Re[2]: [Gnash-commit] gnash ChangeLog gui/fb.cpp gui/fbsup.h gui/gui....

[Udo Giacomozzi]

Hello strk,

Friday, October 13, 2006, 10:12:21 AM, you wrote:
s> I've prepared support for the runtime option, my question now is:
s> should I really see any red rectangle with that enabled ?
s> Because I'm not (gtk/ogl).

I haven't tested this with GTK/OGL but it just calls draw_poly() which
should be supported by any renderer. Please note that you won't see
any red rectangle when the entire frame is re-rendered.

Maybe debug the coordinates using log_msg.

s> I suppose the rectangle is just too big (and seen your 'sligtly enlarge
s> it due to antialiazing, so it might fall outside of visible area).

Only for the FB/AGG combination. Also, the triangulating render
handlers do not support this feature, yet (but it can be done
universally). However, you should see the rectangle anyway.

s> If I'm right the set_invalidated() call might be a bit too generic, 
s> while we might have functions to report exact rectangle being updated
s> (the parent would union oall of invalidated rects in childs).

set_invalidated() just sets the flag of that particular instance. I
wanted a wrapper for m_invalidated=1 because it's easier to debug and
who knows if other actions should be taken one day when invalidating.

Which childs get re-drawn is decided by get_invalidated_bounds(). It
has a "force" parameter which includes all childs even if they have
not been modified themselves. For example, when you rotate a MovieClip
that has 5 childs, they all have to be redrawn.

Udo