>I find a solution.
>I need to do a VFP C++ reverse engineering, and crack my runtime VFP8*.dll.
>
>But this is one unacceptable effort in order to correct this behaviour.
>
>I want VFP9 resolve it.
>
>If VFP9 not resolve it, i change the VFP9*.dll code.

Or, gee, radical concept: maybe, just maybe, you could handle it the way everyone else has been suggesting?

Or is that too simple?