On Mon, Nov 12, 2007 at 10:06:50PM +0100, Udo Giacomozzi wrote: > Hello strk, > > Monday, November 12, 2007, 9:09:26 PM, you wrote: > s> We'd need similar euristic for proper and performant withinSquareDistance > s> test for curves. The classic approach seems to be approximate and compute > s> distances, so knowing the actual rendering scale would help finding > s> an appropriate approximation. I guess this still needs a Gui handler > s> in the movie_root, or we'll miss the actual on-screen pixel scale... > > You need to know the actual rendering scale anyway to know if the > stroke is magnified. Remember all strokes are at least one physical > pixel wide (hairline), even if they are downsized. Relative to the > definition this means the stroke width is increased.
Alright, can I count on you adding that Gui pointer in the movie_root to complete the planned stage matrix stuff ? > Anyway, don't forget that the convert-everything-to-shapes approach > would eliminate the need to check the strokes (and would make handling > of dynamic shapes tricky...aaarghh). The only experience I have with line-to-polygon conversion is an hell of performance, and I still dream about a future 'dashed-line' style I used to want as a flash coder (I wonder if the mouse pointer would hit the dash and not the gap now). --strk; _______________________________________________ Gnash-commit mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnash-commit
