> Can you please shed some light as to why circular exception references do not constitute a bug within the context of the user project?

To be honest it's been so long that I forget the circumstances where we ran into it. But I'm pretty sure that it didn't happen in the context of some dumb programmer mistake. It was more likely some very subtle circular dependence, or perhaps even an intentional/valid one.

Blaming the user doesn't seem like the right answer. That's like saying that we should worry about X-Loop headers to prevent email forwarding cycles, or that garbage collection shouldn't have to deal with cycles in data structures. Accidents happen, and sometimes cycles are valid.