Re: WIN1252 encoding - backend or not? - Mailing list pgsql-hackers

Tom Lane wrote:
> Bruce Momjian <pgman@candle.pha.pa.us> writes:
> > Marc G. Fournier wrote:
> >> On Sat, 4 Dec 2004, Bruce Momjian wrote:
> >>> OK, so what do we want the process to be?
> >> 
> >> Basically, during Beta/Release, we should almost a policy where a third 
> >> party patch needs to be approved by a second committer *before* being 
> >> applied ... and that even applies to Tom >:)  Your own patch, fine ... but 
> >> a third party patch, even submitted by someone who has submitted patch 
> >> previously, should be reviewed/approved by two committers ...
> 
> > Please find a cure that isn't worse than the disease.  I don't have time
> > to apply patches as it is, let alone check with someone else.
> 
> That's a fair objection, but if it means that the default is that
> patches don't get applied during late beta/RC, I'm not sure I'm unhappy
> with that default.

The actual affect would be to make beta last longer because in fairness
to patch appliers you have to give each patch a reasonable chance of
being applied.  While it is temping to say, "No one was around to review
your patch so we couldnt' apply it", it isn't going to be received very
well by the submitters.  This was the issue we had in getting 8.0 to
beta, if you remember.

> In the particular case of this patch, although Bruce said that others
> had already commented on the patch, the only comments I see in the
> pgpatches archives said that the patch was unreviewable because it

Alvaro mentioned a incorrect comment:
http://archives.postgresql.org/pgsql-patches/2004-11/msg00344.php

That's the only correction I remember, aside from the diff vs. new file
discussion.

> wasn't offered as a diff.  I think it would be reasonable to insist on
> at least one concurrence ("looks ok to me") posted to pgsql-patches
> before applying during late beta.  We've gotten into a mode where
> if you like a patch you say nothing, but I wonder whether we shouldn't
> change that habit.

If I get a "looks good to me" email I am much more likely to apply the
patch promptly, that's for sure.  Lacking that, if I have seen someone
comment on the patch and a new version was generated, I assume it is OK.
Now, if I don't understand the patch, we can change the procedure so I
require someone to state it is OK rather than the fallback of quiet
acceptance, especially just before a beta or RC version.

Also, because the bad patch got in there is a temptation to believe our
process is flawed, but that assumes that a perfect procedure exists.  A
perfect procedure doesn't exist because the world isn't perfect.  We can
make adjustments, but never expect any process to be fool-proof.

--  Bruce Momjian                        |  http://candle.pha.pa.us pgman@candle.pha.pa.us               |  (610)
359-1001+  If your life is a hard drive,     |  13 Roberts Road +  Christ can be your backup.        |  Newtown Square,
Pennsylvania19073