Re: BUG #5835: PL/Python crashes - Mailing list pgsql-bugs

From Alex Hunsaker
Subject Re: BUG #5835: PL/Python crashes
Date
Msg-id AANLkTi=6subm+p_zaStpLqJH7O0KGVyngjk0x59=P8hY@mail.gmail.com
Whole thread Raw
In response to BUG #5835: PL/Python crashes  ("Julien Demoor" <jdemoor@gmail.com>)
Responses Re: BUG #5835: PL/Python crashes  ("Julien Demoor" <jdemoor@gmail.com>)
Re: BUG #5835: PL/Python crashes  (Alvaro Herrera <alvherre@commandprompt.com>)
List pgsql-bugs
On Thu, Jan 13, 2011 at 06:57, Julien Demoor <jdemoor@gmail.com> wrote:
>
> The following bug has been logged online:
>
> Bug reference:      5835
> Logged by:          Julien Demoor
> Email address:      jdemoor@gmail.com
> PostgreSQL version: 9.0.2
> Operating system:   Ubuntu Linux
> Description:        PL/Python crashes
> Details:

> BEGIN;
> CREATE OR REPLACE FUNCTION test_none_3(a numeric[]) RETURNS numeric(12,4) AS
> $$
>    return 1;
> $$ LANGUAGE plpython2u;
> SELECT test_none_3('{NULL, NULL, NULL}'::numeric[]) FROM
> generate_series(1,244) ORDER BY generate_series DESC LIMIT 1;
> ROLLBACK;

It looks like we are not incrementing a reference count for NULL
entries in an array.  The attached fixes it for me.  I did look for
other places where we might have missed this and did not find any.  (
the PyDict_SetItemString() that use Py_None look fishy, but apparently
work ).

Can you see if it fixes it for you?

Attachment

pgsql-bugs by date:

Previous
From: Craig Ringer
Date:
Subject: Re: BUG #5829: No buffer Space Available(maximum connection reached?) from postgres edb driver
Next
From: "Tyler Kellen"
Date:
Subject: BUG #5838: tables of the same name in different schemas