Re: branching for 9.2devel - Mailing list pgsql-hackers

From Tom Lane
Subject Re: branching for 9.2devel
Date
Msg-id 2439.1303861189@sss.pgh.pa.us
Whole thread Raw
In response to Re: branching for 9.2devel  (Josh Berkus <josh@agliodbs.com>)
Responses Re: branching for 9.2devel
List pgsql-hackers
Josh Berkus <josh@agliodbs.com> writes:
>> Huh?  We've never guaranteed anyone a regular annual cycle, and we've
>> never had one.  We agreed to use the same schedule for 9.1 as for 9.0;
>> I don't remember anything more than that being discussed anywhere,
>> ever.

> We *want* to have a regular annual cycle which doesn't vary by more than
> a few weeks.

There may be some people who want that, but it's not project policy
and I don't think it will ever become so.  Our policy is "we release
when it's ready".  To allow the development schedule to become purely
calendar-driven would mean a drastic decline in our quality standards.

I suppose we could have something like a predetermined branch-from-devel
date for each major release, with the time from branch to actual release
varying depending on release stabilization progress, while new
development proceeds forward on a regular commitfest clock.  But I fail
to see any significant advantage from doing that.  What it would mostly
do is decouple the development community entirely from release
stabilization work, and I think that would be a seriously bad idea.
Not only from the take-responsibility-for-your-work angle, but because
diverting manpower from release stabilization will also mean that it's
that much longer from feature freeze (or whatever you call the branch
event) to actual release.  I don't think that people will be that happy
about knowing "if I finish this by date X, it will be in release N" if
they have no idea when release N will reach production status.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Alignment padding bytes in arrays vs the planner
Next
From: "Greg Sabino Mullane"
Date:
Subject: Re: Alignment padding bytes in arrays vs the planner