[ty] Make ecosystem-analyzer the default workflow#23996
Conversation
|
| Lint rule | Added | Removed | Changed |
|---|---|---|---|
invalid-await |
40 | 0 | 0 |
invalid-return-type |
1 | 0 | 0 |
| Total | 41 | 0 | 0 |
|
|
I'm not sure I'm the best reviewer here because I'm happy to go with the consensus opinion in the team (I don't feel too strongly), but I think I disagree with the consensus 😆 I'm supportive of having ecosystem-analyzer run by default on PRs, but I personally still find mypy_primer very useful because:
|
|
Given that the main advantage that mypy_primer still has to offer over ecosystem-analyzer is speed, I'm not sure making it a label-based workflow makes sense. By the time I've remembered to add the label to the PR, ecosystem-analyzer might have already run for 1m, so there's no longer any point in me adding it to my PR. I think I'd probably be inclined to just remove the workflow if we don't want it enabled by default. |
I guess it doesn't hurt to leave it in for a few more weeks at least. I was expecting another argument from you where you previously argued that having the "exact same" workflow like typeshed, mypy, etc. would be beneficial for comparison reasons? |
I'd love that. I find having the actual diff really useful.
Oh, sorry, to be clear -- I meant that I found mypy_primer almost entirely useless during that period because the PR-comment diffs were so noisy; I had to rely only on ecosystem-analyzer :) (And I found this very painful because, as you say, the very-useful flaky-diagnostic detection ecosystem-analyzer did made it extremely slow!)
I don't remember making that argument... anyway, it doesn't feel that strong to me now, even if I did make it in the past 😆 I do still wonder whether we could contribute some of these features from ecosystem-analyzer upstream, so that other projects that use mypy_primer could also benefit from them. But that's obviously not high-priority |
Ok, I'll work on this. I'm merging this for now to raise the stakes for ecosystem-analyzer improvements. |
* main: (24 commits) Update astral-sh/setup-uv action to v7.6.0 (#24003) [ty] discover /usr/local/lib dist-packages on Debian/Ubuntu (#23797) [ty] Rename and invert logic of ReturnsNever constraints (#23997) [ty] Make ecosystem-analyzer the default workflow (#23996) [ty] Include CPython projects in ecosystem-analyzer runs (#23995) Update docker/setup-buildx-action action to v4 (#23992) Update actions/download-artifact digest to 484a0b5 (#23974) Update docker/metadata-action action to v6 (#23991) Update docker/login-action action to v4 (#23990) Update docker/build-push-action action to v7 (#23989) [ty] Completely remove the `NoReturn` shortcut optimization (episode 2) (#23994) Update dependency mkdocs-material to v9.7.4 (#23978) Update dependency ruff to v0.15.6 (#23979) Update dependency astral-sh/uv to v0.10.10 (#23977) Update cargo-bins/cargo-binstall action to v1.17.7 (#23975) Update CodSpeedHQ/action action to v4.11.1 (#23976) Update Rust crate jiff to v0.2.23 (#23982) Update Rust crate getrandom to v0.4.2 (#23981) Update actions/setup-node action to v6.3.0 (#23987) Update Rust crate toml to v1.0.6 (#23984) ...

Summary
mypy_primerlabel)Test Plan