* Daniel Farina:
> The idea of canceling a COMMIT statement causing a COMMIT seems pretty> strange to me.
Canceling commits is inherently racy, so I'm not sure if this behaviorso strange after all.