hadoop-pig-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Doug Cutting <cutt...@apache.org>
Subject Re: guideline on branching
Date Tue, 27 Nov 2007 20:24:03 GMT
Olga Natkovich wrote:
> Because for major changes it can break existing functionality and you
> don't want to keep things in your local copy till everything works
> because you don't get any backup for it.

Nigel said, "attach them to the Jira issue".  That's the backup.

I've found that merging is simple when, instead of branching, folks 
regularly post revisions of trunk patches to Jira as they develop things 
and as trunk changes.  But for a large destabilizing changes that 
involve many developers, a branch may be warranted.

In my years at Apache I've only used a development branch once: to port 
Nutch's core algorithms to the MapReduce platform.  We didn't want to 
break the trunk, and knew it would take a while to complete this port. 
The branch lasted several months.

So I don't yet see a case for single-developer branches, but have found 
that branches can be useful for long-term, multi-developer modifications.


View raw message