fix(sanitizeUri): sanitize URIs that contain IDEOGRAPHIC SPACE chars

Chrome 62 was not sanitizing dangerous URLs containing
JavaScript, if they started with these "whitespace" characters.

---

Browsers convert `　javascript:alert(1)` as an attribute value to
`javascript:alert(1)`. So the sanitizer gets the second string and is
able to strip the javascript from the attribute.

But Chrome (<62) only did this after you read it and wrote it back again.
So we added a bit of code that tried to get Chrome to do its conversion
before sanitizing it:
https://github.com/angular/angular.js/blob/817ac567/src/ngSanitize/sanitize.js#L406-L417

I believe that Chrome 62 now does not do this conversion any more, instead
it just leaves the attribute value alone, whatever you do to it.

This fix uses `trim()` to remove the problematic whitespace before
sanitizing, which appears to solve the problem.

Closes #16288

Author: petebacondarwin

src/ng/sanitizeUri.js

@@ -62,7 +62,7 @@ function $$SanitizeUriProvider() {
     return function sanitizeUri(uri, isImage) {
       var regex = isImage ? imgSrcSanitizationWhitelist : aHrefSanitizationWhitelist;
       var normalizedVal;
-      normalizedVal = urlResolve(uri).href;
+      normalizedVal = urlResolve(uri && uri.trim()).href;
       if (normalizedVal !== '' && !normalizedVal.match(regex)) {
         return 'unsafe:' + normalizedVal;
       }

test/ngSanitize/sanitizeSpec.js

@@ -237,11 +237,9 @@ describe('HTML', function() {
       .toEqual('');
   });
 
-  if (isChrome) {
-    it('should prevent mXSS attacks', function() {
-      expectHTML('<a href="&#x3000;javascript:alert(1)">CLICKME</a>').toBe('<a>CLICKME</a>');
-    });
-  }
+  it('should prevent mXSS attacks', function() {
+    expectHTML('<a href="&#x3000;javascript:alert(1)">CLICKME</a>').toBe('<a>CLICKME</a>');
+  });
 
   it('should strip html comments', function() {
     expectHTML('<!-- comment 1 --><p>text1<!-- comment 2 -->text2</p><!-- comment 3 -->')