Hi ... I have been looking into the GFC class system for inspiration, and when looking at this library many things looks almost but not quite like gtkmm. But gtkmm is the most used and supported library so I guess it would be smart to support this, and it certainly do work, I have to say. But this, I don't hope, does not mean that Gtkmm can't improve :-)
There have reasonly been some discussion about overhead, and size of a Gtkmm application, and it seem Gtkmm was guilty in some unnecessary overhead. To solve some of these problems in GFC, it used "virtual signal system" to decouple (see description in : http://gfc.sourceforge.net/faq.html) the signal overhead, from the default widget tree. As I understand it, Gtkmm uses some kind of (pseudo) automatic generated code, so this should be possible to alter, but the compatibility issue still remains, and the model may take some getting use to for the developers using it. The gain from this model would be better performance (not a lot of unused virtual function), and lower memory footprint (no or very small vtables). Is this possible, or are the strong reasons not to do things like this ? /BL _______________________________________________ gtkmm-list mailing list gtkmm-list@gnome.org http://mail.gnome.org/mailman/listinfo/gtkmm-list
