On Thu, Oct 12, 2006 at 09:08:39AM +0200, Udo Giacomozzi wrote: > Hello strk, > > Wednesday, October 11, 2006, 11:25:01 PM, you wrote: > s> This can be a solution for the 'renderer' availability during > s> parse time, anyway this was already fixed by all but GTK/AGG > s> combination. > > Oh. Missed that. How was it fixed?
For the cairo backend, which also needs size to be specified, the render_handler_cairo exposes an additional function to set the drawable. Can we do something similar for the AGG backend ? Splitting movie_def_impl creation from actual firing of the "loader" thread would require modification of the create_library_movie and create_movie and friends, which would be a not-so-trivial change. > s> The other problem we have, and that persist with all renderers > s> (what we call 'font antialiasing'), is the current implementaion > s> of fonts glyph rendering. It relies on the fact that the whole > s> SWF has been parsed before rendering the textured glyphs. > s> It must be modified to allow on-demand rendering of them, > s> or, if you prefer, right-after-reading-definition rendering > s> of each font. > > I understand. I'd say on-demand rendering should be a nice solution. > Tthat way we keep all rendering calls in one thread, right? Right. Anyway, for the short-term problem (font antialiasing) right-after-reading-definition would be good enough. > I think I can do it, but I'm very busy at the moment, so if anyone > else wants to do it, no problem ;) I think you're in the best position to take this task... we can probably wait until you have more time :) --strk; _______________________________________________ Gnash-dev mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnash-dev
