The canvas spec currently defines this:
(http://www.whatwg.org/specs/web-apps/current-work/multipage/the-canvas-element.html#check-the-usability-of-the-image-argument)

"When a user agent is required to check the usability of the image argument, where image is a CanvasImageSource object, the user agent must run these steps, which return either good, bad, or aborted:
1. If the image argument is an HTMLImageElement object that is not fully decodable, or if the image argument is an HTMLVideoElement object whose readyState attribute is either HAVE_NOTHING or HAVE_METADATA, then return bad and abort these steps.
2. If the image argument is an HTMLCanvasElement object with either a horizontal dimension or a vertical dimension equal to zero, then the implementation throw an InvalidStateError exception and return aborted."

This is not matching up with current implementations and the language in http://www.whatwg.org/specs/web-apps/current-work/multipage/the-canvas-element.html#dom-context-2d-drawimage:
"If the image has no image data, throws an InvalidStateError exception. If the image isn't yet fully decoded, then nothing is drawn."

After some discussion, we believe that this section should read as follows:
"When a user agent is required to check the usability of the image argument, where image is a CanvasImageSource object, the user agent must run these steps, which return either good, bad, or aborted:
1. If the image argument is an HTMLImageElement object in the broken state, then throw an InvalidStateError exception, return aborted and abort these steps.
2. If the image argument is an HTMLImageElement object that is not fully decodable, or if the image argument is an HTMLVideoElement object whose readyState attribute is either HAVE_NOTHING or HAVE_METADATA, then return bad and abort these steps.
3. If the image argument has either a horizontal dimension or a vertical dimension equal to zero, then the implementation throw an InvalidStateError exception and return aborted."