Excerpts from Matthijs Bomhoff's message of mié jun 29 15:20:49 -0400 2011:
> Possibly related: In some earlier attempts at reproducing this, I
> actually tried to use length() instead of md5(). It seemed that I
> could not get it to trigger with that. Just selecting * triggers it of
> course, but caused a bit too much clutter in my psql for the required
> long random strings, so I had to find a function that would actually
> need the value itself. Hence the md5().
Yeah, the length can be obtained from the toast header AFAIR, so you
don't need to do a full detoast of the value, which masks the problem.
Álvaro Herrera <alvherre(at)commandprompt(dot)com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
In response to
pgsql-bugs by date
|Next:||From: Tom Lane||Date: 2011-06-29 21:52:22|
|Subject: Re: BUG #6082: server segfault - pg_stat_reset_shared(null) |
|Previous:||From: Matthijs Bomhoff||Date: 2011-06-29 19:20:49|
|Subject: Re: Bug with plpgsql, temp tables and TOAST?|