Fix unicode length comparisons

the test suite had the wrong expected values for max/minLength for
unicode strings, they've been updated.

Use the number of unicode codepoints to compare lengths.

Author: halvorgb

core/src/main/java/org/everit/json/schema/StringSchema.java

@@ -131,7 +131,7 @@ public Pattern getPattern() {
     }
 
     private List<ValidationException> testLength(final String subject) {
-        int actualLength = subject.length();
+        int actualLength = subject.codePointCount(0, subject.length());
         List<ValidationException> rval = new ArrayList<>();
         if (minLength != null && actualLength < minLength.intValue()) {
             rval.add(new ValidationException(this, "expected minLength: " + minLength + ", actual: "

tests/src/test/resources/org/everit/json/schema/draft4/maxLength.json

@@ -28,7 +28,7 @@
       {
         "description": "two supplementary Unicode code points is long enough",
         "data": "\uD83D\uDCA9\uD83D\uDCA9",
-        "valid": false
+        "valid": true
       }
     ]
   }

tests/src/test/resources/org/everit/json/schema/draft4/minLength.json

@@ -28,7 +28,7 @@
       {
         "description": "one supplementary Unicode code point is not long enough",
         "data": "\uD83D\uDCA9",
-        "valid": true
+        "valid": false
       }
     ]
   }