4 ms·

You will miss some. But then you just make a diff on the commits' SHA and only then you clone the repo (or fetch an archive, which is quicker). You save a lot

by fenollp 11y ago

You will miss some. But then you just make a diff on the commits' SHA and only then you clone the repo (or fetch an archive, which is quicker). You save a lot of time / bandwidth / space this way.

oxplot 11y ago
You only delay saving time and bandwidth — you eventually have to clone the repo for any ops other than checking the head SHAs and unless the repo is rarely updated, the delay will be short.
- fenollp 11y ago
  Use git-archive or curl: gzipped source code is always smaller than a full versioned repo…
  - oxplot 11y ago
    You don't need the full versioned repo, just the top commit (ie cloning with depth=1). Also, git compresses the objects before sending them.