Message-ID: <3F3A49BE.8BA38DAA@rodos.fzk.de>
Date: Wed, 13 Aug 2003 16:22:54 +0200
From: Christoph Haller <ch@rodos.fzk.de>
MIME-Version: 1.0
To: pgsql-sql@postgresql.org, surrano@mailbox.hu
Subject: Re: Order of triggers - totally lost
Content-Type: text/plain; charset=iso-8859-2
Content-Transfer-Encoding: 7bit

> The docs say in trigger-datachanges.html:
>
> "Changes made by query Q are visible by queries that are started after
query
> Q, no matter whether they are started inside Q (during the execution
of Q)
> or after Q is done.
>
> This is true for triggers as well ..."
>
> Trying to understand that, I have a feeling that the update in A_BIUD
should
> already see the results of the preceding insert, including the results
of
> triggers activated by that insert.

I've seen that
PostgreSQL Weekly News - July 25th 2003
        Even though we are nearing release, it doesn't hurt to look
ahead, as
evidenced by several additions to the TODO list this week: delay
resolution of array expression type to assignment coercion can be
performed on empty array expressions, have AFTER triggers execute after
the appropriate SQL statement in a function, not at the end of the
function, more details on possible tablespace definitions were added,
add PL/PHP, allow PL/pgSQL to name columns by ordinal position (e.g.
rec.(3)), and prevent COMMENT ON DATABASE from using a database name.
Some of these already have folks working on them, however if you are
interested in lending a hand please don't be afraid to speak up.

Looks like this passage about AFTER triggers is what you're expecting,
but is not implemented yet.
>
> What may be wrong? Any ideas to re-organize parts of the triggers? May

> putting the update to an A_AU trigger help? I tried it, still have
problems
> (not sure it's still the trigger order), but the trigger order is
still
> strange for me:
>
No ideas, sorry. Maybe move over to the [HACKERS] list for more
professional support.

Regards, Christoph