John Naylor <jcnaylor(at)gmail(dot)com> writes:
> I changed the signature of a freespace.c function that the BRIN code
> calls, and this change by itself doesn't cause a crash. With the full
> FSM patch, causing BRIN errors in manual queries in psql doesn't cause
> a crash. However, during the BRIN regression tests, the queries that
> purposely cause errors result in FATAL instead, causing a crash.
Sounds like something's messing up the backend's exception stack,
so that elog.c has noplace to throw the error to. See the
promote-ERROR-to-FATAL logic therein.
regards, tom lane