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
Zinc has a lot hooks to improve incremental compilation performance in case of really big projects. However when we hash classpath we just do tons of IO operations without any way to e.g. reuse libs hashed before.
In case of 100+ projects with 100+ libs and on slower IO (e.g. windows) it takes a while during each compilation (even no-op ones).
Zinc has a lot hooks to improve incremental compilation performance in case of really big projects. However when we hash classpath we just do tons of IO operations without any way to e.g. reuse libs hashed before.
In case of 100+ projects with 100+ libs and on slower IO (e.g. windows) it takes a while during each compilation (even no-op ones).
I haven't found a way to do it without breaking binary compatibility but simple proposal how it can be done is here: https://github.com/romanowski/zinc/tree/hashing-classpath-proposal
@stuhood I think pants may also benefit from caching classpath hashes.
The text was updated successfully, but these errors were encountered: