Re: [Gzz] PEG 1003

[Benja Fallenstein]

Benja Fallenstein wrote:

> The information EVERY vob (I think) needs to draw itself is:
> - a position on the screen
> - a shape for clipping
>
> The clipping shape can be set in the drawing context before the vob is 
> asked to draw itself, so that the vob needn't worry about it.
>
> In affine coordinate systems, for EVERY vob we also need:
> - unit vectors of the affine transformation, for rotation/scaling/skewing
>
> This, too, can AFAIK be set in the graphics context before the vob is 
> called.
>
> Finally, SOME vobs (which draw a border around something) will also 
> need a shape to draw themselves in. This shape will often be the 
> clipping shape for some other vobs.


Going from vobs to coordinate systems, there are several options. For 
example, we could say that coordinate systems are only for 
transformation, and shapes are given in an orthogonal structure.

However, it seems better to couple the two: for example, when we place a 
cell into a window, we both want to translate and clip by the window's 
coordinates, and use the same scaling information as the coordsys the 
window is placed inside.

Ok, how about this:
- Each cs MAY have a shape.
- If it does have a shape, for now it's a box (specified by width and 
height).
- Cell border/background vobs need to be placed into a cs having a 
shape. They draw themselves as that shape.
- As proposed by Tuomas, each cs has a "clippedBy" coordinate set.
- The clipping area of a cs is the intersection of its shape (if any) 
with the shape of its clippedBy cs (if any) and with the shape of the 
clippedBy cs's clippedBy cs (if any) and so on, up to the root cs.

The interface could well be Tuomas' proposed coorders.java, but it would 
be different on the vobs' side: A vob could ask, "What is my position 
and size?" and draw itself into that rectangle. (It could also ask "What 
is my position?" and ignore the size-- this is what text vobs would be 
supposed to do.) They wouldn't have to worry about the "nasty case" from 
the Graphics tech report, because all this would be ORTHOGONAL to 
scaling (and rotation): The affine transformation and the clipping would 
be set in the graphics context before the vob is even called, and they 
would NOT be influenced by the "width" and "height" as passed to the vobs.

- Benja