Gervase Markham wrote:
> This is a "think piece" :-)
>
> This is abridged from this week's Linux Weekly News
> (http://lwn.net). Do we need something like this? Having three
> patches waiting for review and another three for super-review, how
> can we lower the barrier to entry for contribution while
> maintaining code quality?
>
I agree with this, but we need this suggestion posted somewhere a bit
more prominent as many people don't read the spam fest that is .general
perhaps posting the idea to Mozillazine.
I've seen a lot of patches lie in Mozilla which are low risk and are
only chaging something minor but aren't checked in because they're
waiting for review. I'm sure that this will put some people off the idea
of contributing patches.
There needs to be some level of common sense here. The people who are
given permission to check into the CVS should have enough sense to know
whether the changes are cosmetic and have no chance of causing
regressions and should be able to check in these without requiring
review. The review process is a good idea but requiring everything to be
reviewed seems to be just slowing everything down.
A set of rules should be made to roughly decide what needs review and
what doesn't. Personally I believe the following should be true:
* changes to whitespaces and comments (i.e. no risk changes to the
codebase) shouldn't require review and can be checked in by someone with
appropriate CVS access.
* Low risk fixes, that are unlikely to cause regressions should require
one review although the reviewer should pass it to a super review if
they have any doubts.
* Other fixes should go through the normal two stage review processes.
