Yes, it's because it sees the intermediate
state (6,1) when it shouldn't (it processes
each foreign key separately and does the cascade
update which places it in an "invalid" state
since the 1 key no longer exists).
I believe the fixes I've been working on for deferred
constraints will make this case work as well.
On Wed, 18 Jul 2001 pgsql-bugs(at)postgresql(dot)org wrote:
> Bob Soeters (bob(at)iway(dot)nl) reports a bug with a severity of 2
> The lower the number the more severe it is.
>
> Short Description Two foreign keys in one table both referencing same
> record in primary table gives error on update of primary table