Tobias Nygren <tnn%NetBSD.org@localhost> writes: > On Wed, 13 Apr 2022 10:14:15 -0400 > Peter Lai <cowbert%gmail.com@localhost> wrote: > >> Disagree. 3.7 does not EOL until JUN 2023. I recommend we follow >> upstream support schedule. > > This is a valid point. But if upstream module authors of well-known > modules start to deprecate support ahead of the core upstream support > schedule then that is a problem for us. We have to at least acknowledge > that keeping it comes with the cost of intermittent tree breakage. Agreed - what I was trying to say, but more concise :-) > I favor keeping at most three major versions. The fact that we are even having a discussion about dropping the oldest of four versions because many things are desupporting it, when that version has a year to go until EOL is a clue about the python world. It means that even though the base package is not yet EOL, a substantial amount of the python community is treating it as such. Right now, dropping 3.7 to leave .8 (old), .9 (default), .10 (too new to be default) sounds right. I wouldn't want to adopt a "drop 3.8 as we add 3.11" rule, because 3.11 arriving and upstreams of python packages starting to de-support 3.8 aren't really causally related.
Description: PGP signature