NSCompositingOperation is an EnumeratedType defined by NSImage that describes how drawing is to be done.
The constants are described in terms of having source and destination images, each having an opaque and transparent region. The destination image after the operation is defined in terms of the source and destination before images as follows:
An example is on your HD at file:///Developer/Examples/AppKit/CompositeLab.
Many drawing functions let you specify the operation to use, including -[NSImage compositeToPoint:fromRect:operation:fraction:], -[NSImage drawInRect:fromRect:operation:fraction:], and NSRectFillUsingOperation().
Prior to 10.4, functions that don’t take an operation parameter will choose their own. NSBezierPath uses NSCompositeSourceOver, while NSRectFill() uses NSCompositeCopy. See FillRectVsNSRectFill.
On 10.4 and later, functions that don’t take an operation parameter will usually use NSGraphicsContext’s, which can be changed with -[NSGraphicsContext setCompositingOperation:]. The default is NSCompositeSourceOver. NSRectFill() ignores -[NSGraphicsContext compositingOperation] and continues to use NSCompositeCopy.
If you need to draw a NSBezierPath with a specific NSCompositingOperation on 10.3 or earlier, you can do so like this:
[NSGraphicsContext saveGraphicsState]; [bezierPath addClip]; NSRectFillUsingOperation([bezierPath bounds], operation); [NSGraphicsContext restoreGraphicsState];
These are the classic Porter-Duff rules for digital image compositing upon which modern imaging is built. The original 1984 paper can found at http://www.keithp.com/~keithp/porterduff/.