Re: RLS feature has been committed

On Tue, Sep 23, 2014 at 11:23 AM, Dave Page <dpage(at)pgadmin(dot)org> wrote:
> On Tue, Sep 23, 2014 at 4:19 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>> On Tue, Sep 23, 2014 at 11:16 AM, Dave Page <dpage(at)pgadmin(dot)org> wrote:
>>> Regardless of what Robert may feel, review should only generally be
>>> *expected* during a commitfest, but it can be done at any time.
>>> Committers are free to commit at any time. The process was never
>>> intended to restrict what committers do or when - in fact when I
>>> introduced the process to -hackers, it was specifically worded to say
>>> that developers are strongly encouraged to take part in the commitfest
>>> reviews, but not forced to, and may continue to work on their patches
>>> as they see fit.
>>
>> So, just to be clear, you're saying that committers are free to commit
>> things even when other community members (who may themselves be
>> committers) ask them not to?
>
> At what point did I say anything like that? Do not twist my words.

Well, if that's not what you're saying, then good, because I sure as
heck don't think that. I think it's the role of a committer to commit
things that are generally agreed to be good ideas, not things that
that particular committer personally thinks are a good idea regardless
of the opinions of others. The committer is entitled to weigh their
own opinion more heavily than the opinions of others because, hey,
that's why they have the commit bit and the other people don't. But
they're not entitled to run roughshod over contrary opinions.

Now, the way that CommitFests come into it is that people can't do two
things at once. It's completely reasonable, in my opinion, for me to
say, hey look, I really want to review your patch but I don't have
time to do that right now because we're in the middle of a CommitFest;
please therefore submit it to the next CommitFest. And if I say that,
the person should either (1) do it or (2) disagree with the necessity
of doing it not (3) commit anyway.

Whether you've been paying attention or not, most committer patches
ARE submitted to CommitFest and go through the exact same review
process as patches from non-committers. That's a good thing. It
contributes to our delivery of quality software, and it reduces
arguments about whether something was rushed through. There are
appropriate times to leapfrog that process, when it's clear that the
patch is uncontroversial and that delay is merely dithering. But this
is not one of those times. It's a giant patch implementing a complex
feature for which more review time was explicitly requested.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company