Use JCodings to check for uppercase char instead of Java String

This improves on 3ec4747 by:

* Avoiding a Rope to String conversion just to be able to extract the
  first character in an encoding-aware way

* Avoiding issues with surrogate characters -- as pointed out by
  @eregon during PR review -- `String::charAt(...)` can still slice
  characters in half as discussed in
  <https://stackoverflow.com/questions/5903008/what-is-a-surrogate-pair-in-java>

PullRequest: truffleruby/2102

Author: ivoanjo

src/main/java/org/truffleruby/parser/lexer/RubyLexer.java

@@ -1798,7 +1798,7 @@ private int identifier(int c, boolean commandState) {
             }
             tempVal = createTokenRope();
 
-            if (result == 0 && Character.isUpperCase(tempVal.getString().charAt(0))) {
+            if (result == 0 && isFirstCharacterEncodingAwareUppercase(tempVal)) {
                 result = RubyParser.tCONSTANT;
             } else {
                 result = RubyParser.tIDENTIFIER;
@@ -3493,12 +3493,12 @@ public boolean update_heredoc_indent(int c) {
     }
 
     public void validateFormalIdentifier(Rope identifier) {
-        int first = identifier.getString().charAt(0);
-
-        if (Character.isUpperCase(first)) {
+        if (isFirstCharacterEncodingAwareUppercase(identifier)) {
             compile_error("formal argument cannot be a constant");
         }
 
+        int first = identifier.getString().charAt(0);
+
         switch (first) {
             case '@':
                 if (identifier.get(1) == '@') {
@@ -3665,4 +3665,11 @@ protected boolean isSpaceArg(int c, boolean spaceSeen) {
         return isARG() && spaceSeen && !Character.isWhitespace(c);
     }
 
+    /** Encoding-aware (including multi-byte encodings) check of first character of a given rope, usually to determine
+     * if it is a constant */
+    private boolean isFirstCharacterEncodingAwareUppercase(Rope rope) {
+        byte[] ropeBytes = rope.getBytes();
+        int firstCharacter = rope.encoding.mbcToCode(ropeBytes, 0, ropeBytes.length);
+        return rope.encoding.isUpper(firstCharacter);
+    }
 }