[MOAB-dev] MOAB git branching model

[Tim Tautges]

On 09/19/2013 01:16 PM, Jed Brown wrote:
> Tim Tautges <tautges at mcs.anl.gov> writes:
>> My concern with the petsc workflow is about branch merges to next
>> taking awhile to graduate to master.  However, looking over the repo
>> structure, it seems like that happens quickly enough in petsc.
>
> It takes as long or as short as needed.  There is a tradeoff between
> fast graduation (easy for new development to confidently build on it)
> versus stability (only well-tested features make it to master).  Strike
> your balance wherever it suits you.

Yeah.  I'm more worried about not doing it until just before an official release, which has its own problems.

>
>> For releases, do they just get tagged on master, or is there a
>> separate release branch?  If the former, do point releases get inserted
>> into master as well?
>
> Feature releases are tagged on 'master', at which point 'maint' is
> fast-forwarded.  Maintenance releases (bug fixes; no API changes) are
> tagged on 'maint'.  Preview releases or RCs would be tagged on 'master',
> though we don't currently use those.
>

I was asking more about numbered releases, not feature releases.

- tim

>> And, does auto testing happen on next?  I'd assume so (and on master too, of course?).
>
> Yes, most auto-testing is directed at 'next', with some on 'master' to
> confirm that we didn't mess up.  You can also have CI run on branches
> before merging to 'next', but our current testing infrastructure doesn't
> handle that case well at the moment.
>

-- 
================================================================
"You will keep in perfect peace him whose mind is
   steadfast, because he trusts in you."               Isaiah 26:3

              Tim Tautges            Argonne National Laboratory
          (tautges at mcs.anl.gov)      (telecommuting from UW-Madison)
  phone (gvoice): (608) 354-1459      1500 Engineering Dr.
             fax: (608) 263-4499      Madison, WI 53706