# Suggestions for my use case

1 messages from 2006-09-02 to 2006-09-02. Participants: Russell.
Thread: https://gitlist.dev/t/5455

## Russell, 2006-09-02 14:41

Subject: Suggestions for my use case
Message-ID: <c1b8b6670609020741q2d45ea3nf336611edcd452b6@mail.gmail.com>
URL: https://gitlist.dev/e/c1b8b6670609020741q2d45ea3nf336611edcd452b6%40mail.gmail.com

```
Hello,

I've been tring git for a few private projects and have come up with a
way to organise distributed development on multiple machines.  I was
wondering if anyone else could comment on what I've done?  Here it
is...

I want to be able to work on a project called proj on multiple
computers, and have backups on removable media.  Let's say I have
computers c1 and c2, and a removable hard disk labelled disk1.  I've
setup the repositories in

  c1: /home/me/proj
  c2: /home/me/proj
  disk1: /media/disk1/me/proj

In c1:/home/me/proj/.git/remotes/c2
  URL: ssh://c2/home/me/proj
  Pull: refs/heads/master:refs/heads/host/c2
  Push: refs/heads/master:refs/heads/host/c1

In c1:/home/me/proj/.git/remotes/disk1
  URL: /media/disk1/me/proj
  Pull: refs/heads/master:refs/heads/disk/disk1
  Push: refs/heads/master:refs/heads/host/c1

In c2:/home/me/proj/.git/remotes/c1
  URL: ssh://c1/home/me/proj
  Pull: refs/heads/master:refs/heads/host/c1
  Push: refs/heads/master:refs/heads/host/c2

In c2:/home/me/proj/.git/remotes/disk1
  URL: /media/disk1/me/proj
  Pull: refs/heads/master:refs/heads/disk/disk1
  Push: refs/heads/master:refs/heads/host/c2

c1 has these branches:
  master
  host/c2
  disk/disk1

c2 has these branches:
  master
  host/c1
  disk/disk1

disk1 has these branches:
  master
  host/c1
  host/c2

So after doing some work at c1 and committing into master, I can
  c1$ git-push c2
  c1$ git-push disk1

Then when I sit down at c2:
  c2$ git-pull . host/c1

Or if I've forgotten to push from c1:
  c2$ git-pull c1

Or I can cart disk1 around, and
  c2$ git-pull disk1

Each repository stays on its master branch.  As per the git-pull man
page, no development is done on branches that I pull from.

This is working fine for me at the moment, which is single person
development, straight line, no branching.  When I come to branches, I
think the branches will need to be added in each repository, with push
and pull lines in .git/remotes/c1 etc.

Does this look reasonable?  What have I missed that would make this
more manageable, particularly with respect to branching?

-- 
VGER BF report: U 0.500329

```