Stable Loggerhead Branch

[Martin Pool]

On 2 July 2010 13:33, Max Kanat-Alexander <mkanat at bugzilla.org> wrote:
>> This is a bit hypothetical because nobody has actually put up a
>> non-stability patch targeted to 1.18 but if someone did, it would be
>> the right thing to review and merge it so as to unblock them.  If
>> their patch would put extra work on you to stabilize it then you
>> should get them to do that.  You could cajole another bzr dev into
>> doing the review.
>
>        Well, okay, but I can't know in advance if a patch is going to make it
> unstable. I think the purpose of a stable branch is to *not* add
> features, and to *not* refactor, and just add bug fixes. So I wouldn't
> want to accept any new feature or anything that even looks large on the
> 1.x branch.

I agree.  Basically what I meant is that you should manage the
released branches pretty much as we do in bzr which is that clearly
safe bug fixes can come in but nothing else.

I think for any patch someone puts up (fix, feature, stability,
performance, whatever) there at any time there should be _a_ possible
path  to it landing, whether that is into a release branch or trunk.
If the particular implementation looks unsafe then it's ok to ask it
be made more safe or tested etc.  But I wouldn't like to see a state
where new features with sane implementations are not allowed to land
on either 1.18 or trunk.

-- 
Martin