Andrew Dunstan <andrew(at)dunslane(dot)net> writes:
> Something committed in about the last 24 hours or so has caused my
> Windows buildfarm members to hang completely while running the
> regression tests. It's happened for both MinGW and MSVC builds.
baiji and dawn_bat are both green now, so apparently Stefan was correct
to guess that the array indexing bug in psql/print.c was the problem.
I don't understand, however, how that manifested as a hang and not
a core dump. Does Windows not treat bogus pointer dereference the
same as everyone else does?
regards, tom lane