document.documentElement.innerHTML might be better than document.body.parentNode.innerHTML. The latter fails if there is no body element, e.g. when the page is an SVG document.
If you want a closer approximation of the full source, outerHTML might also be better.
I've also seen "new XMLSerializer().serializeToString(document)" suggested. That seems to give the most complete source, but I've also read that it might have problems with things that need escaping. I have no experience with that approach because for what I needed the first thing I found when Googling, document.documentElement.innerHTML, gave me what I needed.
One more thing to consider. All of the above I'd expect give you source that would produce the currently displayed page including any modifications that were made after loading by JavaScript (which is probably what you'd want for cheating on a test so is fine). I'm not sure that is the same as what "view source" gives--does it give the current page or the page as it came over the wire?
