Reland of 'Improve canonicalization of mailto url path components'

third_party/WebKit/LayoutTests/fast/url/script-tests/mailto.js

 description("Tests mailto: URL canonicalization");
 
 cases = [ 
   ["addr1", "addr1"],
   ["addr1@foo.com", "addr1@foo.com"],
 
   // Trailing whitespace is stripped.
   ["addr1 \t ", "addr1"],
   ["addr1?to=jon", "addr1?to=jon"],
   ["addr1,addr2", "addr1,addr2"],
-  ["addr1, addr2", "addr1, addr2"],
+  ["addr1, addr2", "addr1,%20addr2"],
+  ['addr1,"addr2"', "addr1,%22addr2%22"],
   ["addr1%2caddr2", "addr1%2caddr2"],
 
   // U+10300 input as UTF-16 surrogate pair, expected as escaped UTF-8
   ["\uD800\uDF00", "%F0%90%8C%80"],
   ["addr1?", "addr1?"],
 
   // Null character should be escaped to %00
   // Not sure how to test null characters in JavaScript.  They appear to get
   // stripped out.
   // ["addr1\\0addr2?foo", "addr1%00addr2?foo"],
 ];
 
 for (var i = 0; i < cases.length; ++i) {
   shouldBe("canonicalize('mailto:" + cases[i][0] + "')",
            "'mailto:" + cases[i][1] + "'");
 }
   // Invalid -- UTF-8 encoded surrogate value.
 shouldBeFalse("canonicalize('mailto:\xed\xa0\x80') == 'mailto:%EF%BF%BD'");