Sync HTML element interfaces C-D with the spec

Source/core/html/HTMLCanvasElement.idl

Index: Source/core/html/HTMLCanvasElement.idl
diff --git a/Source/core/html/HTMLCanvasElement.idl b/Source/core/html/HTMLCanvasElement.idl
index 5f2ee7c88105497ec14666243e705cdfa4aa9bc2..25e95d7a7ef947231b88b5a120bd7ab22a02fe70 100644
--- a/Source/core/html/HTMLCanvasElement.idl
+++ b/Source/core/html/HTMLCanvasElement.idl
@@ -24,16 +24,16 @@
  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  */
 
+// https://html.spec.whatwg.org/#htmlcanvaselement
+
 typedef (CanvasRenderingContext2D or WebGLRenderingContext) RenderingContext;
 
 interface HTMLCanvasElement : HTMLElement {
-
+    // FIXME: width and height should be unsigned long.
     attribute long width;
     attribute long height;
 
-    [RaisesException] DOMString toDataURL(optional DOMString? type = null, optional any arguments);
-
-    // Note: this differs deliberately from the specified Web IDL for this function:
+    // FIXME: this differs deliberately from the specified Web IDL for this function:
     //     RenderingContext? getContext(DOMString contextId, any... arguments);
     // in order to eliminate the custom binding. It is functionally equivalent.
     //
@@ -47,4 +47,8 @@ interface HTMLCanvasElement : HTMLElement {
     // handled differently the return type could be changed to "RenderingContext?"
     // and the [CallWith=ScriptState] extended attribute removed.
     [CallWith=ScriptState] any getContext([Default=Undefined] optional DOMString contextId, [PermissiveDictionaryConversion] optional CanvasContextCreationAttributes attributes);
+
+    // FIXME: type should not be nullable or have a default value and the
+    // trailing arguments should be variadic.

[Jens Widell, 2015/03/08 09:36:46]

The variadic bit is something we might not want to fix. It's not observable
whether the argument is (handled as) variadic, and implementing it as such just
makes the code more complicated. 

[philipj_slow, 2015/03/08 14:28:58]

Yes, and I think this is a bit weird in the spec, it actually treats all of the
arguments a group, not just argument 2 and following: "Let file be a
serialisation of the canvas element's bitmap as a file, using the method's
arguments (if any) as the arguments."

I'm not sure what the comment should say: "Note that the arguments argument is
variadic in the spec, but that implementing it as such would only complicate the
implementation"?

[Jens Widell, 2015/03/08 15:56:28]

"... since only the first argument is used in our implementation"?

I just don't think we should add a FIXME if we don't think it should be fixed,
since chances are someone will come around and simply fix it.

[philipj_slow, 2015/03/09 10:43:25]

I agree, how's the new comment?

+    [RaisesException] DOMString toDataURL(optional DOMString? type = null, optional any arguments);
 };