Merged in dscorbett/pygments-racket (pull request #358)

Improve RacketLexer

2 files changed, 27 insertions, 11 deletions

diff --git a/pygments/lexers/jvm.py b/pygments/lexers/jvm.py
index 69d90b78..4633c66f 100644
--- a/pygments/lexers/jvm.py
+++ b/pygments/lexers/jvm.py
@@ -34,14 +34,19 @@ class JavaLexer(RegexLexer):
     filenames = ['*.java']
     mimetypes = ['text/x-java']
 
-    flags = re.MULTILINE | re.DOTALL
+    flags = re.MULTILINE | re.DOTALL | re.UNICODE
 
     tokens = {
         'root': [
             (r'[^\S\n]+', Text),
             (r'//.*?\n', Comment.Single),
             (r'/\*.*?\*/', Comment.Multiline),
-            (r'@[a-zA-Z_][\w\.]*', Name.Decorator),
+            # method names
+            (r'((?:(?:[^\W\d]|\$)[\w\.\[\]\$<>]*\s+)+?)' # return arguments
+             r'((?:[^\W\d]|\$)[\w\$]*)'                  # method name
+             r'(\s*)(\()',                               # signature start
+             bygroups(using(this), Name.Function, Text, Operator)),
+            (r'@[^\W\d][\w\.]*', Name.Decorator),
             (r'(assert|break|case|catch|continue|default|do|else|finally|for|'
              r'if|goto|instanceof|new|return|switch|this|throw|try|while)\b',
              Keyword),
@@ -50,20 +55,15 @@ class JavaLexer(RegexLexer):
              r'transient|volatile)\b', Keyword.Declaration),
             (r'(boolean|byte|char|double|float|int|long|short|void)\b',
              Keyword.Type),
-            # method names
-            (r'^(\s*(?:[a-zA-Z_][\w\.\[\]<>]*\s+)+?)' # return arguments
-             r'([a-zA-Z_]\w*)'                        # method name
-             r'(\s*)(\()',                            # signature start
-             bygroups(using(this), Name.Function, Text, Operator)),
             (r'(package)(\s+)', bygroups(Keyword.Namespace, Text)),
             (r'(true|false|null)\b', Keyword.Constant),
             (r'(class|interface)(\s+)', bygroups(Keyword.Declaration, Text), 'class'),
             (r'(import)(\s+)', bygroups(Keyword.Namespace, Text), 'import'),
             (r'"(\\\\|\\"|[^"])*"', String),
             (r"'\\.'|'[^\\]'|'\\u[0-9a-fA-F]{4}'", String.Char),
-            (r'(\.)([a-zA-Z_]\w*)', bygroups(Operator, Name.Attribute)),
-            (r'[a-zA-Z_]\w*:', Name.Label),
-            (r'[a-zA-Z_\$]\w*', Name),
+            (r'(\.)((?:[^\W\d]|\$)[\w\$]*)', bygroups(Operator, Name.Attribute)),
+            (r'([^\W\d]|\$)[\w\$]*:', Name.Label),
+            (r'([^\W\d]|\$)[\w\$]*', Name),
             (r'[~\^\*!%&\[\]\(\)\{\}<>\|+=:;,./?-]', Operator),
             (r'[0-9][0-9]*\.[0-9]+([eE][0-9]+)?[fd]?', Number.Float),
             (r'0x[0-9a-fA-F]+', Number.Hex),
@@ -71,7 +71,7 @@ class JavaLexer(RegexLexer):
             (r'\n', Text)
         ],
         'class': [
-            (r'[a-zA-Z_]\w*', Name.Class, '#pop')
+            (r'([^\W\d]|\$)[\w\$]*', Name.Class, '#pop')
         ],
         'import': [
             (r'[\w.]+\*?', Name.Namespace, '#pop')
diff --git a/tests/examplefiles/example.java b/tests/examplefiles/example.java
new file mode 100644
index 00000000..f2e94322
--- /dev/null
+++ b/tests/examplefiles/example.java
@@ -0,0 +1,16 @@
+class _PostUnico$deClassá
+{void fo$o() {}
+
+  void PostUnicodeFunctioná() {
+  láb$el:
+    break láb$el;
+  
+  }
+}
+
+class áPreUnicode$Class
+{
+  public int $foo;
+  public int á$foo;
+  _PostUnico$deClassá áPreUnicodeFunction() { return null; }
+}