Tom Lane wrote:
> Fran Fabrizio <ffabrizio@Exchange.WebMD.net> writes:
> >> Kinda hard to believe.
>
> > I know it, but that's what I am seeing..... (to recap, my trigger only
> > fires on the first insert per connection.)
>
> > [snip]
>
> > ******* Here is the server log this time around *******
>
> > 2001-05-14 12:02:54 DEBUG: StartTransactionCommand
> > 2001-05-14 12:02:54 DEBUG: query: insert into log values
> > (0,0,lower(''),lower(''),lower(''),lower('hostname.changed.com'),upper('Med'),
> > upper('Apps'), upper('Medssld'), upper('INFO'), '1 medssld(s) running.',
> > '20010513 00:00:00');
> > 2001-05-14 12:02:54 DEBUG: ProcessQuery
> > 2001-05-14 12:02:54 DEBUG: CommitTransactionCommand
>
> > ****** Note, NO TRIGGER firing! So I disconnect and reconnect.... *****
>
> This does NOT show that no trigger fired. What it shows is that no
> queries within the trigger needed to be recompiled. The log entries
> you are looking at come from the parser when a query is compiled, not
> from plpgsql execution.
>
> Try adding some "raise NOTICE" or "raise DEBUG" messages to your trigger
> procedure to see what's really happening.
>
> (Yes, I know that plpgsql could do with some better tracing/debugging
> facilities than it's got. Any volunteers out there?)
Hands off please. I have some drastic changes to PL/pgSQL
outstanding that'll implement *real* cursor support. I'll
look at tracing while at it. Stay tuned...
Jan
--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#================================================== JanWieck@Yahoo.com #
_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com