[cairo] libpixman - libpixregion, libic and slim merged; cairo built with it

[Bill Spitzak]

On Tuesday 09 December 2003 01:45 pm, Keith Packard wrote:

> 'box' objects contain x1,y1 -> x2,y2 values, 'rectangles' contain x,y
> width,height values.  Doing that consistently will at least avoid driving
> me insane.
>
> I don't know which interface to prefer for external use.  Internally,
> boxes are *way* easier to do computations with because the values are
> independent, but I realize that many existing external interfaces have
> rectangles for some reason.

Despite the fact that it will break *my own* code, I would recommend the 
two-point version of boxes. The same form of boxes should be used 
*everywhere*, meaning inside the cairo implementation and also in the public 
interface.

Two points are more easily be intersected, unioned, transformed by orthogonal 
transforms, and converted to line segments. It can define a "winding 
direction" with (x1>x2)!=(y1>y2) which transforms correctly, this is 
necessary for non-zero-winding-rule regions.

Much software disagrees with what negative width or height means (reflected 
rectangles, zero, or is width unsigned?). Therefore it may be best to avoid 
width/height so that nobody makes incorrect assumptions about it.

Although width/height has a history in X, Mac and Windows users are quite 
used to the xy-pair (unfortunately they call it left, bottom, right, top, 
which implies a direction, we must make it clear that these are two points). 
So this will not alienate as many programmers as you may think.

C++ and many other languages provide the inline methods needed so that asking 
for the width and height of a box is no harder to write than getting the 
points.

-- 
                   ,~,~,~,~ ~ ~ ~ ~
     /\_       _|_========___         Bill Spitzak
 ~~~/\/\\~~~~~~\____________/~~~~~~~~ spitzak at d2.com