On Wed, Jan 18, 2012 at 7:25 PM, Robert O'Callahan <rob...@ocallahan.org>wrote:
> On Thu, Jan 19, 2012 at 4:09 PM, Boris Zbarsky <bzbar...@mit.edu> wrote: > > > Whether other UAs can fix this bug on their end faster than they can add > > various canvas APIs is an interesting discussion, I suppose. > > > > Of course it's vastly better for the Web platform for browsers to fix their > bugs than add new APIs to work around the bugs :-) > > I completely agree and it's good to see Gecko has fixed this, we'll have to see what happens in Webkit though. Last I talked with Adam Barth early last year there were still security concerns with allowing SVG images to render to a canvas without tainting it. Gecko completely punts rendering foreignObject to a canvas while Webkit will let you do fun things like render a browser native control (ex. <button>) to a canvas by embedding it into SVG. I suppose the presence of foreignObject could taint it, though at some point we should work out the issue of rendering same domain HTML this way. - E