fix for python 3.11

[kliem]

This shall fix #114.

It is supposed to replace:

• add test for python 3.11 #115, as it also provided the CI tests for python 3.11
• Python 3.10 compatibility #103 (which is included, but fixed in terms of depending only on cython 0.29).

[mkoeppe]

build failure in https://github.com/sagemath/cypari2/actions/runs/3165731973/jobs/5154977510#step:5:1069

[mkoeppe]

Builds with "snapshot" are failing, for example https://github.com/sagemath/cypari2/actions/runs/3166111784/jobs/5155607498

Should a test with pari 2.15 be addded? (See https://trac.sagemath.org/ticket/34537)

[kliem]

@mkoeppe: Regarding your first comment, it is always good to add files to github...

The issue with pari > 2.13 is tracked in #111.

Sure we could add tests for 2.14 and 2.15. However, there is a already a lot of tests. Maybe it makes sense to go away from the 2-dimensional solution and instead test various versions of pari with a default python and various versions of python with a default pari.

Or a different approach would be to reuse pari installations for different python versions (similar to some of your sage workflows). That would be lot's faster, wouldn't it?

[mkoeppe]

LGTM

[videlec]

Builds with "snapshot" are failing, for example https://github.com/sagemath/cypari2/actions/runs/3166111784/jobs/5155607498

Should a test with pari 2.15 be addded? (See https://trac.sagemath.org/ticket/34537)

Yes. Some tests need to be adapted. Nothing dramatic though. See #111

[videlec]

Thanks for the fix.

In the future, please avoid to merge branch that do not pass the test suite. Priority is now #111.

[kliem]

The tests aren't really working at the moment and yes, pari 2.13 is not yet supported.

[videlec]

Sure, I know. I was trying to suggest that we could have merged the (not yet available) fix for pari > 2.13 first.