On 10/12/2005 11:20 PM, Tom Lane wrote:
> I wrote:
>> Bruce Momjian <pgman@candle.pha.pa.us> writes:
>>> It isn't so much the alphabetical order, since there is only one
>>> trigger, but the concept that we now group all the _before_ triggers
>>> before the _after_ triggers.
>
>> But we've always done that. Has the example ever been correct?
>> I was intending to try it on older versions, but I don't actually
>> think it's ever acted like the docs said.
>
> After digging in the CVS archives, I find that it did work like that
> up till this 7.0 patch:
>
> 1999-09-29 12:05 wieck
>
> This is part #1 for of the DEFERRED CONSTRAINT TRIGGER support.
> Implements the CREATE CONSTRAINT TRIGGER and SET CONSTRAINTS
> commands.
>
> So the example was probably correct when put in, but no one's noticed it
> was wrong since 7.0 :-(
IIRC we had discussed that stuff during the RI development and decided
to have ALL _after_ triggers get fired by the deferred queue after the
statement. The example deletes multiple rows in one statement.
Jan
--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#================================================== JanWieck@Yahoo.com #