On Sat, Aug 2, 2008 at 1:52 PM, Nitro <[EMAIL PROTECTED]> wrote: > The new float canvas will output with higher quality than the last one, > mainly due to antialiasing (and things like gradient fills). Performance > is not too bad right now, too.
Great, that's nice to know. > Are you interested only in svg import or also in svg export? Export is > easier to do for us than import, so if the interest is high enough, I'll > try to take a stab at it. Export is not really necessary right now. > Basically yes. Right now there are two modes. One is for "coarse" testing > and uses bounding boxes. Then there's a finer method which tests whether > the hit point is in the "path" of an object. Things like polygons, > ellipses etc. have a path object, the only objects that don't have a path > are bitmaps and fonts. The current method does not take the alpha value at > the hit pixel into account. On the one hand this allows for easy selection > of (nearly) completely transparent object, on the other hand you might > want a different behaviour (NOT selecting the object if the alpha value at > the hit point is below X). What would be your default way of using fc? > Taking the alpha value at the pixel into account or not? well, I won't be dealing with transparent objects. All my objects will have solid borders so a sane default way of hit testing for such objects will suffice. It could be configurable which alpha value corresponds to a hit tested object so the user can change that value and change the way hit testing works. It is important to have future use of floatcanvas in mind. /Chris _______________________________________________ FloatCanvas mailing list [email protected] http://mail.mithis.com/cgi-bin/mailman/listinfo/floatcanvas
