I wonder if the 'signal' lib is thread-safe? i.e. a publisher is doing something, at the same time a subscriber is trying to 'disconnect' from the slot, will it get problem? Found that it only set a flag to indicate the slot in disconnection, but never remove the slot? So that the lib is thread-safe? ********************** It won't remove anything from the 'slots' container: template<typename Function> void do_disconnect(const Function& f, mpl::bool_<false>) { // Notify the slot handling code that we are iterating through the slots BOOST_SIGNALS_NAMESPACE::detail::call_notification notification(this->impl); for (iterator i = impl->slots_.begin(); i != impl->slots_.end(); ++i) { slot_function_type& s = *unsafe_any_cast<slot_function_type>(&i->second); if (s == f) i->first.disconnect(); } }

Check also the Mirror library both in the vault and the Sandbox .. though if you want just the type names Mirror might by too heavyweight. On Sat, Sep 26, 2009 at 8:36 AM, joel <joel.falcou@lri.fr> wrote:

Andrey Semashev wrote:

...

Emil Dotchevski wrote:

...

Do we have in Boost something that would attempt to return the pretty version (meaning, it'd fall back to returning the ugly version) of the std::type_info::name string?

Although it's not official Boost, I have a type_info_wrapper in Boost.Log that does this.

http://boost-log.sourceforge.net/libs/log/doc/html/log/detailed/utilities.ht...

http://boost-log.sourceforge.net/libs/log/doc/html/boost/log/type_info_wrapp...

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Check boost.Identification in the vault

-- ___________________________________________ Joel Falcou - Assistant Professor PARALL Team - LRI - Universite Paris Sud XI Tel : (+33)1 69 15 66 35

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

-- ________________ ::matus_chochlik