Re: Git question on reviewing tests and contributing fixes

On Tue, Dec 16, 2014, at 15:07, Andreas Tolfsen wrote:
> On Tue, Dec 16, 2014 at 1:49 AM, Brad Hill <hillbrad@gmail.com> wrote:
> > Is there any way to do this?  Does bob have to add me as a collaborator on
> > his fork?
> 
> This is a limitation Github imposes, and Bob will have to add you as a
> collaborator on his repository.
> 
> I think the current suggestion is to drop Bob's original review and
> submit a new one under your name, which admittedly isn't great.

But not all bad either, because Bob will still be in the commit as
author, only the PR would be different.  Also, bringing "your" fork over
to wpt-repo then allows any of us with access to wpt to chip in.


**

I know we all dream of this, but I think it would be really good if
GitHub worked the way that you actually could push to the pulls/ heads 
(they store pull requests in refs/pulls/<id>/head in the WPT-repo if I
remember correctly).  I guess one problem would be that you would either
have to also update Bob's branch at that point, which feels very bad and
not distributed at all, or you couldn't track Bob's branch from the pull
request any longer then.

In other words, if someone pushed to   wpt:refs/pulls/<bobs_pr_id>/head 
 the tracking of  bob:refs/heads/my_branch  would be disabled on GitHub
side.   They should probably still allow bob to push into  
wpt:refs/pulls/<bobs_pr_id>/head   though, so that everyone in the
project + the original author can collaborate on the pull request.


I always thought it weird that GitHub has no way to do this.  But the UX
of it all does leave something to be desired.

-- 
  Odin Hørthe Omdal
  odinho@opera.com

Received on Tuesday, 16 December 2014 16:19:09 UTC