Hi,

On Fri, Aug 26, 2011 at 11:31 AM, Gleb Kurtsou <gleb...@gmail.com> wrote:

Here you go:

$ mkdir small-clone $ cd small-clone $ git init # Tell git to fetch only master (AKA head) and stable/8 branches $ git remote add -t master -t stable/8 origin git://git.freebsd.your.org/freebsd.git $ git fetch origin # fetches about 1.6M objects

remote: Counting objects: 1600952, done. remote: Compressing objects: 100% (418377/418377), done. remote: Total 1600952 (delta 1227540), reused 1524707 (delta 1159458) Receiving objects: 100% (1600952/1600952), 632.85 MiB | 3.12 MiB/s, done. Resolving deltas: 100% (1227540/1227540), done. From git://git.freebsd.your.org/freebsd * [new branch] master -> origin/master * [new branch] stable/8 -> origin/stable/8

# Add stable/7 branch and fetch it $ git remote set-branches origin --add stable/7 $ git fetch origin

remote: Counting objects: 35713, done. remote: Compressing objects: 100% (11359/11359), done. remote: Total 27595 (delta 20679), reused 22440 (delta 16043) Receiving objects: 100% (27595/27595), 14.40 MiB | 1.93 MiB/s, done. Resolving deltas: 100% (20679/20679), completed with 2329 local objects. From git://git.freebsd.your.org/freebsd * [new branch] stable/7 -> origin/stable/7 $ git branch -r origin/HEAD -> origin/head origin/master origin/stable/7 origin/stable/8

Practically it does not save you all that much on repo size or the time to fetch stuff. Complete clone is ~700M/1.9M objects and head+stable/8 clone is only about 10% smaller. If you really want to have smaller repo you will need to trip repo history. I'm not sure yet how to do that.

--Artem