This pull request has been mentioned on NixOS Discourse. There might be relevant details there:

https://discourse.nixos.org/t/prs-ready-for-review/3032/5117

nixpkgs-reviewd workflow finished

Status and logs: https://github.com/Bot-wxt1221/nixpkgs-reviewd/actions/runs/12865969944

OK. But that nixpkgs-review is a bit strange. Some package failed on github actions can be built on my self computer. I will try to re-run it.

Sorry for so many changes. All these is need for litestar. I don't think it should be splited into more PRs.

I think it should. Next time you're in this situation, you should open one merge request for each one of dependencies, andopen a draft PR for the package you're really interested in mentioning those as dependencies. It may seem a hassle for now, but believe me, you'll gain tons of time (and we will as well) in the long run:

• one good change won't be blocked by those which are starting debates
• nixpkgs-review will be faster
• you'll get less reviewers by PR (only maintainers of said package), so less reviews, less comments, faster iteration and - - and it'll be easier to track for everybody

So everybody go faster, contributors and reviewers.

Many packages seemed to fail because selenium. Which is update on the python-update branch. Should we target this to 375144?

Many packages seemed to fail because selenium. Which is update on the python-update branch. Should we target this to 375144?

Not sure. Usually what I do is rebase my PR onto the PR I need and mark mine as draft until the other is merged, but there may be a betters way. Anyway, you can try rebasing locally and see if you get lucky and the build works 😆

It's also possible that it is already broken on master (not everybody runs nixpkgs-review) and that the patch update won't fix it.

If the breakage is not directly linked to this MR, it's possible to merge it I think.

@autra I have tested that It succeed in building on python-updates branch.

@autra I'm not sure what you said for numpy. propagatedBuildInputs and dependencies are the same according to

@autra I'm not sure what you said for numpy. propagatedBuildInputs and dependencies are the same according to

nixpkgs/pkgs/development/interpreters/python/mk-python-derivation.nix

Lines 365 to 370 in d9e47de

	propagatedBuildInputs = validatePythonMatches "propagatedBuildInputs" (
	propagatedBuildInputs
	++ dependencies
	++ [
	# we propagate python even for packages transformed with 'toPythonApplication'
	# this pollutes the PATH but avoids rebuilds

yes. It may not be this. There are a few errors about numpy and type mismatch. The other possibility would be because the package get a wrong python version because it is not defined in python-packages. But that is just my intuition, I'm not sure.

@autra Could you list the package here. It's hard to find it.

Many packages seemed to fail because selenium. Which is update on the python-update branch. Should we target this to 375144?

Yes, the packaging issue should be fixed in latest selenium release - SeleniumHQ/selenium#14823