Better error message for bad regexps (Issue #62)

lark/common.py

@@ -1,7 +1,8 @@
 import re
-import sre_parse
 import sys
 
+from .utils import get_regexp_width
+
 Py36 = (sys.version_info[:2] >= (3, 6))
 
 
@@ -95,10 +96,10 @@ class PatternRE(Pattern):
 
     @property
     def min_width(self):
-        return sre_parse.parse(self.to_regexp()).getwidth()[0]
+        return get_regexp_width(self.to_regexp())[0]
     @property
     def max_width(self):
-        return sre_parse.parse(self.to_regexp()).getwidth()[1]
+        return get_regexp_width(self.to_regexp())[1]
 
 class TokenDef(object):
     def __init__(self, name, pattern, priority=1):

lark/parser_frontends.py

@@ -1,5 +1,5 @@
 import re
-import sre_parse
+from .utils import get_regexp_width
 
 from .lexer import Lexer, ContextualLexer, Token
 
@@ -77,7 +77,7 @@ class Earley_NoLex:
         self.regexps = {}
         for t in lexer_conf.tokens:
             regexp = t.pattern.to_regexp()
-            width = sre_parse.parse(regexp).getwidth()
+            width = get_regexp_width(regexp)
             if width != (1,1):
                 raise GrammarError('Scanless parsing (lexer=None) requires all tokens to have a width of 1 (terminal %s: %s is %s)' % (sym, regexp, width))
             self.regexps[t.name] = re.compile(regexp)
@@ -121,7 +121,10 @@ class XEarley:
         self.regexps = {}
         for t in lexer_conf.tokens:
             regexp = t.pattern.to_regexp()
-            assert sre_parse.parse(regexp).getwidth()
+            try:
+                assert get_regexp_width(regexp)
+            except ValueError:
+                raise ValueError("Bad regexp in token %s: %s" % (t.name, regexp))
             self.regexps[t.name] = re.compile(regexp)
 
     def parse(self, text):

lark/utils.py

@@ -112,3 +112,10 @@ except NameError:
             return -1
 
 
+import sre_parse
+import sre_constants
+def get_regexp_width(regexp):
+    try:
+        return sre_parse.parse(regexp).getwidth()
+    except sre_constants.error:
+        raise ValueError(regexp)