CI builds are failing

[hinerm]

Failing since 6b8c32f so poking @gselzer

I tried updating more of the action component but no luck

[gselzer]

Haven't dug too deeply, but the failure on 6b8c32f is not the same as the failure on later commits. I also have a hunch that the former was a one-off, unreproducible error, while the latter looks more troubling (but maybe only occurs on that one entry in the matrix??)

Will definitely peek next week when I have time

[gselzer]

Been digging into this a bit with #325. It seems like part of the problem is napari 0.5.6 - both pinning napari 0.5.5 and the current main branch seem to do better (could be related to napari/napari#7682), but still seeing failures. More digging will be necessary.

[hinerm]

For what it's worth, what I observed was consistent failures on Ubuntu with python 3.12. These seem to happen after the tests have completed, during JVM shutdown.

Then there is a separate failure in the Conda Setup & Code Coverage step which happens after utilties/widget tests. Removing those tests did cause the build to pass.

Failures seemed consistent with Jpype 1.5.2 and 1.4.1

I didn't see any changes in failures with pinning napari 0.5.5 personally

NB: these comments come from a starting point of updating the github action versions.

For the code coverage I also tried updating the xvfb-action and going back to coactions like the rest of the tests, both to no avail