Bugfix in tools.nearley: Backslash now handled correctly in strings (Issue #341)

lark/tools/nearley.py

@@ -32,8 +32,8 @@ nearley_grammar = r"""
     NAME: /[a-zA-Z_$]\w*/
     COMMENT: /#[^\n]*/
     REGEXP: /\[.*?\]/
-    STRING: /".*?"/
 
+    %import common.ESCAPED_STRING -> STRING
     %import common.WS
     %ignore WS
     %ignore COMMENT

tests/test_nearley/test_nearley.py

@@ -75,6 +75,14 @@ class TestNearley(unittest.TestCase):
 
         parse(u'±a')
 
+    def test_backslash(self):
+        grammar = r'main -> "\""'
+        code = create_code_for_nearley_grammar(grammar, 'main', BUILTIN_PATH, './')
+        d = {}
+        exec (code, d)
+        parse = d['parse']
+        parse(u'"')
+
     def test_utf8_2(self):
         fn = os.path.join(TEST_PATH, 'grammars/unicode.ne')
         nearley_tool_main(fn, 'x', NEARLEY_PATH)

tests/test_reconstructor.py

@@ -28,7 +28,7 @@ class TestReconstructor(TestCase):
         item: NL
             | rule
         rule: WORD ":" NUMBER
-        NL: /(\\r?\\n)+\s*/
+        NL: /(\\r?\\n)+\\s*/
         """ + common
 
         code = """
@@ -42,7 +42,7 @@ class TestReconstructor(TestCase):
         g = """
         start: (rule | NL)*
         rule: WORD ":" NUMBER
-        NL: /(\\r?\\n)+\s*/
+        NL: /(\\r?\\n)+\\s*/
         """ + common
 
         code = """
@@ -59,7 +59,7 @@ class TestReconstructor(TestCase):
             | rule
             | "hello" -> hi
         rule: WORD ":" NUMBER
-        NL: /(\\r?\\n)+\s*/
+        NL: /(\\r?\\n)+\\s*/
         """ + common
 
         code = """