Re: 6.4.2 patch for shared-memory hashtable bugs

> Here is a patch for version 6.4.2 that corrects several serious bugs
> in Postgres' shared-memory-hashtable code. These problems are fairly
> harmless until you get to more than 256 shared objects of a given type
> (locks, buffers, etc) ... but then things get nasty. For more info see
> the discussion on the pgsql-hackers list in late Feb. 99 (thread title
> "Anyone understand shared-memory space usage?"). The equivalent
> changes are already in the 6.5 source code, but not in 6.4.*.

Quick question: when you do a query with a join why does the hash table
code need to use shared memory? Can't it do the join within its own memory
space?

[I remember seeing a post recently with someone talking about how
currently the code uses static sized buffers and so thats why we get the
hash table out of memory errors but I didn't quite follow what was going
on]

thanks,
Wayne

------------------------------------------------------------------------------
Wayne Piekarski Tel: (08) 8221 5221
Research & Development Manager Fax: (08) 8221 5220
SE Network Access Pty Ltd Mob: 0407 395 889
222 Grote Street Email: wayne(at)senet(dot)com(dot)au
Adelaide SA 5000 WWW: http://www.senet.com.au