On 25.05.2011 20:11, Tom Lane wrote:
> Heikki Linnakangas<heikki.linnakangas@enterprisedb.com> writes:
>> On 25.05.2011 17:47, Tom Lane wrote:
>>> [ scratches head ... ] Why does the save/restore in ExecEvalCase not
>>> take care of this?
>
>> The mistake happens during planning, when the SQL function is inlined
>> and pre-evaluated.
>
> Hm. I'm inclined to think this may be more of a bug in the inlining
> process than anything else.
Well, if you want to get away without the capability to reference
multiple CaseTestExprs at a time, you'll have to detect the danger and
abort the inlining process. That's a bit pessimal, although this is a
pretty artificial case in the first place so maybe we don't care much.
(I'm still going to need more placeholder slots to handle IN and
BETWEEN. Of course, I can just copy-paste CaseTestExpr into something
like InTestExpr and BetweenTestExpr, but it seems like it would be good
to unite all that infrastructure)
-- Heikki Linnakangas EnterpriseDB http://www.enterprisedb.com