Re: [Boost-users] [MSM] Unable to retrieve correct value of attribute of a state from within actions. (Possible Bug?)

Subject: Re: [Boost-users] [MSM] Unable to retrieve correct value of attribute of a state from within actions. (Possible Bug?)
From: Christophe Henry (christophe.j.henry_at_[hidden])
Date: 2013-02-26 15:16:26

Hi Deniz,

> Just for final clarification three more questions:
>
> 1.
> You mean, that from within the "operator()" of an action I am only allowed 
> to use the parameters for accessing states, events, and the FSM internals? 
> (Neither "event_", "source_", "target_", "fsm_", nor accessing the 
> dummy-instances directly.)

I didn't say this, just the dummy states you declare for the table.
If you had this declared:

BOOST_MSM_EUML_METHOD(ActivateEmpty_ , activate_empty , activate_empty_ , 
void , void )

You could call inside a state action something like:

activate_empty_(target_)(evt,fsm,*this);

or inside an action:

activate_empty_(target_)(evt,fsm,src,target);

But the interest is quite limited. They are made for the transition table. 
But as you see, msm creates some valid functors.

> This is now clear for the current event, current source- and target state 
> as well as for all other states of the FSM (by accessing them with 
> "get_state").
>
> 2.
> However, how to access other events than the current one?
>
> It seems to work as I did before (by accessing the "dummy"-instance of the 
> event), but if you do not recommend it, maybe there is a better solution?
> (I need to access all events directly to initialize their event-number 
> attributes from within the FSM's entry-action.)

Well, you could do this for events. Though as you seem to have a number for 
every type, a static member could be easier.
The normal way is usually in the event constructor but I agree that for 
eUML, it is not so practical, so yes, your trick will do.

> 3.
> And all the underscore-accessors "event_", "source_", "target_", "fsm_" 
> are only allowed to be used from within the transition table?

They are allowed everywhere but their main reason of existence is to make a 
nicer transition table.

>
> Maybe you could emphasize the real usage in the documentation, so that 
> nobody misuses it (like I did).

Ok, will do. Thanks.

Cheers,
Christophe