I wrote: > Another point worth making is that this version of the patch deletes the > tuple tables during AtEOSubXact_SPI(), earlier in cleanup than would > happen with the prior version. That increases the risk that external > code might try to delete an already-deleted tuple table, if it tries > to call SPI_freetuptable() during subxact cleanup. The new code won't > crash, although come to think of it it will probably throw an error > because you're not connected anymore. (Maybe this is a reason to not > insist on being connected, but just silently search whatever the top > stack context is?)
After further reflection I think that's the prudent way to do it, so I've adjusted SPI_freetuptable to not insist on being connected. Pushed with that change: http://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=3d13623d75d3206c8f009353415043a191ebab39 regards, tom lane -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers
