Alex Peshkov wrote:

>
> I see no problems doing all possible verifications.

Me neither. We surely cannot track all possible logical changes but, if
possible, we should validate the formats a bit better than just
comparing lengths and data types.


Dmitry