[Biopython-dev] Possible re-import from CVS to git

[Peter]

On Wed, Apr 22, 2009 at 5:32 AM, Eric Talevich <eric.talevich at gmail.com> wrote:
> On Tue, Apr 21, 2009 at 2:18 PM, Peter <biopython at maubp.freeserve.co.uk>wrote:
>
>> So, if we are going to re-do the git import (and we may have to fix
>> the tag history), ...
>
> Eek. Now that the Summer of Code is under way, I guess this is a good time
> to bring up the question of how Nick and I should be following the Biopython
> trunk and publishing our own code.
>
> In spite of the warning that the CVS tracker in GitHub was tentative, I was
> getting comfortable with the setup we had. Should I (we) hold off on pushing
> anything substantial to GitHub until this tagging situation is resolved, or
> is there a better way to approach this? For example, does anyone know if
> it's straightforward to back up a branch's recent history with
> git-format-patch and apply it directly onto a new repository with different
> references?

Bartek is looking into fixing the existing CVS to git mirror on github, but
that may not be possible.  And I do think it is worth fixing the tag history
even at the cost of some upheaval in the short term.

In terms of you and Nick, for now carry on using github if you are comfortable
with it.  The new phylogenetics stuff will I assume be mostly new python
modules, or modifications to a couple of existing ones (e.g. Bio.Nexus).
Merging this later shouldn't be too bad - you should be able to generate a
diff against CVS (or its current mirror in git) and we can apply that to CVS
(or a new git repository).

Peter