[DGD] kotaka work

[Jared Maddox]

> Date: Thu, 12 Jul 2012 23:49:55 -0700
> From: Shentino <shentino at gmail.com>
> To: All about DGD and Hydra <dgd at dworkin.nl>
> Subject: Re: [DGD] kotaka work
> Message-ID:
> 	<CAGDaZ_r6Ax4Z8M01gEYnx+6S1BnJEQAvn=5trNcbWjYf86hZOA at mail.gmail.com>
> Content-Type: text/plain; charset=UTF-8
>
> So basically do I keep features in branches until they're finished,
> only push stable stuff to a release, or both?
>

I suggest having one single 'main' branch where the most up-to-date
code is located. Whenever you think you've gotten something to a
release stage, branch it off into a new numbered branch (e.g. 8.9) for
the sake of bookmarking it, and keep working in the main branch on
your next feature/bug/whatever.

In Linux, big projects apparently sometimes get their own branch for
development, which gets merged back into the main branch when it's
ready, but Linux is a big project with a lot of littler projects going
on all the time.

The only reason to keep new features in auxiliary branches before
they're ready is if they'll interfere with some other development
that's happening at the same time.

If you wanted to rewrite DGD from scratch, while maintaining the
original code, then the new version would probably belong in it's own
branch. However, if you wanted to add code to auto-save state whenever
an appropriate Power Management signal was received, then that would
probably be done in the main branch.