Bruce Momjian <maillist@candle.pha.pa.us> writes:
> I have fixed the problem I introduced with aggregates. They should work
> now, and the HAVING regression test should work too.
Also, I put in Vadim's recommended fix for the subplan problem.
The regression tests look a lot better than they did. The "union"
test is still failing by adding a bunch of
NOTICE: equal: don't know whether nodes of type 600 are equal
lines to the expected output. Anybody know what's causing that?
regards, tom lane