Add support for unicode letters in Java identifiers

Additionally, allow $ in middle of variable names

1 files changed, 10 insertions, 10 deletions

diff --git a/pygments/lexers/jvm.py b/pygments/lexers/jvm.py
index 06dcc81b..b2fda960 100644
--- a/pygments/lexers/jvm.py
+++ b/pygments/lexers/jvm.py
@@ -34,14 +34,14 @@ 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_][a-zA-Z0-9_\.]*', Name.Decorator),
+            (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),
@@ -51,9 +51,9 @@ class JavaLexer(RegexLexer):
             (r'(boolean|byte|char|double|float|int|long|short|void)\b',
              Keyword.Type),
             # method names
-            (r'^(\s*(?:[a-zA-Z_][a-zA-Z0-9_\.\[\]<>]*\s+)+?)' # return arguments
-             r'([a-zA-Z_][a-zA-Z0-9_]*)'                      # method name
-             r'(\s*)(\()',                                    # signature start
+            (r'^(\s*(?:[^\W\d][\w\.\[\]<>]*\s+)+?)' # return arguments
+             r'([^\W\d]\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),
@@ -61,9 +61,9 @@ class JavaLexer(RegexLexer):
             (r'(import)(\s+)', bygroups(Keyword.Namespace, Text), 'import'),
             (r'"(\\\\|\\"|[^"])*"', String),
             (r"'\\.'|'[^\\]'|'\\u[0-9a-fA-F]{4}'", String.Char),
-            (r'(\.)([a-zA-Z_][a-zA-Z0-9_]*)', bygroups(Operator, Name.Attribute)),
-            (r'[a-zA-Z_][a-zA-Z0-9_]*:', Name.Label),
-            (r'[a-zA-Z_\$][a-zA-Z0-9_]*', 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),
@@ -71,10 +71,10 @@ class JavaLexer(RegexLexer):
             (r'\n', Text)
         ],
         'class': [
-            (r'[a-zA-Z_][a-zA-Z0-9_]*', Name.Class, '#pop')
+            (r'[^\W\d]\w*', Name.Class, '#pop')
         ],
         'import': [
-            (r'[a-zA-Z0-9_.]+\*?', Name.Namespace, '#pop')
+            (r'[\w.]+\*?', Name.Namespace, '#pop')
         ],
     }