[boost] Re: [fsm] Why exit actions must not fail

From: Andreas Huber (ah2003_at_[hidden])
Date: 2004-05-29 03:29:56

Robert Bell wrote:
> F2 defines termination as "making a transition to the final state".
> It's
> far from clear that destroying a state machine should make a
> transition
> to the final state. Specifically, if the machine is in a state which
> has
> no transition to the final state, what does it mean to terminate the
> machine? If the machine is in such a state when it is destroyed, does
> it
> make sense to pretend it can transition to the final state when no
> such transition exists?

You can see this as another assumption related to A1...

> I would think that when a state machine object is destroyed, no
> actions (exit or otherwise) should be executed; rather, it should
> just tear down
> the state machine and release its resources (state objects,
> transitions, whatever). What am I missing?

You're not missing anything. If you disagree with the assumptions I make
then I will not be able to convince you. The only hope I have is that you
will see that my assumptions do make sense *in practice*.

Regards,

Andreas