Jim,

We agree to disagree as far as the RELEASE command is concerned.

I feel that you are emulating COM's handling of RELEASE commands within VFP, and it is my opinion that we should not be doing so EXCEPT when we explicitly code the management routine(s) to go with that practise. In which case there would only be one actual RELEASE for any object.

Regards,

jim N

>>>>As regards breaking existing code I have two things to say:
>>>1) The VFP team didn't appear too worried about this when they changed the ERROR return information between V3 and V5. Or they felt they had no choice. Could end up the same here. Sh** happens sometimes.
>Jim,
>
>I totally disagree with your obervations regarding the RELEASE command. The RELEASE command releases a memory variable (as well as other things). It is not supposed to release an object unless the variable is the last reference to the object. In OO one should not be concerned with what other objects are doing when writing the code in an object. If I decide to release a variable in one object's method it should have no effect whatsoever on any other variable in any other object. Depending on the fact that it has no effect on other variables is not doing something wrong, it is depending on the designed behavior of the release command to only affect the variables explicitly listed on its command line.
>
>The release method is another story since it is the object itself that is doing the releasing. However, I don't have any problems with dangling references. Probably because the objects clean up after themselves and there aren't a load of references floating around.