Re: [fxtf-drafts] [geometry] DOMRect: use of unrestricted doubles - min(x, NaN)

So I think the relevant testcase here is sort of like this:

      var r = new DOMRect();
      r.x = 5;
      r.width = NaN;

here Firefox nightly and Chrome dev alert `5`, while Safari Tech Preview alerts `NaN`.

It does seem to me that reliably returning `NaN` here would be somewhat better, but I'm not an expert on this code or how people use these APIs.  If they use them.

GitHub Notification of comment by bzbarsky
Please view or discuss this issue at using your GitHub account

Received on Tuesday, 17 April 2018 20:58:12 UTC