Good work tracking down this obscure failure!
Aug 22 2020
Aug 10 2020
Aug 8 2020
A similar patch already landed. I'm going to commandeer and close.
I'm not sure what the README.rst rename was doing in the diff. So I dropped it in-flight.
Debian's perturbations to how Python is packaged continue to confound me. There's aren't enough Picard facepalm memes to express my feelings on the matter.
@durin42 what's the state of this series? Can the bottom part of it be reviewed or do you plan to revisit the entire series? Perhaps a rebase would be in order, just in case?
I'm going to approve but will hold off landing until the following patch solidifies a bit.
FWIW I encountered the following test failure when running this series locally:
I guess we lose some typing as part of this change. But that only guards against unknown actions, which I suppose shouldn't happen.
Yeah, I suppose mutation during iteration will make a lot of noise. So as long as we have test coverage of this code, we should be fine.
The introduction of sorting here is interesting. I guess we never sorted filenames before performing working directory modifications before? But surely there is sorting happening elsewhere because anything with ui output would need to be sorted to ensure determinism. And surely tests would have caught that.
Aug 6 2020
I left a handful of comments.
Aug 5 2020
I need to think about the implications of this change a bit more: the requires file plays an important part in how clients open repositories and there are significant backwards and forwards compatibility concerns.
Aug 3 2020
Having thought about this a bit more, I think using the importlib.resources high-level APIs are better. i.e. this series as-is is fine.
Aug 2 2020
Oh yay - I should have looked at the next commit before commenting on the prior one ;)
I'm only +0 on this refactor, as I don't think it changes readability much. But whatever.
This is a nice cleanup and improves readability!