[RFD] Merge proposals and code reviews too private

[Martin Pool]

2009/8/20 Matthew D. Fuller <fullermd at over-yonder.net>:
>> I don't see how asking people to submit their code to a code review
>> system is inconsistent with code review being separate from the
>> general discussion list.
>
> The point is we've been asking people to submit their code to a code
> review system long before the point where their message is "Please
> consider this for merging", way back at the point where their message
> is "Hey, I made this $THING, let's talk about whether we want it
> {,like this}".  Reviews by all appearances fits the former well, and
> having it not overlap the list may make sense.  Having the latter
> divorced from the dev list where the same question is raised for
> things that DON'T have code written, though, is uncomfortable to say
> the least.  I don't think our current setup serves that case very well
> at all.

I think the distinction is between a conceptual review (rfc on the
list) and a code review (on the list).

In general if there's any chance the concept would be controversial
I'd encourage people to post an rfc to the list as soon as they become
aware of this - typically/hopefully before starting code, but possibly
while they're writing it or even during code review, if it turns out
the concept is more controversial than expected and you want a larger
audience to review the concepts.  We should do more of these.

I think doing code review before the code is actually ready is also
fine, if you want to see if the basic code is ok perhaps before you've
updated everything or before you've fixed all the tests.

I don't think there's any point doing a merge proposal before you have any code.

There are at least a few non-active/core developers subscribed to all
reviews.  I'm not sure how much they actually read them.

-- 
Martin <http://launchpad.net/~mbp/>