geda-developers team mailing list archive

[Peter TB Brett <peter@xxxxxxxxxxxxx>]

On Fri, 30 Aug 2013 19:50:05 +0100, "Peter C.J. Clifton" <pcjc2@xxxxxxxxx>
wrote:
> On 2013-08-30 19:10, Peter TB Brett wrote:
>> On Fri, 30 Aug 2013 13:57:58 -0400, DJ Delorie <dj@xxxxxxxxxxx> wrote:
>>> So let's figure out what *our* workflows require, and write up the
>>> best way to do those workflows.  I'm certainly not the one to train
>>> new developers on how to use git ;-)
>> 
>> The geda-gaf official git repository's workflow is:
>> 
>> 1) Check 'make distcheck' succeeds *before* pushing to the repository.
>> 2) Commit messages describe "why"; "what" is covered by the commitdiff.
>> 3) 'git push --force' is an admission that you totally screwed up.
>> 
>> Within those guidelines, anything goes really.
> 
> 
> I concur... I don't want to see --force used on PCB's repository either.
> 
> We might consider allowing a naming convention for more transient, or 
> prone-to-rewriting branches, but I'd probably prefer those to be private 
> branches.

You could call them "wip-<name>", for example.

> freedesktop.org (which hosts a lot of open source repositories) actally 
> host separate git repositories for individual users of the form 
> ~username/repo_name, e.g.:
> 
> OFFICIAL:      xorg/driver/xf86-video-intel
> Chris Wilson:       ~ickle/xf86-video-intel
> Keith Packard:     ~keithp/xf86-video-intel
> 

https://git.kernel.org/cgit/ is another similar example.  Lots of copies of
the Linux repo all over the place, named "kernel/git/<name>/<foo>.git"

Peter

-- 
Dr Peter Brett
http://peter-b.co.uk/