You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Since adding a new Transpiler, we have some buggy code with CacheLayer I believe. It doesn't know when to break the cache for pre-commits. @samreid is this right?
The text was updated successfully, but these errors were encountered:
I'm more scared about it this now, because I really don't want to expose the need to regenerate all phet-io APIs for each repo. It seems like that cacheing, for multiple pre-commit tasks run in a row, is still working. Perhaps we need to keep that feature?
I would prefer not to have a watch process that invalidates caches
I would prefer not to try to leverage the SWC watch process for caching
Adding a new cache strategy like "did this test already pass within 5 minutes" could be buggy, and could "kick off" multiple tests at once if precommit hooks run in parallel anyways.
We recently got a speed boost in generating phet-io apis by caching in the web server layer.
We already opt out of testing when running in phet-io-sim-specific itself.
Pre-commit tests are supposed to be fast and independent. I believe that a more sophisticated "multi-repo-aware" strategy is the correct solution to this problem. But I do not know how developers could make that automatic.
Since adding a new Transpiler, we have some buggy code with CacheLayer I believe. It doesn't know when to break the cache for pre-commits. @samreid is this right?
The text was updated successfully, but these errors were encountered: