[Biopython-dev] biopython on github

Peter biopython at maubp.freeserve.co.uk
Tue Mar 17 20:21:45 UTC 2009


2009/3/17 Tiago Antão <tiagoantao at gmail.com>:
> OK, in order to exercise and try github development I have forked a
> branch to work on the PopGen code. The idea of the branch is to serve
> as a platform for merging with the "official" branch. So, the idea is:
>
> 1. Official branch - The stable thingy
> 2. PopGen stabilizer branch - The place to merge contributions from
> PopGen development branches. The idea is that people can go crazy on
> their own branches and this intermediate one serves as a point to
> stabilize (unit test, documentation, QA, ...) before the commit to the
> official one
> 3. Crazy branches - Develop your crazy idea. I have 3 ideas myself:
> One for Jason's structure code, one for my LDNe code and another for
> statistics. Many more welcomed....
>
> The development procedure would be like this:
> A. People would have all the fun on their development branches
> B. When they felt confident they would submit their code to the
> stabilizer branch, where we would check that all the important things
> were there: unit test, code comments, QA, documentation
> C. When things were in good shape, we would propose changes to the
> official branch
>
> And, by the way, bug fixes of existing production would also be done
> on the stabilizer branch.
>
> Does this make any sense?

Totally.  But keep in mind the current "official" git branch (the one
being updated from CVS) may get nuked if we have to redo the import to
fix the missing version tags - so I would suggest you name your
branches with "test" or "provisional" or something temporary in the
text for now.

> In my view, with things like git, a policy like this encourages both
> innovation while preserving stability and robustness of the official
> branch.

Yes - and this like the right approach for Bio.PopGen, with you acting
as the gatekeeper.

Peter




More information about the Biopython-dev mailing list