Re: tracking two git branches at once

[Graham Percival]

Mats Bengtsson wrote:
> I know very little about GIT, but if I understand you correctly, the 
> main reason
> to do the clone locally is to save disk space and bandwidth, right? 

Yes.  Since git keeps all changes for the entire repo, there's no point 
in storing all that data twice.  If we used a separate system for 
web/master, I'd just leave them separate, but there's no point in 
storing an extra copy of all the changes to the lilypond source tree 
just so I can make one-line changes to html files.


I have it all set up now, and it works great!  I heartily recommend it 
to all developers.

cd /path/to/main/source/tree
git checkout web/master
git clone -l -s -n . web
git checkout master   (or whatever branch you normally track)

# make sure that web/.git/config contains lines that look something like 
this:
[remote "origin"]
        url = /Users/gperciva/usr/src/lilypond/.git
        fetch = +refs/heads/*:refs/remotes/origin/*
[branch "web/master"]
        remote = origin
        merge = refs/heads/web/master


Now you can make changes to web/whatever, do a git commit, and then do a 
git push.  That sticks your changes in your main lilypond git directory; 
from there you can do a git push git+ssh+//blahblah  like normal.

Cheers,
- Graham