From: | "Marko Kreen" <markokr(at)gmail(dot)com> |
---|---|
To: | "Jonathan Gray" <jon5pg(at)streamy(dot)com> |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Pl/pgsql functions causing crashes in 8.2.2 |
Date: | 2007-02-06 10:14:37 |
Message-ID: | e51f66da0702060214m1c3b0a5n47de1fbfa6c21a19@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On 2/6/07, Jonathan Gray <jon5pg(at)streamy(dot)com> wrote:
> Following an upgrade to 8.2.2, many of my plpgsql functions started to cause
> server process crashes.
>
> I make use of a custom data-type "uniqueidentifier", available here:
>
> http://gborg.postgresql.org/project/uniqueidentifier
> ftp://gborg.postgresql.org/pub/uniqueidentifier/stable/uniqueidentifier-0.2.
> tar.gz
>
> This type has given me the same kind of process crash once before, but that
> was related to NULL values in a foreign-key referenced field (unresolved to
> this day, but behavior is allowed for all builtin types).
Indeed, the code can crash on NULL values as the NULL checks
are missing or wrong in the functions. Actually all the various
functions except newid() should be declared STRICT IMMUTABLE
thus immidiately avoiding problems with NULLs.
Could you reproduce the crash with this change? I'll try
to play with this myself too.
--
marko
From | Date | Subject | |
---|---|---|---|
Next Message | Simon Riggs | 2007-02-06 10:14:45 | Re: Proposed adjustments in MaxTupleSizeandtoastthresholds |
Previous Message | Simon Riggs | 2007-02-06 10:05:53 | Re: Proposed adjustments in MaxTupleSize andtoastthresholds |