Neil Conway <neilc(at)samurai(dot)com> writes:
> BTW, both of our fixes suffer from the deficiency that they will
> actually reject input one variable too early: we disallow a SQL
> statement with 1023 variables that we strictly speaking could store.
Right. I thought about putting the overflow checks inside the switches
so that they wouldn't trigger on the case where we don't need another
variable ... but it doesn't seem worth the extra code to me either.
regards, tom lane