On 5/07/09 17:00, Luke Kanies wrote:
> On Jul 5, 2009, at 3:40 AM, Brice Figureau wrote:
> 
>> On 5/07/09 4:39, James Turnbull wrote:
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
>>>
>>> Luke Kanies wrote:
>>>>> Yes, it was my understanding that #2033 was a feature for 0.26 (I
>>>>> don't
>>>>> feel it would be right to add new feature while we're in beta).
>>>>> It just happened  that I was cut off from the Internet a couple of
>>>>> weeks
>>>>> ago, so that was the perfect time to work on a bigger thing not
>>>>> involving redmine :-)
>>>> Makes sense.
>>>>
>>>> I may end up getting more features in 0.25, also, if clients  
>>>> change my
>>>> priorities, which looks like it's possible.
>>>>
>>> At some point we need to call "enough".  We're up to b2. I don't
>>> want to end up like Dovecot with b15.  I think one (or at the most
>>> two more) betas would be a reasonable limit.
>> Yes, I completely agree. We should focus on getting to 0 high/ 
>> important
>> bugs. What we can do if we absolutely want to add new features or only
>> track them for later is to create a 0.26 branch now and commit those
>> there. Or do the reverse and branch 0.25 now and keep 0.26 for the
>> development. Choose your poison of course.
> 
> First, I don't think we should have any more betas.  We should focus  
> on getting tickets closed and push for an rc1, rather than more betas,  
> if at all possible.

Agreed, 0.25 seems to be in good shape. As soon as the more important 
bugs are fixed (I'm still worried by 2296), then I think we could have a 
rc1.

> As to later releases, I really think we need to switch to something  
> like the kernel uses, where the 'next' branch is recreated  
> periodically from the code that is accepted for release.  This puts a  
> higher burden on James as the release maintainer (although not a ton,  
> since it could pretty easily be turned into a straightforward rake  
> task), but it makes it trivial to add and remove code chunks from the  
> branch.
> 
> E.g., the daemontools code, which had broken tests for months, would  
> have been added initially, but then once the tests were found to be  
> broken, it would have been removed from the list of code to merge and  
> the branch would have been recreated.  Boom, code gone, and no extra  
> commits or anything.

Yeah, code Darwinism, I like that :-)
-- 
Brice Figureau
My Blog: http://www.masterzen.fr/


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Puppet Developers" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/puppet-dev?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to