Composition vs. Deriviation in Gtkmm

[kenton wilson gmail com]

Hello All,

In the Gtkmm tutorial (http://developer.gnome.org/gtkmm-tutorial/2.24)
many of the examples are implemented by having the user defined
example class derive from a Gtkmm class. Most likely many (if not
all), of these examples could of been implemented by using some form
of composition instead of derivation. As you all know, in C++,
composition is generally preferred over derivation. Which leads me to
ask, in the context of Gtkmm, is there a rationale for preferring
derivation over composition (or vice-versa). What are the pros and
cons of each approach in implementing your UI using Gtkmm?

TIA,
Kenton