Re: [Fenfire-dev] Re: [Fencommits] libvob: various fixes: hw<->sxsw, concatInverse, TranslationLob and RectBg

[Matti Katila]

On Tue, 22 Feb 2005, Benja Fallenstein wrote:
> On Sat, 19 Feb 2005 18:57:22 +0200, Matti J. Katila <address@hidden> wrote:
> > Sat Feb 19 17:11:56 EET 2005  Matti J. Katila <address@hidden>
> >   * various fixes: hw<->sxsw, concatInverse, TranslationLob and RectBg
> >
> >      Trans root = new Trans(){
> >             public String toString() { return "root"; }
> > -           float sx() { return 1; }
> > -           float sy() { return 1; }
> > -           float w() { return width; }
> > -           float h() { return height; }
> > +           float sx() { return width; }
> > +           float sy() { return height; }
> > +           float w() { return 1; }
> > +           float h() { return 1; }
> 
> What?!? *Wrong*! That's reverting a bugfix I did a week ago! :-o

Was it already bugfixed?-)  

> And why did you revert the w/h of the root cs to 1/1? IMO it's
> "obvious" that it should be the w/h of the scene.

OpenGL and AWT must return the same answer from getSqSize() and they need 
to rerender same kind of view for same code. No exceptions.
 
> The root cs is *definitely* not scaled as that sx()/sy() says -- that
> would mean that translate(0, .5f, .5f) would go to the center of the
> window. It doesn't.

We need a specification right now. And both APIs need to follow it.
 

> > -               into.setX(cs(), 
> > into.x(p1)-into.x(p2)*into.sx(p1)/into.sx(p2));
> > -               into.setY(cs(), 
> > into.y(p1)-into.y(p2)*into.sy(p1)/into.sy(p2));
> > +               into.setX(cs(), 
> > into.x(p2)*into.sx(p1)/into.sx(p2)+into.x(p1));
> > +               into.setY(cs(), 
> > into.y(p2)*into.sy(p1)/into.sy(p2)+into.y(p1));
> 
> That's not a fix, it's putting in a bug, but I guess you fixed that
> with your other commit =)

Please make it look so that I can understand it. "x+y/sx*yw+.-asdf+*asdf" 
is complicated and it doesn't say wat it does ;) Second, it didn't work =)


   -Matti