> -----Original Message-----
> From: George Pavlov [mailto:gpavlov@mynewplace.com]
> Sent: 16 October 2006 19:09
> To: Dave Page
> Subject: RE: pgAdmin beta 3 query bug (got a reproducible case!)
>
> ok, after a lot of messing around i think i have it down to
> one totally
> reproducible case (may not be the only case of a crash, but
> this one is
> 100% crasha-ble for me). basically when you have an error message in
> between two queries that differ in their result set shape you get a
> crash. here are the steps:
>
> 1, from pgadmin connect to any DB and open the query tool
> 2. paste the following queries into the query pane
>
> select * from information_schema.columns;
> select * from my_bogus_table_name; -- this table does not
> exist, just a
> way to cause an error message
> select 1 from information_schema.columns; -- could be any query that
> produces a result set different from the first one
>
> 3. execute these three in sequence by highlighting and hitting f5 on
> each one
> - the first one returns a result set as expected
> - the second one produces an error message as expected
> - the last one produces a crash
>
> the above may seem like a contrived scenario, but i narrowed it down
> from some actual queries i had.
Oh, well spotted. I've committed some code to fix this and hopefully
make it easier to catch similar issues in the future. Do you want an
updated .exe?
This also sounds like it might have been an issue that Harald saw - he
was running some experimental queries with occasional errors iirc.
Regards, Dave.