Russell Keith-Magee

> This further muddies what exactly a "Merge All Windows" menu command should, by default, do. tl;dr - it should do whatever macOS does. "Merge all windows of the same...

> Not sure what was going on with artifact downloading, but I guess it's working now? Yeah - that's all fallout from the recent `actions/download-artifact@v4` update. Github thinks it's *amazing*,...

> Incidentally — and probably relatedly — I notice the method is "private". If I get it working on both normal and native handler items, and now that it's being...

> > Not to derail this PR...but just wanted to note I've seen [Wandalen/wretry.action](https://github.com/Wandalen/wretry.action) implemented with `download-artifact@v4`....not great...but technically another option rather than rolling back. > > This is still...

I believe this is due to the emulator you have selected. The testbed is designed to run on a very specific emulator - in particular, a device with an Android...

You can also make the tests pass locally by creating an Android image using the Pixel 3a skin. The CI configuration has the invocation to do this.

Yes - because you haven't escaped the JSON for use in Windows. The CI configuration is running on Ubuntu, so it's able to use Unix quoting. The value that is...

This is a weird error - it shouldn't be possible for `self.interface` to be None. I can only assume that it happens as some sort of garbage collection edge case...

Another variation on theme from [this run](https://github.com/beeware/toga/actions/runs/8870287628/job/24353355496?pr=2539) ``` _______________________ ERROR at setup of test_load_url ________________________ Traceback (most recent call last): File "/Users/runner/work/toga/toga/testbed/build/testbed/macos/app/Toga Testbed.app/Contents/Resources/app_packages/pluggy/_callers.py", line 156, in _multicall teardown[0].send(outcome) File "/Users/runner/work/toga/toga/testbed/build/testbed/macos/app/Toga...

Thanks for the report. I've seen variations on this problem in the past; in the past, the issue has been that the underlying container on the non-visible tab isn't getting...