python312Packages.matplotlib: 3.8.4 -> 3.9.0

[doronbehar]

Changelog: https://github.com/matplotlib/matplotlib/releases/tag/v3.9.0

Description of changes

Currently still a draft, due to this upstream issue (that apparently only we experience?). Also intentionally I target master and not staging, so that I'll be able to use the CI more easily once this gets ready.

I'm still opening this, to let the maintainers know that I'm "on it", and to share the slightly non-trivial work needed to get this working.

As for the last commit that enables the tests, I'm fairly sure that at least we won't get circular import errors as described in #255262 , but perhaps many tests fail due to our freetype library different then upstream (and upstream compares the images bit by bit). So I'll see how CI runs the tests for all platforms and decide than (once upstream issue will be resolved).

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandboxing enabled in nix.conf? (See Nix manual)
  • sandbox = relaxed
  • sandbox = true
• Tested, as applicable:
  • NixOS test(s) (look inside nixos/tests)
  • and/or package tests
  • or, for functions and "core" functionality, tests in lib/tests or pkgs/test
  • made sure NixOS tests are linked to the relevant packages
• Tested compilation of all packages that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage
• Tested basic functionality of all binary files (usually in ./result/bin/)
• 24.11 Release Notes (or backporting 23.11 and 24.05 Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
  • (Module updates) Added a release notes entry if the change is significant
  • (Module addition) Added a release notes entry if adding a new NixOS module
• Fits CONTRIBUTING.md.

---

Add a 👍 reaction to pull requests you find important.

[doronbehar]

First of all, I managed to fix the mentioned compilation issue. Besides that, instead of enabling tests by default, I decided to put a passthru.tests.withOutdatedFreetype derivation that will use the same freetype that upstream used to generate the original images they compare bit by bit..

I plan to put submit a version of this PR to staging-next or staging after CI is green here.

[doronbehar]

The x86_64-darwin build and tests CI job is green :) However for aarch64-linux there is a single test failure (from matplotlib.passthru.tests.withOutdatedFreetype):

=================================== FAILURES ===================================
_________________________ test_preset_clip_paths[png] __________________________

args = ()
kwds = {'extension': 'png', 'request': <FixtureRequest for <Function test_preset_clip_paths[png]>>}

    @wraps(func)
    def inner(*args, **kwds):
        with self._recreate_cm():
>           return func(*args, **kwds)
E           matplotlib.testing.exceptions.ImageComparisonFailure: images not close (RMS 0.015):
E               result_images/test_axes/preset_clip_paths.png
E               result_images/test_axes/preset_clip_paths-expected.png
E               result_images/test_axes/preset_clip_paths-failed-diff.png

../czlizmr5dmpxi7ysqknxys82776hldgk-python3-3.11.9/lib/python3.11/contextlib.py:81: ImageComparisonFailure

It would have been interesting to inspect the diff, but I suppose we don't have anybody with an aarch64-darwin machine to do it for us. We also can't submit a bug report upstream about it due to this - they would probably want the images we generated.

In a dilemma whether to submit a PR to staging or to staging-next (#316075) .

[doronbehar]

staging PR for reference: #319357 .