DM-45794: Update the shared Ruff configuration #266
Conversation
- Exclude ASYNC109, which required adjusting the comment column - Drop the S113 exclusion, since its httpx bug was fixed - Add additional per-file exclusions useful for vertical monorepos - Drop the McCabe complexity override The exclusions for vertical monorepos unfortunately appear to require duplicating the exclusion lists, or at least I don't know a way to write a glob that matches the paths both with and without a prefix. Since they have to be duplicated, I made them exact copies so that future maintenance is an easy cut and paste. The McCabe complexity threshold increase from 10 to 11 was something I added for Gafaelfawr, but I found a better way to break apart the relevant Gafaelfawr function and other packages appear to not have trouble with a limit of 10.
There was a problem hiding this comment.
Is the idea that ruff-shared.toml should be the same between the PyPI and FastAPI templates? If that's the case we could explore turning ruff-shared.toml into a file template and then experiment with either a symlinking or adding something to the post gen hook that copies the file into place so that it's ultimately single-source. I don't know if that makes it harder to keep track of or not, though.
|
Yes, indeed, I think we can just share the same copy everywhere, and I haven't figured out the best way to do that yet. I'm on board with doing it -- it just wasn't immediately obvious to me how that worked. Is there a place where we're already using that pattern? If I have a model to follow, I can make that change as a separate PR. |
|
I thought I had done this with the license/COPYRIGHT files and perhaps the |
The exclusions for vertical monorepos unfortunately appear to require duplicating the exclusion lists, or at least I don't know a way to write a glob that matches the paths both with and without a prefix. Since they have to be duplicated, I made them exact copies so that future maintenance is an easy cut and paste.
The McCabe complexity threshold increase from 10 to 11 was something I added for Gafaelfawr, but I found a better way to break apart the relevant Gafaelfawr function and other packages appear to not have trouble with a limit of 10.