#How does function caching interact with
slate moat
They both exist. The function caching applies to whether the entire function will execute, whereas layer caching is the same as before.
Are you asking whether the function cache TTL/no-cache/etc. applies to containers created inside the function? If so, no it doesn't, at least not yet. It only applies to whether the function itself is invoked to get a result or just read from cache.
If you were asking something else lemme know 🙂
frank lance
Ok I think that makes sense. Yes, that's what I was asking. For example, I have a "artifactPush" that I want to always run. But just setting // +cache="never" isn't sufficient according to the above. I'd also have to do the timestamp env var to actually ensure that the withExec is always run.
I think the function caching docs would benefit from an update and cross link to - https://docs.dagger.io/features/caching with another cross link to invalidating the cache - https://docs.dagger.io/cookbook/builds/#invalidate-cache.
Docs around caching are a bit scattered at the moment and the multiple layers of complexity makes it more confusing 🙂
slate moat
Ok I think that makes sense. Yes, that's what I was asking. For example, I have ...
Agreed, adding fixing that to my TODO list 🙂
slate moat
