> > The basic question is whether this is the proper way to do temp
> > tables.
>
> I haven't looked at the patches, but fwiw I would have tried it about
> the same way. No need to touch pg_class if the info is
> session-specific...
Yes, my feeling is that the code is complicated enough without having
the temp table stuff adding complexity. What I did is that a cache
lookup returns a fake pg_class tuple. The only code changes are a few
function calls in the cache routines to insert my fake tuples, and some
code in the heap_create_with_catalog/heap_create/heap_destroy code to
create temp tables with unique names. A new istemp flag in a few
structuers. The rest of the code is untouched.
-- Bruce Momjian | http://www.op.net/~candle maillist@candle.pha.pa.us | (610)
853-3000+ If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill,
Pennsylvania19026
