Normalize symbol expressions

[fur] / tokenization.py

diff --git a/tokenization.py b/tokenization.py
index ff79307..3c4dc6f 100644 (file)
--- a/tokenization.py
+++ b/tokenization.py
@@ -31,18 +31,19 @@ def _make_token_matcher(definition):
 
     return token_matcher
 
-
 _TOKEN_MATCHERS = [
+    ('keyword',                         r'(def|end)(?![a-z_])'),
     ('open_parenthese',                 r'\('),
     ('close_parenthese',                r'\)'),
     ('comma',                           r','),
     ('integer_literal',                 r'\d+'),
-    ('symbol',                          r'[a-z]+'),
+    ('symbol',                          r'[a-z_]+'),
     ('single_quoted_string_literal',    r"'.*?'"),
-    ('equality_level_operator',         r'(<=|>=|==|!=|<|>)'),
+    ('comparison_level_operator',         r'(<=|>=|==|!=|<|>)'),
     ('assignment_operator',             r'='),
     ('addition_level_operator',         r'(\+|-)'),
     ('multiplication_level_operator',   r'(\*|//|%)'),
+    ('newline',                         r'\n'),
 ]
 
 _TOKEN_MATCHERS = list(map(_make_token_matcher, _TOKEN_MATCHERS))
@@ -57,6 +58,12 @@ def tokenize(source):
             index += 1
             continue
 
+        if source[index] == '#':
+            while index < len(source) and source[index] != '\n':
+                index += 1
+
+            continue
+
         success = False
 
         for matcher in _TOKEN_MATCHERS:
@@ -67,11 +74,13 @@ def tokenize(source):
                 break
 
         if not success:
-            raise Exception('Unexpected character "{}"'.format(source[index]))
+            raise Exception('Unexpected character "{}" on line {}'.format(
+                source[index],
+                line,
+            ))
 
-        while index < len(source) and source[index] in set(['\n']):
+        if token.type == 'newline':
             line += 1
-            index += 1
 
 if __name__ == '__main__':
     import unittest
@@ -202,7 +211,7 @@ if __name__ == '__main__':
             self.assertEqual(
                 tokenize('=='),
                 (Token(
-                    type='equality_level_operator',
+                    type='comparison_level_operator',
                     match='==',
                     index=0,
                     line=1,
@@ -213,7 +222,7 @@ if __name__ == '__main__':
             self.assertEqual(
                 tokenize('>='),
                 (Token(
-                    type='equality_level_operator',
+                    type='comparison_level_operator',
                     match='>=',
                     index=0,
                     line=1,
@@ -224,7 +233,7 @@ if __name__ == '__main__':
             self.assertEqual(
                 tokenize('<='),
                 (Token(
-                    type='equality_level_operator',
+                    type='comparison_level_operator',
                     match='<=',
                     index=0,
                     line=1,
@@ -235,7 +244,7 @@ if __name__ == '__main__':
             self.assertEqual(
                 tokenize('>'),
                 (Token(
-                    type='equality_level_operator',
+                    type='comparison_level_operator',
                     match='>',
                     index=0,
                     line=1,
@@ -246,7 +255,7 @@ if __name__ == '__main__':
             self.assertEqual(
                 tokenize('<'),
                 (Token(
-                    type='equality_level_operator',
+                    type='comparison_level_operator',
                     match='<',
                     index=0,
                     line=1,
@@ -257,19 +266,19 @@ if __name__ == '__main__':
             self.assertEqual(
                 tokenize('!='),
                 (Token(
-                    type='equality_level_operator',
+                    type='comparison_level_operator',
                     match='!=',
                     index=0,
                     line=1,
                 ),),
             )
 
-        def test_handles_trailing_newline(self):
+        def test_tokenizes_newline(self):
             self.assertEqual(
-                tokenize('print\n'),
+                tokenize('\n'),
                 (Token(
-                    type='symbol',
-                    match='print',
+                    type='newline',
+                    match='\n',
                     index=0,
                     line=1,
                 ),),
@@ -296,6 +305,12 @@ if __name__ == '__main__':
                         index=0,
                         line=1,
                     ),
+                    Token(
+                        type='newline',
+                        match='\n',
+                        index=5,
+                        line=1,
+                    ),
                     Token(
                         type='open_parenthese',
                         match='(',