bump CodeMirror to 4.0.x

[jokeyrhyme]

https://groups.google.com/forum/#!topic/codemirror-announce/5WCakVK_xnY

As I write this 4.0.3 is available.

I did just try to bump this in bower.json and move on, but there were some unit tests that failed in PhantomJS.

If someone with more experience with ui-codemirror can accomplish this, that would be terrific, however, I may look at this next week myself if I get a chance.

[whitehat101]

I've been using CodeMirror 4 without issues. The errors in the tests are only happening in PhantomJS, when CM is trying to do some IE feature detection.

CodeMirror 4.0.3 or 4.1.0
Chrome 34.0.1847 (Mac OS X 10.9.2): Executed 20 of 20 SUCCESS (1.516 secs / 0.661 secs)
Firefox 18.0.0 (Mac OS X 10.9): Executed 20 of 20 SUCCESS (2.448 secs / 0.568 secs)
PhantomJS 1.9.7 (Mac OS X): Executed 20 of 20 (3 FAILED) (0.89 secs / 0.291 secs)

All three failures have a common cause:

TypeError: 'null' is not an object (evaluating 'r0.left')
    at hasBadBidiRects (.../codemirror.js:7180)

// Feature-detect IE's crummy client rect reporting for bidi text
var badBidiRects;
function hasBadBidiRects(measure) {
  if (badBidiRects != null) return badBidiRects;
  var txt = removeChildrenAndAdd(measure, document.createTextNode("A\u062eA"));
  var r0 = range(txt, 0, 1).getBoundingClientRect();
  if (r0.left == r0.right) return false;                      // line 7180
  var r1 = range(txt, 1, 2).getBoundingClientRect();
  return badBidiRects = (r1.right - r0.right < 3);
}

[whitehat101]

This codemirror bug was reported, and this was the maintainer's response:

codemirror/codemirror5#2409 (comment)

PhantomJS is starting to feel increasingly shaky. I don't really want to add too much code to the core to deal with its bugs. In our own test suite, tests that fail on Phantom and work on real browsers are simply disabled when running on Phantom. I suggest you do the same in your tests.

[jokeyrhyme]

Ah, okay. Well, I'll close this, as this isn't a bug in this project. Cheers!