incubator-cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Prasanna Santhanam <...@apache.org>
Subject Re: [DISCUSS][INFRA] Setting up gerrit
Date Fri, 15 Feb 2013 05:19:07 GMT
On Fri, Feb 15, 2013 at 03:37:43AM +0530, Chiradeep Vittal wrote:
> 
> 
> On 2/14/13 1:43 AM, "Prasanna Santhanam" <tsp@apache.org> wrote:
> 
> >
> >Personally I'm okay with either tool. IMO the real issue is in people
> >reviewing a patch within time for a release. On the review I've cited
> >we've got a lot of things to learn from -
> >
> >a) timely response from blueprint (FS) owners and/or asking for time
> >for a thorough review
> >b) sign-off from multiple committers
> >c) incremental patch sets
> >
> 
> I do feel that 'enough' committers problem highlighted by Pranav is a real
> problem with the Gerrit workflow. But if the contributor asks for multiple
> reviews then Gerrit is better than RB. In RB it is not clear when the
> patch is ready for commit since it is not clear which of the previous
> comments have been addressed.

True - for core components with single owners it might be hard but
even a smoke test pass can be considered a review. Contributors
can (and should?) review code, so their review counts as well. And
that's been happening on RB which is a good sign. On the OS lists
(*hates taking the same example*) committers are also chosen based on
good code review. You don't have to push code necessarily. If there is
a component you've been aiming to contribute to, one can review
patches coming in that area. 

In summary while it may be true that gerrit outweighs the features
when it comes to managing patch workflow with git - we will together
as a community have to promote that good culture for stronger reviews.

As Ahmad would put it - 'Nuff said, Let's do it!

-- 
Prasanna.,

Mime
View raw message