Hi Christophe humm...... but In our design We actually do thinks in the constructor of the back end. and We use multiple inheritance as well where derived is inherited from msm:backend and some other base classes to couple MSM with our existing message dispatching FW and call process_event. Is it not allowed to use derived classes as backends ?. I was not a ware of it. Is it mentioned in the documentation ? It gives me a lot headache to change our design now ... :(. we have spent now about 2 man month altogether to implement base classes around msm now if this is a case I have to change all of it. is there a way to allow such a implementation ?. Br. Richie On 19 July 2010 11:28, Christophe Henry <christophe.j.henry@googlemail.com> wrote:
Hi All
I have run in to the following problem : if I have 3 state machines where 1st has a composite state which has one more composite state than the most inner one enters to a pseudo exit state which sends finish event. The middle one has a transaction from the most inner pseudo exit with even finish to a pseudo exit middle state which sends finish again. The most outer state machine has a transition from the middle pseudo exit with event finish to a internal state. when the most inner pseudo state sends event finish it ends in a NO transition on the most outer state machine instead of triggering the transition in the middle state machine. <snip>
Hi Richard,
Remove these lines:
//struct test1 : msm::back::state_machine< test1Sm > //{ //}; //typedef test1 Test1Player;
And replace them with:
typedef msm::back::state_machine< test1Sm > Test1Player;
And it should work.
Regards, Christophe _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost