boost::thread, condition, mutex, and locks. The acyclic_mutex stuff uses
thread_specific_ptr.
> Since I am using a RTOS my plan is replace the active object scheduler
> with the RTOS scheduler and use the RTOS queue in place of the
> activation queue. Are there any major issues involved with making
> those kinds of substitutions?
I don't know, as I don't know what your RTOS scheduler/queue do. But I'd
blindly guess your RTOS scheduler deals with handing out timeslices to
different threads/processes, which isn't really what poet::scheduler is
doing.
> One other idea is to use state machines from boost::statechart as the
> servant. I can't find mention of the servant in the libpoet code so I
> am assuming that the servant implementation does not matter. Is my
> assumption correct?
Your servant object can be used to provide the "passive function" parameter
when creating an active_function. The active_object_example.cpp example
program in the documentation may be helpful.