From: "someuserat" <marktutt@excite.com>
This is getting harder.
bind(&DataMember::equals, bind(&WidgetMap::getIndexedMember, _1), d)
would be my guess. Easier would be to define a helper function
bool memberEquals(WidgetMap const & m, DataMemberPtr d) { return m.getIndexedMember().equals(d); }
and then use
bind(memberEquals, _1, d) in dropWidgetMember.
I actually ended up figuring out how to write a function object class to do something similar. This has been a good learning experience! Is there a down side to the function object approach over the helper function? From what I could tell, the helper function had to be a global function, which just didn't seem right.
The two approaches are about equal. The function object requires slightly more effort since you need boilerplate code to hold the state ('d' in the example); boost::bind is good at holding state. The function object is also less flexible, the function is more likely to be reusable in other contexts. On the other hand the function will typically not be inlined, although this is rarely a concern. Of course you can use a stateless function object similar to 'memberEquals', too. Both the function and the helper class can be defined in an unnamed namespace; the function can be made static. Neither is more global than the other, so this is not a deciding factor. You can make the function member of WidgetMap, too, if you prefer the OO style, although this is a bit controversial. :-) boost::bind just gives you more options; it's up to you to choose what works best.