Sorry about the delay. I haven't got the hang of this UT thang yet :-(

More info?
Well, there's a main application object which is the main framework, and has lots of data hung off it too.

The data refers to a locking mechanism that is also hung off the framework (as a method of it, of course)

The last "CLEAR ALL" gets upset in the "destroy"s of the data hanging off the framework when they try to refer to the locking mechanism, claiming that the ojbect doesn't exist.
It *may* be when these objects are in an array of objects (I think we've had problems with releasing arrays of object pointers before) but I'm not sure if that's a pertinent point.

Enough extra?

Regards
Mark