>>But when I execute the seek I get "Uniqueness of index violated" or something to thst effect. After some research, I discovered that I should chamge the buffering for the violation table from optimistic record buffering to optomistic table buffering, because when the record pointer moves for the seek, foxpro attempts to save the record when "optimistic record buffering" is used. The problem is that I now get "Uncommitted changes in table 'violation'".
>
>the VFP generated RI code has a problem with compund indexes
>I was told ths was not fixed with 5.0 as well
>
>Arnon
has anyone tried xcase w/ 5.0?