[openstack-dev] When should things get added to Oslo.Incubator

Jay Pipes jaypipes at gmail.com
Tue Dec 3 22:41:13 UTC 2013


On 12/03/2013 05:30 PM, Mark McLoughlin wrote:
> On Wed, 2013-12-04 at 10:42 +1300, Robert Collins wrote:
>> Hey - https://review.openstack.org/#/c/57022/7//COMMIT_MSG - I
>> strongly suggested here that reusing the Nova object code is the first
>> step towards an objects library, and that we should be putting it in
>> olso; there are some reasonable concerns about this being experimental
>> but...
>>
>> The Oslo wiki page says:
>> The process of developing a new Oslo API usually begins by taking code
>> which is common to some OpenStack projects and moving it into the
>> oslo-incubator repository. New APIs live in the incubator until they
>> have matured to meet the criteria described above.
>>
>> So if this code lands in solum as-is, its now common - to me that
>> means we should be going straight to oslo.incubator and working on
>> shared code.
>>
>> However there's clearly an expectation mismatch - so - any thoughts?
>
> Speaking to Devananda about the copy of the code in Ironic, he made the
> case that Ironic was just going to sync the code from Nova without
> making modifications to it until the code was ready to graduate as a
> standalone library or move to oslo-incubator.
>
> That actually does make some sense - I've often described oslo-incubator
> as "managed copy and paste". It still sucks, but not as much as the copy
> and paste we've done previously where the multiple copies of the same
> code evolved independently.
>
> Now, in the case of this Solum import, it does look like this is a fork
> of the Nova code. Unless there's some realistic plan to reconcile these
> forks, then I think it is a problem and quite short-sighted.

Not a problem that cannot be resolved, and quickly :) Heck, we are about 
as early in the process as you can possibly be! Robert is right to bring 
this important topic up now in the early code reviews.

So, there was some discussion with Clayton Coleman and Dan Smith around 
reasons for diverging [1]. Some of the reasons were due to stuff that 
Dan himself noted in code reviews existed for backwards compatibility or 
hack reasons in Nova and that did not apply to Solum. Some of the 
reasons were around development/design patterns that folks thought were 
too magical. In either case, you and Robert are probably right that it 
would be good to plop this code into oslo incubator and iterate on it 
there, instead of in one of the projects...

Best,
-jay

[1] 
http://lists.openstack.org/pipermail/openstack-dev/2013-November/018480.html




More information about the OpenStack-dev mailing list