"Han Holl" <han(dot)holl(at)prismant(dot)nl> writes:
> So it looks like using the pure SQL version here is not a good idea.
> I'm quite willing to produce an example of this behaviour, but I
> suspect that it's 'known behaviour' to experts.
No, not to me anyway. I would not be surprised if there's some leak in
the SQL-function code, but could I trouble you for a complete example,
rather than having to reverse-engineer one?
regards, tom lane