Richard Guo <guofenglinux@gmail.com> writes: > I'm reproducing this bug in HEAD with the repro given by Martijn, but > with some additional GUC sets:
> set enable_hashjoin to off; > set enable_mergejoin to off; > set enable_indexscan to off; > set enable_bitmapscan to off;
Hm, does not work for me ... although I assume you mean bc76f5ac4 or before? It should definitely not happen after 1aa8dad41, unless there's an additional bug ...
Yeah, I tested in bc76f5ac4. Attached is a regression test I composed. I tested it locally and it can catch this bug before 1aa8dad41 and give the expected answer after 1aa8dad41.
