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
Affects 1.3.0 (maybe more).
Requirements: JVM project with zinc (see https://github.com/jsirois/lf-example for example) with multiple branches affecting the same class (in this case Example class)
Start off in branch A. Do a ./pants compile ::. See the Example class appear in the cache directory
Switch to branch B. Do a ./pants compile ::. See both versions of the Example class in the cache directory
Delete the cache directory for just the Example class
Switch back to branch A. Do a ./pants compile ::. Pants should recompile Example class but it does NOT write it to the cache.
It seems like the build_invalidator thinks that the class is in the cache, and when it isn't it gets recompiled but nothing gets pushed to cache. I believe the class should be written to the cache. Running pants invalidate between switching branches fixes the problem.
The text was updated successfully, but these errors were encountered:
how does pants determine that this is an "incremental" compile? Because it is not in a clean state?
the thing is, if i dont do step #3, everything works fine. It pushes the classes to the cache fine.
Affects 1.3.0 (maybe more).
Requirements: JVM project with zinc (see https://github.com/jsirois/lf-example for example) with multiple branches affecting the same class (in this case Example class)
It seems like the build_invalidator thinks that the class is in the cache, and when it isn't it gets recompiled but nothing gets pushed to cache. I believe the class should be written to the cache. Running pants invalidate between switching branches fixes the problem.
The text was updated successfully, but these errors were encountered: