lix-project/lix

Fork 48

add ruff linter to functional2 pytests #812

New issue

Closed

opened 2025-04-28 14:55:58 +00:00 by commentatorforall · 11 comments

commentatorforall commented

2025-04-28 14:55:58 +00:00

Member

Currently there is no linting or anything else denoting common code styling for functional2 tests.
Hence, there is also no pre-commit/mr check if the guidelines are followed

Describe the solution you'd like

As has been discussed in the Lix-Development channel somewhen, ruff should be added as a linter.
It supports pre-commit hooks, which should be enabled as well.
The style configuration can be added to the pyproject.toml

I would suggest having all configuration options enabled and clearly set to avoid discussions and differentiating code-styles in the future

Describe alternatives you've considered

black has also been suggested, but disregarded in advantage of ruff, as the latter is faster and a drop-in replacement for black

Additional context

ruff has also been mentioned in the wiki as to be added at a later point in time

My intention for this issue is for it to be a discussion point for which style guides to use and to keep track of who is currently working on the implementation

## Is your feature request related to a problem? Please describe. Currently there is no linting or anything else denoting common code styling for functional2 tests. Hence, there is also no pre-commit/mr check if the guidelines are followed ## Describe the solution you'd like As has been discussed in the Lix-Development channel somewhen, `ruff` should be added as a linter. It supports pre-commit hooks, which should be enabled as well. The style configuration can be added to the pyproject.toml I would suggest having all configuration options enabled and clearly set to avoid discussions and differentiating code-styles in the future ## Describe alternatives you've considered `black` has also been suggested, but disregarded in advantage of `ruff`, as the latter is faster and a drop-in replacement for `black` ## Additional context ruff has also been mentioned in the [wiki](https://wiki.lix.systems/books/lix-contributors/page/code#bkmrk-python) as to be added at a later point in time My intention for this issue is for it to be a discussion point for which style guides to use and to keep track of who is currently working on the implementation

commentatorforall added the

testing

Language/Python

labels

2025-04-28 14:55:58 +00:00

commentatorforall added this to the functional2 project

2025-04-28 14:55:59 +00:00

lix-bot commented

2025-05-04 20:50:03 +00:00

Member

This issue was mentioned on Gerrit on the following CLs:

commit message in cl/3085 ("functional2: Added ruff formatter")

This issue was mentioned on Gerrit on the following CLs: * commit message in [cl/3085](https://gerrit.lix.systems/c/lix/+/3085) ("functional2: Added ruff formatter")

commentatorforall commented

2025-05-05 09:34:45 +00:00

Author

Member

should ruff be scoped to only run on functional2 or be scoped to include all python files within lix?

commentatorforall self-assigned this

2025-05-05 09:34:53 +00:00

jade commented

2025-05-05 18:36:31 +00:00

Owner

Eventually having more linting is definitely a good thing, and I'm in favour, but starting with functional2 is a great idea as it'll let us make incremental progress.

commentatorforall commented

2025-05-05 18:40:08 +00:00

Author

Member

i definitely see the point, but if you check out the pr (the pre-commit stuff) it does introduce some imo hacky stuff

i definitely see the point, but if you check out the pr (the [pre-commit stuff](https://gerrit.lix.systems/c/lix/+/3085/4/misc/pre-commit.nix#122)) it does introduce some imo hacky stuff

jade commented

2025-05-05 18:52:28 +00:00

Owner

That's totally okay for now.

kfearsoff commented

2025-05-05 19:18:45 +00:00

Member

Love linting, love pre-commit hooks. I'll check out the CL later and try my best to review.

Open Q: can Ruff be beat into working with Xonsh, or is it completely impossible? It might give us clues as to where to go long-term (like maybe massively shrinking Xonsh code size in favor of pure Python).

Love linting, love pre-commit hooks. I'll check out the CL later and try my best to review. Open Q: can Ruff be beat into working with Xonsh, or is it completely impossible? It might give us clues as to where to go long-term (like maybe massively shrinking Xonsh code size in favor of pure Python).

jade commented

2025-05-06 02:14:56 +00:00

Owner

@kfearsoff wrote in #812 (comment):

Open Q: can Ruff be beat into working with Xonsh, or is it completely impossible? It might give us clues as to where to go long-term (like maybe massively shrinking Xonsh code size in favor of pure Python).

It doesn't matter much, there's nearly zero logic in xonsh files today, so a linter doesn't achieve very much. Trying to get it working on xonsh doesn't seem like a particularly important job for that reason imo.

@kfearsoff wrote in https://git.lix.systems/lix-project/lix/issues/812#issuecomment-10578: > Open Q: can Ruff be beat into working with Xonsh, or is it completely impossible? It might give us clues as to where to go long-term (like maybe massively shrinking Xonsh code size in favor of pure Python). It doesn't matter much, there's nearly zero logic in xonsh files today, so a linter doesn't achieve very much. Trying to get it working on xonsh doesn't seem like a particularly important job for that reason imo.

commentatorforall commented

2025-05-06 15:59:59 +00:00

Author

Member

About the docstrings:
as of rn, pydoclint doesn't support sphinx docs yet; a pr is open though (last interaction in september 2024)
supported styles are google, numpy and pep275

should we move to one of those or wait until sphinx is supported?
personally i am not a fan of the other docstring styles.

About the docstrings: as of rn, pydoclint doesn't support sphinx docs yet; a pr is open though (last interaction in september 2024) supported styles are google, numpy and pep275 should we move to one of those or wait until sphinx is supported? personally i am not a fan of the other docstring styles.

commentatorforall commented

2025-05-07 13:33:16 +00:00

Author

Member

continuing on that, while pydoclint itself supports sphinx styling, the ruff implementation doesn't yet (see https://github.com/astral-sh/ruff/issues/12434)

jade commented

2025-05-08 03:53:54 +00:00

Owner

goofy! can we ignore the problem for now? it seems plausible that we could just not lint doc strings until they invent a linter that supports them; it's not like sphinx doc strings are a rare use case.