Open Stack

On 09/08/2014 11:07 AM, Alexis Lee wrote:
> Sean Dague said on Mon, Sep 08, 2014 at 09:22:56AM -0400:
>>> On 09/08/2014 05:17 AM, Steven Hardy wrote:
>>>> I think this may be a sensible move, but only if it's used primarily to
>>>> land the less complex/risky patches more quickly.
>>
>> 2 +2 has been part of OpenStack culture for a long time, and there is a
>> good reason for it, it really does keep bugs out.
>>
>> It should also be clear that the subject of this email really should
>> have been "merging code faster", because nothing in here doubles the
>> review bandwidth, it just provides us with less review coverage.
> 
> For these reasons, I'm also wary of changing this in general.
> 
> Sometimes I yell in IRC if I +2 something important, this shortens loop
> time as hopefully I already understand the patch and can answer
> questions.
> 
> Single-approvals for small + simple changes could be worth trying.
> Perhaps also for large + simple changes like whitespace fixes.

Realistically core teams are already doing that. Being considered an
exception people typically provide a good reason in the review as to why
it's a fast approve, which is fine.

The large cross cutting whitespace fixes are actually completely
problematic in their own way, because they typically force a non trivial
rebase on 10s - 100s of patches in flight, so they generate *a ton* of
work for other people.

	-Sean

-- 
Sean Dague
http://dague.net