Update tests

[glatterf42]

The current CI failures have multiple causes. One is pytest 8.0.0, for which this PR adjusts. The affected test was working properly, as far as I can tell, by raising a ValueError before the DeprecationWarning. In the past, since a ValueError was given as the expected result of the test, this was fine; but now, pytest wants to see the DeprecationWarning.
On other tests, one can run with pytest.raises(ValueError, match="something") to specify which ValueError is raised -- which would have made us aware of this issue long ago since Scenario.add_horizon() raises ValueErrors on multiple occasions. From pytest's docs, it doesn't look like we can pass a match= string to the xfail mark, so I'm wondering how much refactoring if would take to implement this. Another option would be to subclass ValueError and raise custom unique errors for the various fail cases, which we can then tell pytest to expect.

This PR also temporarily pins genno < 1.23 and can be used to upgrade Node.js actions as GHA recommends.

How to review

• Read the diff and note that the CI checks all pass.

PR checklist

• Continuous integration checks all ✅
• Add or expand tests; coverage checks both ✅
• [ ] Add, expand, or update documentation. Just updating tests
• [ ] Update release notes. Just updating tests

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (7ffee7f) 95.2% compared to head (96041ca) 95.2%.

Additional details and impacted files

@@          Coverage Diff          @@
##            main    #789   +/-   ##
=====================================
  Coverage   95.2%   95.2%           
=====================================
  Files         46      46           
  Lines       4335    4335           
=====================================
  Hits        4130    4130           
  Misses       205     205           

Files	Coverage Δ
message_ix/tests/test_core.py	100.0% <100.0%> (ø)

[khaeru]

Seems good! One change and one question, maybe a change.