Haml on Git

I’ve been playing around with distributed revision control recently. I have another post in the works all about that, so I’ll skip over the details for now.

The upshot of my adventures is that I’m pretty much convinced that there’s almost no reason to use a client-server system rather than a distributed one¹. Distributed systems have all of the capabilities of client-server systems, and more besides.

So you can imagine my pleasure when Hampton IMed me a few days ago saying “What do you think about moving Haml to Git?” I replied that I thought that was a fantastic idea.

I imported the Haml Subversion history and set up the repository last night. Today I set up the web interface. Haml is now offically stored in a Git repository. Check it out via your broser through the web interface.

To get the development branch of the code, just run

git clone git://github.com/nex3/haml.git

To get the stable branch, cd into the new haml/ directory and run

git checkout origin/stable

When creating patches, the process goes like so:

emacs                   # Make your changes
git commit -a           # Commit all changes to your local repository
git format-patch origin # Create a patch: 0001-your-commit-message.patch

You can then email, pastie, or otherwise send the resulting patch file to Hampton or me. We’ll then apply it² (with git am), which will not only preserve your commit message but your authorship as well.

Git is awesome. Enjoy it.

¹ The only exception to this that I can think of is that more people have Subversion installed than any distributed tool. But if current availability was all that mattered, we’d all still be using CVS.

² Assuming it’s a good patch for something we want, of course. For the record, we accept many more patches than we reject.