[Biojava-dev] Cloning repositories in eclipse

[Andy Yates]

Hi,

I've also been doing Git based migration investigations recently and I feel it's appropriate to share some of my observations which I hope can help you to come to your own conclusions about how to progress

1). Git developers seem to avoid pre-allocating branches. This seems sensibles as you only want to branch when there's a good reason to and then they attempt to close the branch as fast as possible (that's not saying long-term branches are not liked just that they will suffer from drift)

2). Small projects use master like CVS's HEAD i.e. all small commits are done directly there (sometimes even big ones) and only substantial changes necessitate a branch. Bigger projects sometimes split the development branch from master & have master acting as somewhere with *just* the live release (GitFlow promotes this) and acts as a release integration point.

3). When you need to do commits to previous a releases then go back to the release tag, create the branch, patch & push. But that said the chances of you requiring to do this are low.

There are a few other things about keeping master as clean as possible since it helps with merges but it isn't essential. Git can be made more complicated if you want to follow a very specific flow which keeps master as clean as possible or it can be made to act just like subversion (without as many branches). 

Hope this helps,

Andy

On 11 Apr 2013, at 02:54, Andreas Prlic <andreas at sdsc.edu> wrote:

>> I don't get why we have or need the release branch. Other projects hosted
>> on Github don't do it like that, and in my day job, we use maven and Git
>> with the release plugin without such a branch too.
>> 
> 
> 
> So the release plugin then tags the code, but does not create a branch? (in
> SVN tags and branches are essentially the same thing)  We could have one
> branch per release, but it sounds like this is not necessary. Also, looking
> back at past releases I can't remember that there was ever a commit on a
> release branch.
> 
> Andreas
> 
> 
> 
>> 
>> It makes pull requests a bit complicated too, as the web UI defaults to
>> the release branch. I had to force push something to my forked release
>> branch because I accidentally added a file to the wrong branch at first.
>> 
>> The single release branch also won't allow fix releases for older
>> releases, or am I wrong?
>> 
>> Hannes
>> 
>> Am Donnerstag, 11. April 2013 schrieb Andreas Prlic :
>> 
>> Same here, I got my eclipse setup to work with egit, but I thought that it
>>> was not as easy as I would have hoped.   Let's try to keep the entry
>>> barrier as low as possible.
>>> 
>>> This leaves the question: should we rename the devel or release branch to
>>> master? I'd say devel, since by default this would give you the latest
>>> code
>>> (similar to trunk checkouts in svn)
>>> 
>>> Andreas
>>> 
>>> 
>>> 
>>> On Wed, Apr 10, 2013 at 4:08 PM, Steve Darnell <darnells at dnastar.com>
>>> wrote:
>>> 
>>>> Eclipse user here. +1 for making BioJava easier to setup in Eclipse.
>>>> 
>>>> ~Steve
>>>> 
>>>> -----Original Message-----
>>>> From: biojava-dev-bounces at lists.open-bio.org [mailto:
>>>> biojava-dev-bounces at lists.open-bio.org] On Behalf Of Spencer Bliven
>>>> Sent: Wednesday, April 10, 2013 5:57 PM
>>>> To: biojava-dev
>>>> Subject: [Biojava-dev] Cloning repositories in eclipse
>>>> 
>>>> When I first drafted the instructions for cloning biojava in eclipse<
>>>> http://biojava.org/wiki/BioJava3_eclipse>,
>>>> I was frustrated that the simple 'Checkout Maven projects from SCM'
>>>> doesn't work with our layout. The alternate checkout method is about 4
>>>> times as long and much less intuitive.
>>>> 
>>>> I had some time to debug the problem today, and it seems that the
>>> e2m-egit
>>>> connector doesn't handle branches. It fails with a misleading message on
>>>> any repository without a master branch. I've submitted a bug report, but
>>>> I'm not optimistic about it getting patched quickly.
>>>> 
>>>> What proportion of BioJava users are using eclipse? Would fixing this be
>>>> worth renaming one of the branches to master?
>>>> 
>>>> -Spencer
>>>> _______________________________________________
>>>> biojava-dev mailing list
>>>> biojava-dev at lists.open-bio.org
>>>> http://lists.open-bio.org/mailman/listinfo/biojava-dev
>>>> 
>>>> _______________________________________________
>>>> biojava-dev mailing list
>>>> biojava-dev at lists.open-bio.org
>>>> http://lists.open-bio.org/mailman/listinfo/biojava-dev
>>>> 
>>> 
>>> 
>>> 
>>> --
>>> -----------------------------------------------------------------------
>>> Dr. Andreas Prlic
>>> Senior Scientist, RCSB PDB Protein Data Bank
>>> University of California, San Diego
>>> (+1) 858.246.0526
>>> -----------------------------------------------------------------------
>>> _______________________________________________
>>> biojava-dev mailing list
>>> biojava-dev at lists.open-bio.org
>>> http://lists.open-bio.org/mailman/listinfo/biojava-dev
>>> 
>> 
> 
> 
> -- 
> -----------------------------------------------------------------------
> Dr. Andreas Prlic
> Senior Scientist, RCSB PDB Protein Data Bank
> University of California, San Diego
> (+1) 858.246.0526
> -----------------------------------------------------------------------
> _______________________________________________
> biojava-dev mailing list
> biojava-dev at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/biojava-dev