>Hi all,
>
>we think we have found an interesting error in the VFP RI-code. Maybe many of you have found this before, but we thought we'd share anyway.
>As VFP goes through the RI-code, it tries to reuse tables it has opened before. To do that, it builds and reads from a string (pcRiCursors) that holds the used tables and their workareas. It has the form:
>
>devconpeople*12345people*12346
>
>Now, when a table has to be used, it looks for that table in this string the following way:
>
>atc(tcTable+"*",pcRIcursors)
>
>This goes wrong when you look for "people" in the string above.
>The solution is simple of course. Still, i think this could account for some 'vague' faulty transactions (that WE have had anyway).
>
>regards,
>derk schutte

It's widely known that the VFP RI code is not very good and has problems Steve Sawyer has written a replacement that is very good. You get it when you buy his book from www.hentzenwerke.com