Context: I'm trying to work around the 1GByte limit for caches. That 1GByte limit is really painful, creating a lot of work to try to come up with work-arounds.
Anyway, I' tried to split up a cache by using Globs. According to the documentation you can use globs in a cache path.
So, I declared a cache, like this:
coursier-cache-orgs:
key:
files:
- "**/*.sbt"
path: '~/.cache/coursier/v1/https/*.org'
It does cache the data in that glob. However, it completely screws up the restoring of the cached items.
The original layout of the directory is something like:
- ~/.cache/coursier/v1/https/repo1.maven.org
- ~/.cache/coursier/v1/https/repo.jenkins-ci.org
So, it does cache these *.org directories, but it fuses them. So, when restoring it is:
- ~/.cache/coursier/v1/https/*.org/
That is of course completely useless, as the layout got changed.
