Tags: lloydchang/facebook-pyrefly
Tags
Fix incremental panic Summary: This should fix facebook#199 Unfortunately the issue is not easily reproducible, but I'm fairly certain I understand what is happening (we're mixing type values coming from incompatible versions of the same module, and getting arity mismatches) The only reported panic caused by this happens in `subset`, so we can just make this robust. The only obvious way to deal with it is to say that if arities mismatch, then the two classes are from different versions of the code and for now they are not linearly orderable (trying to do anything else is tricky because we'd have work out downstream logic). Reviewed By: ndmitchell Differential Revision: D74819844 fbshipit-source-id: 718dea117b61001ca02eb7934af433a452cb81e7
Increase timeout for extension test Summary: test_extension is sometimes failing: https://github.com/facebook/pyrefly/actions/runs/14940864076/job/41977764460. The failure message says that we're hitting a timeout of 2000 ms. The test seems happy after I increased the timeout to 5000 ms: https://github.com/rchen152/pyrefly/actions/runs/14940926912/job/41977931542. Reviewed By: yangdanny97 Differential Revision: D74520364 fbshipit-source-id: edd0613e9f52f2ea0dcbf7d343026d9e0221fe3a
Bump patch version Summary: Two days ago, we restructured the typeshed directory used in Pyrefly's build. This means it is not possible to run mypy_primer without a lot of extra flags, because the defualt behavior is to compare the latest release vs trunk for a typechecker, and one or the other is going to panic (because the build scripts are coupled to the directory layout). Triggering a patch version bump because there are no major changes, we just need a tag to exist so mypy_primer can find it. Reviewed By: rchen152 Differential Revision: D73692380 fbshipit-source-id: 549f9df493f0417e10da7e855f4c6367fbcfe229
PreviousNext