[cp-patches] FYI: More geom bugfixes.

[Sven de Marothy]

Hi,
I just commited the following fixes to Arc2D.
This should wrap up the remaining awt.geom failures in mauve, 
with the exception of the following two:
FAIL: gnu.testlet.java.awt.geom.Arc2D.getEndPoint (number 1) 
FAIL: gnu.testlet.java.awt.geom.Arc2D.getPathIterator (number 8)

Which both are due to small floating-point deviations (< 1E-14). I think
the tests should be less strict here. Even a 1E-5 deviation would be 
tolerable here, IMHO. I mean, for the PathIterator deviation to be
noticable, you'd either have to scale up the Arc 100 trillion times, or
be drawing your arc on a device with that kind of resolution. Like an
electron microscope. I'll fix it when we port to that platform. :-)

/Sven

2004-11-17  Sven de Marothy  <address@hidden>

        * java/awt/geom/Arc2D.java,
        (setAngleStart): Corrected (wrong sign on atan2 y parameter)
        (setAngles): Likewise
        (containsAngle): Return false on zero extent, don't include     final
angle
        (contains): Treat OPEN-type arcs like CHORD ones, not as PIE    ones.

arc2d.patch
Description: Text Data