I read through the thoughtful discussion on GitHub and I wouldn't even imagined some of the factors, like XP in China, but it seems like the right thing to do, as people will still have access to py2 and sdl1.
~ Michael
On Saturday, October 9, 2021 12:43:08 AM PDT you wrote:
> hey hey,
>
> so, we've been talking about stopping support for these for a _while_.
> Should the next pygame release 2.0.2 be the last one that supports them?
> That means the pygame 2.1.0 series will be without SDL1 and py2 support.
>
>
> ---
>
> There's still lots of python 2.7 pygame users. Maybe more than mac users
> for example.
> But it's at least not increasing anymore, like it was only six months ago
> (near the number of python 3 users... yes really).
> https://github.com/pygame/pygame/issues/2326
>
> We get a cleaner code base, but also get to use modern C (or is it post
> modern C?). Mainly because python 2.7 on windows has to use an ancient
> compiler that doesn't support even all of C89. This means things like
> variable declarations anywhere we want, and better const/struct/union
> handling. Also a lot fewer ifdefs, and compatibility hacks.
>
> 3.5 was EOL'd about a year ago, and not so many people use it.
> With 3.6 being the minimum supported python, we'd also get to use
> contemporary python. Things like types, and even module level getattr.
>
> Probably it could be done in this order.
> - ripping out SDL1 (there's a PR for that
> <https://github.com/pygame/pygame/pull/2692> already)
> - ripping out pre-3.6 isms (mostly removing 2.x compat stuff).
>
>
> What about previous users? SDL1, and python 2.7 users have access to
> previous releases. They aren't going anywhere. New features, and bug fixes
> would be a problem of course. The Debian universe is still on pygame 1.9.6
> for example. A lot of people will just keep using that.
>
>
>
> There's some discussion on gh here:
> https://github.com/pygame/pygame/issues/2326
>
Attachment:
signature.asc
Description: This is a digitally signed message part.