Re: pg_group problem

"Nigel J. Andrews" <nandrews@investsystems.co.uk> writes:
>> Is this something to do with the recent changes to array_in/out perhaps?

Yup, it is.  The special-purpose code for manipulating group membership
lists (which are just int4[]) hadn't got updated.  Fix committed ---
thanks for the report!

> One other thing, there appears to be no regression test for this situation

I think we have regression tests for users but not any for groups.  Feel
free to contribute one ...

            regards, tom lane

PS: bugs in CVS tip are probably more profitably discussed on -hackers
than on -general.