Thanks for bringing this to our attention. I have passed it up the line for you.

For me, this causes VFP to terminate back to VFP3.

Thanks again!

>Hi,
>
>If you want put VFP on memvar inconsistent status,
>run this, and IGNORE errors.
>
>

>PUBLIC X2
>X2=CREATEOBJECT('pippo')
>RELEASE X2
>* now X2 have on internal array memvars name and pointer defined, but heap storage is released, and internal vartype char is bad.
>
>* x3 for compare error trpping
>? TYPE('M.X3')		&& OK, U
>? VARTYPE(M.X3)		&& OK, U
>? VARTYPE(M.X3,.T.)	&& BUG, Variable not found
>=m.X3  			&& OK,  Variable not found
>? m.X3			&& OK,  Variable not found
>
>* if you use x2:
>
>? TYPE('M.X2')		&& ??? U
>? VARTYPE(M.X2)		&& ??? Data Type Mismatch
>? VARTYPE(M.X2,1)	&& ??? Data Type Mismatch
>=m.X2  			&& ??? Data Type Mismatch
>? m.X2 			&& Print fire a strong C0005 error.
>
>DEFINE CLASS pippo as Custom
> PROCEDURE init
>   RETURN
> ENDPROC
>
> PROCEDURE Destroy
>   * now X2 not exists
>	PUBLIC X2
>   * now X2 is correct
>        ? 'var x2 = ', m.X2
> ENDPROC
>
>ENDDEFINE
>

>
>then VFP RELEASE memvar is not correct
>
>Fabio