Re: Completing PL support for Event Triggers - Mailing list pgsql-hackers

From Peter Eisentraut
Subject Re: Completing PL support for Event Triggers
Date
Msg-id 1380673807.22785.15.camel@vanquo.pezone.net
Whole thread Raw
In response to Completing PL support for Event Triggers  (Dimitri Fontaine <dimitri@2ndQuadrant.fr>)
List pgsql-hackers
On Fri, 2013-09-13 at 22:40 +0200, Dimitri Fontaine wrote:
> Please find attached to this email three patches covering the missing PL
> support for Event Triggers: pltcl, plperl and plpython.

For plperl, the previous reviews mostly apply analogously.  In addition,
I have these specific points:

- In plperl_event_trigger_build_args(), the hv_ksplit() call is probably
unnecessary.

- plperl_call_perl_event_trigger_func and plperl_call_perl_trigger_func
as well as plperl_event_trigger_handler and plperl_trigger_handler have
a lot of overlap could perhaps be combined or refactored differently.

- In plperl_event_trigger_handler(), a point is being made about setting
up current_call_data before SPI_connect.  Other handler functions don't
do this, though.  It's not quite clear to me on first readong why it's
done differently here.

- Like I pointed out for the pltcl patch, calling trigger functions as
event triggers and vice versa is not detected in
compile_plperl_function, but I guess this can only happen if you
manually butcher the function after you have set up the trigger.




pgsql-hackers by date:

Previous
From: Peter Geoghegan
Date:
Subject: Re: pg_stat_statements: calls under-estimation propagation
Next
From: Peter Eisentraut
Date:
Subject: Re: [PATCH] Revive line type