Steps to reproduce: Open the testcase from Bug 391028 and check the error console: Error: uncaught exception: [Exception... "Component returned failure code: 0x80040111 (NS_ERROR_NOT_AVAILABLE) [nsIDOMCanvasRenderingContext2D.drawImage]" nsresult: "0x80040111 (NS_ERROR_NOT_AVAILABLE)" location: "JS frame :: https://bugzilla.mozilla.org/attachment.cgi?id=275356 :: anonymous :: line 9" data: no] reproduced on trunk and 1.8 branch
The testcase is supposed to do that; the png image isn't valid. The spec says drawImage should throw INVALID_STATE_ERR however.
Created attachment 549180 [details] very simple test case It's my impression from http://www.whatwg.org/specs/web-apps/current-work/multipage/the-canvas-element.html#dom-context-2d-drawimage that we should never produce NS_ERROR_NOT_AVAILABLE from drawImage; the only licensed failure modes are INVALID_STATE_ERR, TYPE_MISMATCH_ERR, INDEX_SIZE_ERR, and silently doing nothing. The test case I'm attaching uses an uninitialized Image object (I presume this is the same as an HTMLImageElement under the hood). I'm not sure what the *correct* behavior for this test case is; I'm guessing either INVALID_STATE_ERR or silent nop. However, looking at http://mxr.mozilla.org/mozilla-central/source/content/canvas/src/nsCanvasRenderingContext2D.cpp#3212 I see several places where we return NS_ERROR_NOT_AVAILABLE inappropriately; they should all be audited and corrected (and the analogous code in the Azure reimplementation too).
I can reproduce this problem with Firefox 16.0.2 on Mac OS X 10.8.2.
The fixes for bug 1003380 do not appear to have fixed this bug; if you load the test case attached here, you will see that it still generates NS_ERROR_NOT_AVAILABLE. Again, as I understand the spec, NS_ERROR_NOT_AVAILABLE should *never* be generated by this API. (I observe this with 45.2.0ESR/Linux.)
Fixed in bug 1297211. I didn't see this bug when I reported that one.