>>>As Aleksey explained in Re: Error 2066 (undocumented) Index File xxx is corrupted... Thread #810798 Message #811607 it may not be even index corruption. He also gave some suggestions on how to avoid this error.
>>>Tell you truth, I belive that it's a bug in VFP because VFP should handle this situation internaly not generate an error. I had to move one of my applications back to VFP7 because of this error.
>>
>>Interesting, this is why I was wondering if that is fixed in VFP 8 SP1. We'll have to wait a few weeks before knowing more about it.
>
>No it's not fixed in SP1 because MS doesn't consider it a bug.

I have seen one particular instance where this error looks like a false positive. In the case I have seen, you can trap for error 2066, execute SYS(1104), and then RETRY, and the command or function that triggered the error works without error the 2nd time. If you implement this, you would need to configure your error handler to only retry a few times on that error, or you could get stuck in an endless loop should you encounter a case where the error was truly due to index corruption.

The workaround I mentioned above should be safe, because the same index testing is performed each time the command or function that caused the error is retried, it's not like we are ignoring the error. We're just giving the command another chance to succeed after flushing the data buffer.

I consider the case above to be a bug, but I'm not the final arbiter on that. Unfortunately, this came in quite recently and we did not have a dependable repro for it until after we had released SP1.