>>>I still think that VFP, regardless of version, should start with a clean slate.
>>>
>>
>>The problem is a C..5 error is an abrupt termination of the running process - it leaves DLLs and Windows objects in an unstable state since all execution related to VFP is abruptly stopped. Without restarting, the shared components and objects will not be reset.
>
>
>Since these are shared objects, it must be a possibility that other tasks would be using them. Do the objects appear unstable to these other tasks?
>
Yes, since they can never clear the semaphore for the VFP context or release resources allocated in the VFP process context.
>Would it be possible for VFP to get a fresh invocation of these objects and leave behind any left over problems from earlier sessions?
>
If you like committing suicide, sure. If there are any things given to VFP and it dies unnaturally, other things start to hang forever.