Prefer using a commit hash over a branch name

[sr229]

Tracking a cache by commit hash is much more consistent caching-wise and wouldn't require the system to constantly pull data every time a reload is needed since you only need to check for the latest commit every time.

[Earu]

It wouldn't change anything from what's already happening really, just a different hash.

If I wanted to do this I would probably have to do another request during init towards the GitHub API, and let's just say the less requests are done the better.

There are also other considerations,
Like what if I wanted to add a source that's not GitHub ?