I don't understand the reason for mapping boost::functions with different return type signatures ? What does it get you since to pick up the correct return value you must know exactly what each return type is for each map entry and there is nothing in the map entry itself to tell you this except the name. So you end up hard-coding knowledge of each entry in the map based on the name into your code. Doesn't sound like it gets you anything as far as design goes. Perhaps there is something wrong in your design that necessitates something like like this ? "Julia Donawald" <yg-boost-users@m.gmane.org> wrote in message news:al2tap$599$1@main.gmane.org...
Hi, I found the mistake why it throws me an exception... so now it works quite well.... what do you think about that solution in general?