkgsrc, but no python-depending packages for 37 >> >> python3.7 in pkgsrc, with 37 packages not built unless added? >> >> also some python-depending packages >> >> or really: if we removed it anyway, how much extra work would that >> impose on you? (to be compared to the work on the part of others to >> deal with the breakage we have now) > > Yes, I would have to fork a lot of packages further increasing the pain. Or you could just do your 2 porting steps with 2022Q3, and then move to HEAD. > So far I haven't seen that many cases of difficult to fix breakage... The point is that the breakage is showing up and not getting fixed, so it is de facto too hard, relative to the effort that is being applied. We need a get-well plan from that. We'll see if anybody else speaks up, or if it's only you that wants to port (something) from 2.7->3.7->3.11. The week is young :-) |