As discussed in PR #229 , Python 3.9 has been released upstream, on AD, on CF and on GH Actions for some time, so it would be seem to be worth bumping the upper bound of the CI matrix at least to it (which would expand our support range while not proliferating CI jobs, per the current build matrix strategy) and fixing any issues that are found, as well as declaring support in the Trove classifiers. Furthermore, Python 3.10 was just formally released a couple weeks ago and is now available on all the same channels, so it should be usable in our CIs as well. If supported by PyQt/Pyside (at least the Qt6-based versions; we could potentially set up a custom build matrix if needed) and our other test deps, we could consider adding support for it as well and having it be the upper bound version in our matrix instead. This might be a simple change, or it would be rather involved (though more important) if we don't fully support it already.
As discussed in PR #229 , Python 3.9 has been released upstream, on AD, on CF and on GH Actions for some time, so it would be seem to be worth bumping the upper bound of the CI matrix at least to it (which would expand our support range while not proliferating CI jobs, per the current build matrix strategy) and fixing any issues that are found, as well as declaring support in the Trove classifiers. Furthermore, Python 3.10 was just formally released a couple weeks ago and is now available on all the same channels, so it should be usable in our CIs as well. If supported by PyQt/Pyside (at least the Qt6-based versions; we could potentially set up a custom build matrix if needed) and our other test deps, we could consider adding support for it as well and having it be the upper bound version in our matrix instead. This might be a simple change, or it would be rather involved (though more important) if we don't fully support it already.