Bruce Momjian <maillist@candle.pha.pa.us> writes:
> Yep. Wouldn't the best way be to have the temp system record the
> transaction id used, and to invalidate all temp entries associated with
> an aborted transaction. That is how the cache code works, so it seems
> it should be extended to the temp code.
Yeah, that would work -- add an xact abort cleanup routine that goes
through the temprel list and removes entries added during the current
transaction.
AFAICS this only explains the coredump-at-exit business, though.
I'm particularly baffled by that
ERROR: cannot find attribute 1 of relation pg_temp.24335.3
in my last example --- do you understand why that's happening?
regards, tom lane