<yazicivo(at)ttnet(dot)net(dot)tr> writes:
> I've been planning to implement a caching mechanism for non-volatile
> (and non-SRF) procedures in PL/scheme, that'll return pre-computed
> values in case of a procedure signature overlap[1].
This has generally been considered a bad idea even if implemented for
all function types. It seems even less sane to implement it on a retail
PL-by-PL basis.
> And after reading backend/utils/mmgr/README, I think
> TopTransactionContext is the way for me to go.
You realize that that goes away at the end of each transaction?
regards, tom lane