[Bioperl-l] Merge branch 'master' ?

Chris Fields cjfields at illinois.edu
Tue Jun 8 09:55:51 EDT 2010


On Jun 8, 2010, at 6:35 AM, Jay Hannah wrote:

> On Jun 8, 2010, at 5:49 AM, noreply at github.com wrote:
>> Commit: 0e70e88a76d638d9f0406643c37691bb20d60ceb
>>   http://github.com/bioperl/bioperl-live/commit/0e70e88a76d638d9f0406643c37691bb20d60ceb
>> Author: Florent Angly <florent.angly at gmail.com>
>> Date:   2010-06-08 (Tue, 08 Jun 2010)
>> 
>> Merge branch 'master' of github.com:bioperl/bioperl-live
> 
> I'm fascinated by these commits, and my git-fu is still weak.
> 
> I think what's happening here is that any time anyone commits anything fangly is merging those changes into his own repo, which he then merges back to bioperl/bioperl-live again. 

Yes, via git pull.

> So fangly's procedure (whatever it is), is re-committing other people's commits? Making history twice as long with (empty?) "Merge branch 'master'" messages? The diff of these commits reports that fangly is the author of other people's changes(!) yet somehow git annotate still reports that t/data/ZABJ4EA7014.CH878695.1.blast.txt was authored by Razi Khaja yesterday (correct).
> 
> Am I reading that correctly? I find that history very confusing.
> 
> In #moose they taught me to merge other people's commits using the procedure below. This is what I did yesterday to merge rkhaja/bioperl-live into bioperl/bioperl-live (per conversations in IRC). 
> 
>   git remote add rkhaja git://github.com/rkhaja/bioperl-live.git
>   git fetch rkhaja
>   git checkout -b rkhaja-merge rkhaja/master
>   git rebase master
>   git checkout master
>   git merge rkhaja-merge
>   git branch -d rkhaja-merge
> 
> That procedure did not create a "Merge branch 'master'" commit. So is that procedure cleaner than fangly's? Is it the rebase command that makes the difference?

This is probably the best way, sure, to avoid those additional 'Merge branch' messages.  The rebase is the key part, which is where one is just setting the current branch pointer to point to the same place as master HEAD.  

So, what would be the workflow for a simple commit like florent's?  Or one from a topic branch?  Maybe something to add to the 'Using git' page?

> I'm not picking on fangly here, I'm simply struggling to improve my own git-fu. 
> 
> Thanks,
> 
> Jay Hannah
> seeker of git enlightenment
> http://biodoc.ist.unomaha.edu/wiki/User:Jhannah

I think we all are :>

chris




More information about the Bioperl-l mailing list