Re: clone branching?

From: Karl Hasselström <kha@treskal.com>
Date: Aug 12, 2008, 07:34 UTC
Message-ID: <20080812073435.GA30811@diana.vm.bytemark.co.uk>
In-Reply-To: <4b6f054f0808111324x5a7c44afj6f93bc1e36fc481a@mail.gmail.com>

On 2008-08-11 16:24:57 -0400, Trans wrote:

Show 18 quoted lines

> On Mon, Aug 11, 2008 at 4:39 PM, Karl Hasselström <kha@treskal.com>
> wrote:
>
> > git supports multiple branches per repository/working copy, but
> > doesn't force you to have more than one. And it's very easy to
> > pull changes from one repo to another.
>
> In retrospect, It's more the tags I need rather then branches, but
> just the same... So if I have a repository like:
>
>   myapp/
>     1.0.0/
>     1.0.1/
>     1.1.0/
>     current/
>
> This would be okay? No other "giter" is going to look at this and
> think "what the hell?"

That would be OK. But "myapp" wouldn't be a repository -- it'd be a directory that contains four repositories. Each of those repositories would have a specific branch checked out, but they'd probably all contain the full set of tags and branches.

> Hmm... actually I don't see how this can work. If .get is under
> current/ than no one will be able to see the tags on my shared repo,
> but if it is under myapp/ then I'd be branching and tagging my tags
> --that's doesn't make any sense. I'm confused.

All your repositories have a .git directory. That's what makes them git reppositories.

I think what you need to do to get un-confused is to learn about how git stores history. Specifically, that

  1. the history consists of a DAG of commits, and

  2. that branches and tags are just named pointers to commits.

http://eagain.net/articles/git-for-computer-scientists/ is a short and sweet introduction.

-- 
Karl Hasselström, kha@treskal.com
      www.treskal.com/kalle