Try the following in the destroy method of the form:
THISFORM.SetAll("DynamicBackColor","","Column")
THISFORM.SetAll("DynamicForeColor","","Column")
THISFORM.SetAll("DynamicCurrentControl","","Column")
THISFORM.SetAll("DynamicFontName","","Column")
...

So everything is reset before the form starts to unload
the objects which may sometimes be used by dynamic expressions
due to a VFP6/SP3 bug.

Please let me know if it helps.

Andreas



>"Expression is invalid. Use a valid expression for DYNAMICCURRENTCONTROL". Not shouting, just typing message as it appears. I have been running the VFP app since 1997 ( release 5.0 ), using several of the 'dynamic' values for grid columns. In the past few months ( can't say for sure that it has only been since SP3 for 6 ), of the approximately 10,000 seats using the program, a couple dozen will get the message 1 to 2 times per week, typically near the end of a 'session' using the grid -- saving / copying the contents.
>
>The users who do get the problem are fairly state-of-the-art systems 400+ megahertz Intel Pentium III class, with 128+ megs of RAM, linked to networks using 100 base T connections, but some are single users with local data base. I have encountered this message many times while developing, and fully realize the necessity for proper expressions. I have even written a PRG that looks for grids, and analyzes the validity of the various 'dynamic' values -- spits out error message if expression results in invalid data type or cannot be evaluated externally ( contains 'this' or 'thisform' or some other such object reference ).
>
>What I am looking for is more information about trapping the grid / column / expression involved in the error. There does not seem to be an error number, or any further help in locating the problem. That is what I am looking for, not an analysis of the 30+ dynamic expressions in the grid, but how to know which one went bad. As you can tell from the feed back I have gotten, the error is highly random, occurs infrequently, but as more and more users get faster and more capable workstations, I see it occuring more frequently. I have manually and programatically analyzed each of the expressions until I am "blue in the face" ( lack of oxygen to the brain ? )
>
>Any help / suggestions for trapping the message? Since the message does not trigger the VFP error routine, I am somewhat at a loss for further data gathering on the problem.