Re: java.awt.geom.AffineTransform

[Mark Wielaard <mark at klomp dot org>]

Hi Dave,

On Sun, 2004-08-15 at 00:14, David Gilbert wrote:
> I've attached some tests for the java.awt.geom.AffineTransform class. 
> With JamVM 1.1.4 and Classpath from CVS about 1 week ago, I get the
> following results:
> 
> FAIL: gnu.testlet.java.awt.geom.AffineTransform.createTransformedShape
> (number 10)
> FAIL: gnu.testlet.java.awt.geom.AffineTransform.inverseTransform (number
> 4)
> got -16.0 but expected 16.0
> FAIL: gnu.testlet.java.awt.geom.AffineTransform.inverseTransform (number
> 6)
> got -24.0 but expected 24.0
> FAIL: gnu.testlet.java.awt.geom.AffineTransform.inverseTransform (number
> 8)
> got -32.0 but expected 32.0
> FAIL: gnu.testlet.java.awt.geom.AffineTransform.inverseTransform (number
> 12)
> got -8.0 but expected 8.0
> FAIL: gnu.testlet.java.awt.geom.AffineTransform.inverseTransform (number
> 14)
> got -8.0 but expected 8.0
> 6 of 205 tests failed

Besides these I also get the following failures with transform (at least
with kaffe from CVS):

FAIL: gnu.testlet.java.awt.geom.AffineTransform.transform (number 11)
got 0.0 but expected 1.0
FAIL: gnu.testlet.java.awt.geom.AffineTransform.transform (number 12)
got 0.0 but expected 2.0
FAIL: gnu.testlet.java.awt.geom.AffineTransform.transform (number 13)
got 12.0 but expected 6.0
FAIL: gnu.testlet.java.awt.geom.AffineTransform.transform (number 14)
got 36.0 but expected 12.0
FAIL: gnu.testlet.java.awt.geom.AffineTransform.transform (number 15)
got 20.0 but expected 10.0
FAIL: gnu.testlet.java.awt.geom.AffineTransform.transform (number 16)
got 54.0 but expected 18.0
FAIL: gnu.testlet.java.awt.geom.AffineTransform.transform (number 17)
got 28.0 but expected 14.0
FAIL: gnu.testlet.java.awt.geom.AffineTransform.transform (number 18)
got 72.0 but expected 24.0
FAIL: gnu.testlet.java.awt.geom.AffineTransform.transform (number 19)
got 0.0 but expected 9.0
FAIL: gnu.testlet.java.awt.geom.AffineTransform.transform (number 20)
got 0.0 but expected 10.0

This is from the second transform using floats as result.
Shouldn't the input array d[] be reset before the transform?

I also have a bigger total of tests:
16 of 257 tests failed

> The first error is some unusual behaviour in Sun's implementation where
> the createTransformedShape(Shape) method will accept a null argument.
> Classpath throws a NullPointerException.  I searched Sun's bug parade
> and it is in there as report 4190350, status "will not fix".

So it is a bug they refuse to fix.
Shouldn't the test then fail on their implementation?

Cheers,

Mark

Attachment: signature.asc
Description: This is a digitally signed message part