Re: pgsql: Add sql_drop event for event triggers

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>
Cc: pgsql-hackers(at)postgresql(dot)org, pgsql-committers(at)postgresql(dot)org
Subject: Re: pgsql: Add sql_drop event for event triggers
Date: 2013-03-29 23:22:24
Message-ID: 6676.1364599344@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-committers pgsql-hackers

Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> writes:
> Add sql_drop event for event triggers

The buildfarm members that use -DCLOBBER_CACHE_ALWAYS don't like this
patch:

***************
*** 760,771 ****
FROM generate_series(1, 50) a;
BEGIN;
SET TRANSACTION ISOLATION LEVEL SERIALIZABLE;
UPDATE serializable_update_tab SET description = 'no no', id = 1 WHERE id = 1;
COMMIT;
SELECT description FROM serializable_update_tab WHERE id = 1;
! description
! --------------------
! updated in trigger
(1 row)

DROP TABLE serializable_update_tab;
--- 760,773 ----
FROM generate_series(1, 50) a;
BEGIN;
SET TRANSACTION ISOLATION LEVEL SERIALIZABLE;
+ ERROR: SET TRANSACTION ISOLATION LEVEL must be called before any query
UPDATE serializable_update_tab SET description = 'no no', id = 1 WHERE id = 1;
+ ERROR: current transaction is aborted, commands ignored until end of transaction block
COMMIT;
SELECT description FROM serializable_update_tab WHERE id = 1;
! description
! -------------
! new
(1 row)

DROP TABLE serializable_update_tab;

I suspect you have inserted a snapshot-capturing operation into
someplace it mustn't go during transaction startup, but only in a path
that is triggered by an immediately preceding cache flush.

regards, tom lane

In response to

Responses

Browse pgsql-committers by date

  From Date Subject
Next Message Andrew Dunstan 2013-03-30 00:04:36 Re: [COMMITTERS] pgsql: Add parallel pg_dump option.
Previous Message Tom Lane 2013-03-29 23:03:05 Re: Fix for pg_upgrade and invalid indexes

Browse pgsql-hackers by date

  From Date Subject
Next Message Andrew Dunstan 2013-03-30 00:04:36 Re: [COMMITTERS] pgsql: Add parallel pg_dump option.
Previous Message Jeff Davis 2013-03-29 23:20:44 Re: Hash Join cost estimates