Compact Git repository with history depth

Question

In Git, there is a command git clone --depth=<depth> to retrieve historical data only specific length. there are also a command to gather more historical data by use git fetch --depth=<depth> .

How about when we want to free some spaces from large repository? I know that we may use git gc or git prune but are there other way to specific like --depth=<depth> to reduce number of commit store in local repository? AND it also should keep SHA1 to be able continue to working with it.

Answer 1

The easiest way would be to:

• delete entirely the current local repo
• git clone --depth=n /url/of/remote/repo

That would clone the last n commits, while allowing fetch/pull/psuh to still work with the remote repo.

since Git 2.5, you can fetch a single commit , but unless that commit is the latest one (which is like a git clone --depth=1 ), that would not allo for fetch/pull/push.

The other approach to make sure a given local repo is as lean as possible is to use a combination of gc/prune/repack :

git gc
git repack -Ad      # kills in-pack garbage
git prune           # kills loose garbage