Marked the static instances as final

Marked ALL_MODIFIERS_MASK as final

Added back in check for modifiers applying to input events. Turns out some of the examples expect this behaviour (NavigationExample being one).

Change path back to private scope. Accident caused while trying to figure out a way of unit testing the generated path.

Removed some unnecessary imports from PCanvas

Refactored mousePressed and mouseReleased so that code common to each is now in a helper method. Net result is a much lower cyclomatic complexity for each.

Unit tests

Added getInputEventListeners to PCanvas and PNode making them return an array of PInputEventListener objects. This makes it fall inline with the other listeners inherited from JComponent and it also makes listener methods testable.

Reverted override of createInverse in PAffineTransform since the fact that it didn't throw anything was being overlooked in Hudson, but honoured in Eclipse

Removing some calls to printStackTrace(), replacing them with appropriate responses Issue 101

Fixing bug with PImage being run in Headless Mode

Unit Testing for PImage and removed a little cruft

Some Unit Tests for PInputEvent

Adding tests for PText

Adding tests for PText and making it return empty string by default rather than null. More inline with JTextField

Fixed a memory leak with PCanvas instances registering themselves with KeyboardFocus and never unregistering themselves. ...

Added a Hierarchy listener so that the inputsources are installeded and uninstalled when a canvas gets added or removed to a parent.

Fixed Issue 100; Plastered PInputEventFilter with Unit Tests and then made it consider the masks conditionally. ...

andMask if it's non-zero
notMask if it's non-zero
orMask if it's not ALL_MODIFIERS (I can't think of a case where you might want to say, "Only allow events as long as they have a modifier, I don't care which one")

Fixed Issue 91; deprecated repaintFromlayer(PBounds, PNode) in favor of the more specific repaintFromlayer(PBounds, PLayer) making it impossible to accidentally pass a camera say to this method.

Fixed Issue 96; Made PAffineTransform throw PAffineTransformException (a subclass of RuntimeException) whenever a non-invertible transform is encountered. ...

In the process, I removed 2 printStackTrace calls from the code base.

Fixed Issue 92; make it check for empty stack before popping the cursor

Fixed Issue 94; made removeLayer(PLayer) "fail" silently. Also added some additional tests and refactoring. Sorry for such a large commit, I'd try to split it up, but it's hard to tell what's safe to leave out and what's not. I know that the whole together unit tests fine. I have to remember to checkin more frequently.

Simplifying getInverseTransform and making it not call printStackTrace() since it was handling that case anyway. Issue #96 may make this a non-issue anyway.

Added some trivial tests for PPanEventHandler and added come accesors to it for maxAutoPanSpeed and minAutoPanSpeed. ...

Also, cleaned up test for PAffineTransform.

Folding nested if statements into themselves. Makes PMD stop complaining so loudly.

Refactored PDebug code so that PMD didn't complain about two ifs that could be joined. In the process making it slightly clearer.

Replaced complex xor logic with Java's xor operator.

Part of a solution to Issue #53; made toImage not use GraphicsEnvironment in the case of running in Headless mode.

Added accessors for render quality on PCanvas (Issue 93) , and also wrote more tests to confirm that Issue 87 is fixed.

Added some move tests to PCanvas

made toImage(int, int, Paint) return an image of the size requested. Since it seems that the resizing is already done in PNode.paint it was just a matter of deleting the resize code. ...

Also, I've added 3 tests to PNodeTest that try and confirm the behavior.