[cairo] Re: GTK-Canvas

[Rick L Vinyard Jr]

On Tue, 2006-02-28 at 18:10 +0000, Gustavo J. A. M. Carneiro wrote: 
>   My proposal would be rather different.  Have a single CanvasItem
> object, which would emit a "draw" signal with a cairo_context_t as
> argument.

I too needed a canvas, so I put one together late last year.

If anyone's interested I have YACL (Yet Another Canvas Library; these
things are breeding like rabbits) at http://libpapyrus.sourceforge.net/

Actually, I think the fact that numerous canvas projects are popping up
speaks to the need for a powerful Gtk+ canvas.

I took that exact approach with a virtual draw method that takes a cairo
context as a parameter (actually, since I moved it to cairomm it takes a
Cairo::RefPtr<Cairo::Context> ) simply because I wanted to be able to
extend the drawing shapes as I needed. 

It has resulted in putting some basic shapes into the library like
GnomeCanvas has, but the focus was on the ability to inherit from
Papyrus::Drawable (or further down the hierarchy like Papyrus::Shape)
and extend the draw() or draw_shape() methods.

One of the ways that I handled the hit detection was through a virtual
method called pick(x,y). The default implementation was to check the
bounding box, but where important I left it open for the derived classes
to use more appropriate methods; say Preparata's trapezoid method for
point-location in polygons and such.

I also took the approach of separating the canvas library itself from
the Gtkmm widgets; the canvas itself only depends upon cairomm.

Anyway, if anyone has any feedback or comments I'd welcome the input.