[Cosmo-dev] Bug fixes and unit tests

[Ted Leung]

On Sep 28, 2006, at 4:12 PM, Brian Moseley wrote:

> On 9/28/06, Ted Leung <twl at osafoundation.org> wrote:
>
>> Also,  when Chandler goes into feature freeze, all checkins must be
>> code-reviewed before they can be committed -- patches containing the
>> fixes are attached to bugzilla bugs and the bugzilla mechanisms are
>> used to review the patches.    I think that this is a very helpful
>> process and wanted to know what people thought about adopting this  
>> idea.
>
> i've never been a fan of this practice. it's too heavyweight. i for
> one review every svn commit message, and i squawk when i see something
> i don't understand or think could be done better. our team is still
> small enough that i think this practice will keep enough eyeballs on
> the code without going through the tedious bugzilla review process.

I think that there are some situations where RTC (review then commit)  
makes sense, but it's also a cultural thing.   If the Cosmo project  
is actively practicing CTR (commit then review) -- and actually doing  
the reviews, then I think that's fine.    I just figured that as long  
as we were talking about process that we might as well see if people  
wanted to do RTC.   I'm not particularly vested in doing it as long  
as commits are reviewed in a timely way.

Ted