shindig-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Franklin, Matthew B." <mfrank...@mitre.org>
Subject RE: RTC vs CTR was ( Review Request: Allow container implementations to more easily override and extend rpc registered service handlers. )
Date Wed, 25 Jul 2012 21:05:25 GMT
>-----Original Message-----
>From: Henry Saputra [mailto:henry.saputra@gmail.com]
>Sent: Wednesday, July 25, 2012 4:15 PM
>To: dev@shindig.apache.org
>Subject: Re: RTC vs CTR was ( Review Request: Allow container
>implementations to more easily override and extend rpc registered service
>handlers. )
>
>I am thinking about having Apache Shindig bylaws similar to what
>Apache Hadoop has: http://hadoop.apache.org/bylaws.html which govern
>how code commits should be conducted.

+1, though I would use a different community's bylaws as an example [1].  Their definition
of Lazy consensus is a little off to me.   Ross Gardler wrote Rave's and it covers the concept
well[2].

[1] http://hc.apache.org/bylaws.html  (note the section on #Code_Review)
[2] http://rave.apache.org/docs/governance/lazyConsensus.html

>
>I'd like the simplicity of CTR but it needs to have good boundaries. I
>really dont want us to come back to the old model where commits and
>reviews just done with some people working in the same companies.
>Reviews could be done early with some people but at the end should
>targeted to dev list for final approval.
>
>- Henry
>
>On Wed, Jul 25, 2012 at 12:07 PM, Franklin, Matthew B.
><mfranklin@mitre.org> wrote:
>> Most communities I have seen eventually adopt a Commit Then Review
>model over a Review Then Commit model.  Due to the complexity of Shindig, I
>can understand wanting to make sure that bigger changes are reviewed;
>however, for trivial changes such as this, would it be easier to just commit the
>change?
>>
>> I am not a committer, so it is really up to you all.  IMO, it is a lot of overhead
>to review everything :) .  If you do move to a CTR model, I would suggest
>setting some boundaries so that you work into the model.  Maybe saying that
>any change with x lines, etc.
>>
>>>-----Original Message-----
>>>From: Dan Dumont [mailto:noreply@reviews.apache.org] On Behalf Of Dan
>>>Dumont
>>>Sent: Wednesday, July 25, 2012 2:28 PM
>>>To: shindig; Dan Dumont
>>>Subject: Review Request: Allow container implementations to more easily
>>>override and extend rpc registered service handlers.
>>>
>>>
>>>-----------------------------------------------------------
>>>This is an automatically generated e-mail. To reply, visit:
>>>https://reviews.apache.org/r/6141/
>>>-----------------------------------------------------------
>>>
>>>Review request for shindig.
>>>
>>>
>>>Description
>>>-------
>>>
>>>Change rpc registration to return the old handler if there were any so that
>>>container implementations may call into the previously registered handler if
>>>they wish to extend the existing behavior.
>>>
>>>
>>>This addresses bug SHINDIG-1827.
>>>    https://issues.apache.org/jira/browse/SHINDIG-1827
>>>
>>>
>>>Diffs
>>>-----
>>>
>>>
>>>http://svn.apache.org/repos/asf/shindig/trunk/features/src/main/javascri
>pt/
>>>features/container/container.js 1365569
>>>
>>>http://svn.apache.org/repos/asf/shindig/trunk/features/src/main/javascri
>pt/
>>>features/rpc/rpc.js 1365569
>>>
>>>Diff: https://reviews.apache.org/r/6141/diff/
>>>
>>>
>>>Testing
>>>-------
>>>
>>>Tests pass.
>>>
>>>
>>>Thanks,
>>>
>>>Dan Dumont
>>
Mime
View raw message