Name: gsC80088 Date: 11/19/98


The documentation states a Rectangle with a negative
width or height is empty;
But, the isEmpty() function states it as less than
or EQUAL to zero;
But, an isEmpty() Rectangle of 0 width and height
does draws a one pixel dot;
But, with ANTIALIASING on the 0 Rectangle is NOT drawn;
The same occurs for 0 width lines.
A Rectangle with a width and height of 1 draws
a 2 pixel square Rectangle with or without
ANTIALIASING.
Rectangle.contains(int X, int Y, int W, int H)
incorrectly returns false when W or H equals 0 if
the point (X, Y) is in the rectangle(Look at the
source).

The following form a consistant set of rules.

Rectangle empty = new Rectangle(-1, -1);
Rectangle point = new Rectangle(0, 0);
Rectangle dot = new Rectangle(1, 1);
Rectangle rect = new Rectangle(0, 0, 2, 2);

A rectangle with negative width or height is
never drawn and rect.contains(empty) return false
for all posible locations of rect and empty.

A rectangle with 0 width or height is never drawn.
However, rect.contains(point) returns true when
point is at (0, 0), (0, 1) or (1, 0).

A rectangle with a width and height of 1 appears
as a single pixel assuming a 1 to 1 user to
device scaling. rect.contains(dot) returns true
for the same locations given for "point".

A rectangle with a width and height of 2 should
appears twice the size of a rectangle with a
width and height of 1 given any user to device
scaling that doesn't introduce rounding errors.

The following should hold for a point and dot
both located at (0, 0):

dot.contains(point) TRUE
dot.contains(dot) TRUE
point.contains(point) TRUE
point.contains(dot) FALSE

For rigor one could document the handling of
Rectangles of size (0, 1) or (1, 0) but the
behaviors should be clear at this point.

END OF RULES -------

Basicly a Rectangle of size 2x2 should appear
twice the size of a Rectangle of size 1x1 and
a Rectangle of size 0x0 should not appear but
should act like a Point and be able to be contained
in another Rectangle. The actual implementation
violates all three of these rules and is
inconsistant when used with ANTIALIASING.

- dwood
(Review ID: 42018)
======================================================================