Re: planner crash/assert hit in 8.4B1 - Mailing list pgsql-hackers

From Tom Lane
Subject Re: planner crash/assert hit in 8.4B1
Date
Msg-id 13604.1240154778@sss.pgh.pa.us
Whole thread Raw
In response to Re: planner crash/assert hit in 8.4B1  (Bernd Helmle <mailings@oopsware.de>)
List pgsql-hackers
Bernd Helmle <mailings@oopsware.de> writes:
> --On Sonntag, April 19, 2009 15:29:01 +0200 Stefan Kaltenbrunner 
> <stefan@kaltenbrunner.cc> wrote:
>> Program received signal SIGSEGV, Segmentation fault.

> Please note that it doesn't raise the Assert, it crashes because rel is 
> NULL, hence the assertion expression crashes with a SIGSEGV. There seems to 
> be something wrong with the RelOptInfo in estimate_num_groups().

It looks like I neglected to test this code while writing the
PlaceHolderVar patch --- it needs to handle those specially, and fails
to do so :-(.  Thanks for the report!
        regards, tom lane


pgsql-hackers by date:

Previous
From: Brendan Jurd
Date:
Subject: Re: to_timestamp() changes in 8.4 release notes
Next
From: Peter Eisentraut
Date:
Subject: Re: Unicode support