virtually anything I find in the October 2001 MSDN Library on this subject is marked in red with "[This is preliminary documentation and subject to change]".

On Sunday I described a test I concocted to learn how (HRESULT) returns are handled by this feature. The result of that test strongly suggested that the returned (HRESULT) value was random - probably the first or the last subscriber executed has its code returned.

I have also been through both the publisher properties and the subscriber properties using the "Component Services" console and can find no attribute that might be useful in governing the handling and/or interpretation of subscriber return codes.

So my question is this:

Where and how can I submit a ER to have either:
1) Any subscriber "FAIL" HRESULT be THE VALUE passed back to the publisher; or
2) Some mechanism to control this in a more flexible and controlled manner.

I feel real stupid asking for a ER for something that I KNOW so little about, but I see no other way.

My motivation is a simple one: While this is a new feature made available through VFP 7, I see one of its major potential benefits being the ability to develop pieces of applications (and implement them) much more independently, but this all (seems to) goes out the window if there cannot be a RELIABLE way for the publisher(s) to learn of any problem on the subscriber side of things.
I wouldn't want to implement important stuff in subscriber components under the (apparent) current conditions, but if I could be sure of feedback of failure then there would be no impediment.

Thanks

Jim Nelson