I wrote:
> I've been chasing Chris Bitmead's coredump report from earlier today.
> I find that it can be reproduced very easily. For example:
> regression=> select f1 from int4_tbl group by f1;
> < no problem >
> regression=> select f1 from int4_tbl* group by f1;
> < core dump >
We had tentatively agreed not to fix this for 6.5, but I got more
worried about it when I noticed that this particular simple case
worked in 6.4.2. I don't like regressing ... so I dug into it and
have just committed a fix.
It turns out that pretty much *anything* involving grouping or
aggregation would fail if the query used inheritance, because the
necessary preprocessing wasn't getting done in that case. I think
that's a big enough bug to justify fixing at this late date. (Besides,
the fixes do not change the non-inheritance case, so I don't think
I could have broken anything that was working...)
regards, tom lane