Andres Freund <andres(at)anarazel(dot)de> writes:> A reasonable rule would actually be to only use [u]int32 and> sig_atomic_t, never bool.
Yes, I'd agree with that.
regards, tom lane