Re: Support for QNX6, POSIX IPC and PTHREAD-style locking

"Igor Kovalenko" <Igor(dot)Kovalenko(at)motorola(dot)com> writes:
> The trouble is, there seems to be NO rules. 'Too late' does not qualify as
> rule, IMHO. Not as formal one anyway and informal ones are hard to follow
> precisely. There should be formal mechanism for acceptance of patches,

No, we don't have formal rules, and it's unlikely that you'll convince
many people in this project that we should. The one formal rule that
we do adhere to is "no new features during beta". I'm not sure if we've
ever really decided whether a new port is a new feature, but I'd lean
to the view that it is.

I realize that it seemed that you were facing increasingly higher
demands; in fact you were, because the closer we get to release the
less inclined we are to apply non-essential patches. It was strictly
a schedule thing. I'm still not sure that you've quite absorbed the
point that we're trying to have a release candidate out *this week*.
Two weeks ago there was more flexibility, but we're down to the wire
now and only important bug fixes are going in.

In retrospect we should probably have told you to start with that there
was no point in trying to get QNX6 support into 7.2. A localized patch
(only adding port/qnx6/ directory and contents, no changes to shared
files) perhaps would have been accepted two weeks ago, but by the time
you had refined your submission down to that we were a lot closer to
release than we were at the start of the discussion.

The fault for the miscommunication is probably mine, in that I didn't
make it clear to you in the beginning that time was of the essence.
I would like to apologize for that and move on. Let's focus on how we
can make 7.3 better, rather than worrying about what might or might not
have gotten into 7.2 if the calendar had been a little different.

regards, tom lane