Hi Paul,

Someone else pointed out that your present logic does NOT test the success of the FLOCK. They suggested some alternate code and it looked much more thorough to me.

As regards SET REFRESH, I agree that the documentation says screen updates, but there could have been a connection with data updating too. The VFP documentation isn't known for its thoroughness < s >. For instance, the last line of the CURRENT documentation for SET REFRESH says: "SET REFRESH can also be used to specify how often data buffered locally on your workstation is updated" despite stating up until that point that it applies ONLY to BROWSE windows. Here, again, it *could* be applicable both ways (though apparently it is not, based on your trials).

Still a mystery as far as I'm concerned. My best *guess* at this point is that the FLOCK logic **MAY** be implicated, but who knows!?!?!

Cheers,

JimN

>We tried several different settings for refresh, and none of them fixed the problem. We didn't expect it to because SET REFRESH should only affect screen updates. What seemed really strange is that we were locking the file, and unlocking it which is supposed to issue a flush. We also tried closing and re-opening the table. None of these things worked. One of the responses in this thread indicated that an explicit FLUSH command might function a little bit differently form the implicit one with an UNLOCK. That is the only explanation that makes sense to me.
>
>
>>Paul,
>>
>>Just wondering what your value for SET REFRESH is in the application.
>>
>>If it's zero (the default) I suspect that that could be the problem too.
>>
>>According to the Help for "FLUSH" your UNLOCK should have done the equivalent (possibly this is where REFRESH makes a difference???).
>>
>>Cheers,
>>
>>JimN
>>