[boost] [statechart] Problems if I make dtorS virtual?

From: BRIDGES Dick (Dick.Bridges_at_[hidden])
Date: 2005-08-24 13:16:10

When compiling the examples (statechart library downloaded from sf 18
Aug),
I get the following warnings:

<warnings format=edited>
/usr/local/boost_1_33_0/boost/statechart/detail/state_base.hpp:
In instantiation of
`boost::statechart::detail::state_base<std::allocator<void>,
boost::statechart::detail::rtti_policy>':
/usr/local/boost_1_33_0/boost/statechart/state_machine.hpp:443: 
instantiated from 
`boost::statechart::state_machine<Machine, Greeting,
std::allocator<void>, 
boost::statechart::null_exception_translator>'
../main.cpp:12: instantiated from here 
/usr/local/boost_1_33_0/boost/statechart/detail/state_base.hpp:74:
warning: `class
boost::statechart::detail::state_base<std::allocator<void>, 
boost::statechart::detail::rtti_policy>' has virtual functions but 
non-virtual destructor

[snip]

/usr/local/boost_1_33_0/boost/statechart/simple_state.hpp:189:
instantiated from 
`boost::statechart::simple_state<Greeting, Machine,
boost::mpl::list<mpl_::na,
mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na,
mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na,mpl_::na, mpl_::na,
mpl_::na,
mpl_::na, mpl_::na, mpl_::na, mpl_::na>, has_no_history>'
../main.cpp:14:   instantiated from here
/usr/local/boost_1_33_0/boost/statechart/detail/leaf_state.hpp:27:
warning: `class
boost::statechart::detail::leaf_state<std::allocator<void>,
boost::statechart::detail::rtti_policy>' has virtual functions but
non-virtual destructor
</warnings>

I changed ~simple_State() and ~state_base() to virtual and the warnings
disappeared.  However, ~state_base() contains the following annotation:

// This destructor is not virtual for performance reasons. The library
// ensures that a state object is never deleted through a state_base
// pointer but only through a pointer to the most-derived type.
~state_base() {}

As a quick, not-definitive check, I built PingPong* with both the
virtual
and non-virtual dtorS using gcc 3.4.2 on an old i386-redhat-linux
machine.
The results appear to be the reverse of what I would expect, but I
checked
to verify that I have the correct result/config correspendence.

..................non-virtual.....virtual
PingPongSingle....  4.08     .....  3.71
PingPongMulti1....  6.01997  .....  5.85999
PingPongMulti2.... 19.53     ..... 19.55

Reminds me of the old joke:  "I see fine.  I just don't understand what
I'm looking at."  Two questions:  1) will making the dtorS virtual lead
to adverse consequences I simply haven't encountered yet; and 2) can
someone help me understand what I did wrong in setting up this test?

<extra_credit ot=true my_understanding=faulty>
With a well-developed pipelined architecture and a reasonably rich set
of addressing operations, shouldn't there be little or no difference
between direct and indirect access times - given that the bookeeping
normally associated with virtual classes has already been paid for?
</extra_credit>

Thanks to any/all for help with this.

Regards,
Dick Bridges

"Multithreading is just one damn thing after, before, or simultaneous
with another."  Scott Meyers and Andrei Alexandrescu