Tobias Nygren <tnn%NetBSD.org@localhost> writes: > On Fri, 3 Apr 2015 12:43:30 -0400 > rodent%NetBSD.org@localhost wrote: > >> > Removal proposals should explain what what's officially EOL and actually >> > unmaintained and for how long, will be lost, and what the positive >> > effects on maintenance issues will be. (I have not noticed any real >> > pain from python26 still being in pkgsrc.) >> >> python26 was EOL'd in 2013Q4. $X py26 packages will be lost. Maintenance will >> be easier due to having fewer packages to maintain and not backporting fixes to >> packages which might never be fixed upstream. Fewer breaks in bulk builds due >> to upstream dropping support for py26 (sometimes, this is not documented). >> etc. etc. The real issue is how much pain it will cause people, and whether we can reasonably tell anyone people using py26-anything that the time is long past when that is a reasonble thing to do. Probably we're there, but I should have expressed that gauging the pain imposed on people who are not being unreasonable is part of the point. > Technical merits and drawbacks aside, do we expect we have any users > that still set PYTHON_VERSION_DEFAULT=26 and what would their reason for > doing so be? I don't know how big a difference 2.6 vs 2.7 is but if the > answer to this question is "no", then there's no point to discuss > technicalities further. Freeze aside, nobody objected on this thread > for two weeks. Personally I think that is plenty of heads-up. I would expect that no one needs to use 26. The usual issue is not "I feel like using 26" but "package X works with 26 but not 27". There's generally a lot of that early on, but probably very little, maybe none of that, in packages currently in pkgsrc. So if no one speaks up in a week that explaining why we should keep 26, I won't object to removal.
Attachment:
pgpfSV4vuZHhu.pgp
Description: PGP signature