Josef Gschwendtner wrote:

>
> Now the problem arises:
> =======================
> We have been able to add a new field "xx" in table T2 without any
> problem (I think adding new field "xx" in T2 should fail with an
> appropriate message).

No, it shouldn't fail. The compiled BLR of your procedure explicitly
references T1.xx, so there's no ambiguity for the engine.

> That there is a problem with this database we found out doing a
> backup/restore of this database.
> The restore failed because of some weird problems.

What _exactly_ problems?


Dmitry