Re: GiST insert algorithm rewrite

On 09.01.2011 07:05, Tom Lane wrote:
> I just found out that the "benchmark" test script in
> contrib/intarray/bench/ crashes HEAD in gistdoinsert() --- it looks like
> it's trying to pop to a stack entry that isn't there.
>
> Run it per the instructions in the intarray documentation:
>
> $ createdb TEST
> $ psql TEST< ../_int.sql
> ...
> $ ./create_test.pl | psql TEST
> CREATE TABLE
> CREATE TABLE
> CREATE INDEX
> CREATE INDEX
> server closed the connection unexpectedly
> This probably means the server terminated abnormally
> before or while processing the request.
> connection to server was lost
>
> The script generates randomized data, so possibly it won't fail every
> time, but it failed three out of three times for me. The changes I'm
> about to commit in intarray don't seem to make any difference.

Thanks, fixed. Apparently my testing never hit the case where an update,
rather than an insert, into the root page causes it to split.

--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com