On Wed, Jul 2, 2025 at 6:42 PM Branko Čibej <br...@apache.org> wrote:
> On 3. 7. 25 01:05, Greg Stein wrote: > > On Tue, Jul 1, 2025 at 9:54 AM Nathan Hartman<hartman.nat...@gmail.com> > > wrote: > >... > >> IMHO supporting Python 2 can only get more difficult with time. > >> Eventually it may become hard to build or install Python 2 on modern > >> systems. The added friction will be a disincentive for devs to invest > >> time in it. So, yes, unless Someone (tm) steps forward to maintain > >> that, it makes sense to phase out support for Python 2. Leaving it for > >> 1.5 and moving forward on trunk seems reasonable to me. > >> > > Seems very appropriate to say "Py2 was EOL'd a decade ago. If you want > > that, then use serf 1.3 ... 1.5 *builds* require py3". > > > > Note that we're talking about packaging/building. Not necessarily the > > runtime environment. I could not explain why our build system requires > py2, > > for any modern packager. > > It doesn't require it. It supports it. Every single bit of python we > have now works equally well with 3.14.0-beta3 and 2.7. Currently, that's > not a burden. We have more trouble dealing with fairly recent versions > of SCons that have a broken CheckFunction() :( Well, not so recent, but > only recently fixed, TBH. > > I'm not advocating for clinging to staying compatible with Python2, just > being pedantic. Thanks for the clarification! ... "It happens to be true, but is not a goal" is what I'm hearing. Cheers, -g
