How did you find that refresh 'leaks' value to dynamicbackcolor? It might happen that dynamicbackcolor expression contains a syntax error (extra bracket, datatype mismatch, etc.) This case you will really get weird error messages in run-time (no check during compiling) during form refresh, but it should be found and fixed in dynamicbackcolor code. Look carefully what you have there. If it's a complex expression then remove it, firstly, and then start to add it step by step.

>I don't understand either, I am not attempting to retrieve the dynamicbackcolor value anywhere. I am letting the grid do what it does. The dynamicbackcolor set in the grid as as you say (RGB(nnn,nnn,nnn) returning integer). This value is what is "leaked" to the return value of the form's refresh().
>
>Charles Richard
>
>>>Hi all
>>>
>>>I have a problem with dynamicbackcolor in a grid: When I define one, I get a datatype mismatch in the return value of the baseclass of my form refresh. FYI I normally use the following structure in my methods:
>>>
>>>LOCAL llRetval
>>>llRetval = DODEFAULT()
>>>IF llRetval
>>> *My code here
>>>ENDIF
>>>RETURN llRetval
>>>
>>>In this instance, the value llRetval receives from DODEFAULT() happens to be the color code in the dynamicbackcolor. If I don't use dynamicbackcolor, all is ok. I find this wierd! I have debugged and the value returned in the parent method is a boolean (.T.) just before control is returned to the child method. Stepping once to return to the child method and the value returned is the number.
>>>
>>>So two questions:
>>>
>>>1) WHY!!!!!!!!!!!!!!????
>>>2) Do I really need to validate the boolean return value in a form refresh? What is the advantage?
>>>
>>>Thanks in advance
>>>Charles Richard
>>
>>I don't quite follow. Dynamicbackcolor, if retrieved by some method, should be Character data type expression returning Integer.