>We hit this problem last year inside several scans that host many function calls etc... and the only way we beat was re-selecting the table at the end of the scan. Even if not an internal VFP thing ( which we felt it was ) it is a surefire way of not having that problem bite you unexpectedly, just like the _tally thing. We are now fighting with relations that are being lost unexpectedly in a process and having to reset them. Version 6 never did that to us now 7 is, but since we cannot duplicate it with any regularity I guess you could say that isn't a VFP thing either.
>
>MLW

I've almost always found that when VFP is doing something strange like that, the usual cause is something stupid that I've done. Rarely has it ever been VFP's fault. That was the case in a bunch of the C...5 type errors, it was usually a failure to do proper "housekeeping" on my part, even though it didn't crash consistantly or repeatably, it was still my fault, not VFP's.