Question

Doing stuff like "git ls-files" on big repos I have noticed that git is quite fast in this regard, subjectively faster than "find . -type f" would be I think (although I haven't done benchmarks).

Is git using some sort of binary key-value data store internally for storing blobs/deltas, like berkeley db or tokyo cabinet?

Était-ce utile?

La solution

yes Git is basically a key-value store. A more technical name would be a content-addressable file system. Any kind of data stored within git can be retrieved using a key, which is a SHA-1 hash. There is a whole section on git objects in the git book

Licencié sous: CC-BY-SA avec attribution
Non affilié à StackOverflow
scroll top