Re #2 I think the only thing is someone to make it happen. AFAIK nothing prevents it, and we should almost certainly do it.

@curved-television-6568 @broad-processor-92400 Am I out to lunch in suggesting there was a slack thread or GH ticket where one or both of you had a prototype of this? It’s somewhere in my head, but my GitHub search isn’t finding it. This is also why I ask about #1, as the way I would do this in my production code is that I would roughly just update interpreter constraints, and then update the PBS inside a lift.toml, and run my tests. A few minutes of work

There’s a Discussion somewhere about upgrading Python and definite work towards it too, such as that in scie-pants and even prep in the pants repo itself

Ahh, yeah, there are deprecations in Python 3.12 (at least) that we use - so not as straight forward. Womp womp .....

do we have a procedure/GH Issue for upgrading the version? Working through deprecations sounds like something many people could do

I’m more curious if no-gil in 3.13 will provide some longer term benefits - given how we currently operate

Which deprecations are these?

something about pkgutil (or importlib), at the very least,,,, Hilariously, it might have been undeprecated in 13 (need to check which items in the module, specifically)

Prior discussion: https://github.com/pantsbuild/pants/discussions/20287 We've done more work since then, e.g. the scie-pants change, and https://github.com/pantsbuild/pants/pull/20317

do we have a procedure/GH Issue for upgrading the version? Working through deprecations sounds like something many people could do

Are you in a position to synthesise the prior discussions, make a plan and write it up? Because, as you say, sounds like this is something people can chip away at if we have guidance