Murray Cumming wrote:
I would be surprised if so.On Wed, 2008-07-09 at 17:37 +0300, Michael Boccara wrote: As soon as you want a clutter actor with a more complex geometry than just a rectangle, the only way is to subclass ClutterActor and override the 'paint' method. In clutter, it would mean defining and binding the clutter_custom_actor_paint() function, with all the GObject subclassing nightmare it involves. Or am I missing something ? Actually just recently I got pretty optimistic about the possibility of relying only on the C++ inheritance in the cluttermm level only, since I noticed an interesting change in the clutter_actor_paint() definition: void clutter_actor_paint (ClutterActor *self) { // was like this in clutter-0.6.0 // if (G_LIKELY (klass->paint)) // klass->paint (self); // and now in clutter-0.8.0 it is: g_signal_emit (self, actor_signals[PAINT], 0); } Up until clutter-0.6.0, the paint virtual method was hard-assigned to clutter_custom_actor_paint, which was making it impossible to get to use Clutter::CustomActor::paint() from the clutter_main() loop. But now that clutter-0.8.0 is emitting a Glib signal I guess it makes it possible for cluttermm to bind the PAINT signal to the Clutter::CusomActor::paint() virtual method. Do I get it right ? However, I recommend waiting for us to update cluttermm to 0.8 first though. We didn't do that yet because it currently depends on other clutter-* libraries that were no available in 0.7/0.8 versions (which I consider to be a rather strange release schedule). I guess we'll have something this week. -- Michael Boccara Graphtech Herzliya, Israel-- To unsubscribe send a mail to [EMAIL PROTECTED] |
- [clutter] Implementing a new actor via cluttermm Michael Boccara
- Re: [clutter] Implementing a new actor via cluttermm Armin Burgmeier
- Re: [clutter] Implementing a new actor via cluttermm Murray Cumming
- Re: [clutter] Implementing a new actor via clutte... Michael Boccara
- Re: [clutter] Implementing a new actor via cl... Murray Cumming
