I suspect that the unit tests are failing due to a bottle update.

I suspect that the unit tests are failing due to a bottle update.

Is the failure coincidental because of an update in bottle, or is this directly related to the Python 3.13 update? (Which seems ... aggressive, since we're mostly still using 3.10 or even 3.9 ... but maybe not.)

The bot thinks that we should be using the "current" version of Python for our base image (hence this PR). Opening the PR triggered a CI run which installed a fresh version of the dependencies, which includes the "current" version of bottle, which, apparently, has the same incompatibility as Pbench faced. And, so, the unit tests failed.

I don't think this has anything to do with Python 3.13. I think it's all about bottle and how we use it. Since we're not rev-locking our dependency on bottle, I guess there was no need for the bot to propose an update to it, so we didn't notice the problem until a different update came by and prompted us to rebuild.

At this time, I'm not particularly excited about updating Python. We probably should make a change here to address the bottle problem, one way or another, though. But, I'm not feeling particularly motivated to do that, right now, either. So, I made my comment above as a note to a future reader (who might well be me...), and I figured I would leave the PR open until we feel like dealing with it.