Bugfix in scanless

lark/load_grammar.py

@@ -357,6 +357,8 @@ class Grammar:
                         exp.children = list(interleave(exp.children, expr))
                         if r == start:
                             exp.children = [expr] + exp.children
+                    for exp in tree.find_data('expr'):
+                        exp.children[0] = T('expansion', list(interleave(exp.children[:1], expr)))
 
                 x = [T('expansion', [Token('RULE', x)]) for x in ignore_names]
                 _ignore_tree = T('expr', [T('expansions', x), Token('OP', '?')])

lark/parsers/earley.py

@@ -107,6 +107,7 @@ class Column:
 
 class Parser:
     def __init__(self, parser_conf):
+
         self.analysis = GrammarAnalyzer(parser_conf.rules, parser_conf.start)
         self.start = parser_conf.start
 
@@ -154,7 +155,7 @@ class Parser:
                         next_set.add([item.advance(stream[i])])
 
             if not next_set and token is not END_TOKEN:
-                expect = {i.expect for i in cur_set.to_scan}
+                expect = {i.expect[-1] for i in cur_set.to_scan}
                 raise UnexpectedToken(token, expect, stream, i)
 
             return cur_set, next_set

tests/__main__.py

@@ -5,7 +5,7 @@ import logging
 
 from .test_trees import TestTrees
 # from .test_selectors import TestSelectors
-from .test_parser import TestLalrStandard, TestEarleyStandard, TestLalrContextual, TestParsers, TestEarleyScanless
+from .test_parser import TestLalrStandard, TestEarleyStandard, TestLalrContextual, TestParsers, TestEarleyScanless, TestEarley
 # from .test_grammars import TestPythonG, TestConfigG
 
 logging.basicConfig(level=logging.INFO)

tests/test_parser.py

@@ -48,7 +48,7 @@ class TestEarley(unittest.TestCase):
                      B: ("ab"|/[^b]/)*
                   """, lexer=None)
 
-        assertEqual( g.parse('abc'), 'abc')
+        self.assertEqual( g.parse('abc').children[0], 'abc')
 
     def test_earley_scanless(self):
         g = Lark("""start: A "b" c
@@ -58,6 +58,20 @@ class TestEarley(unittest.TestCase):
         x = g.parse('aaaababc')
 
 
+    def test_earley_scanless2(self):
+        grammar = """
+        start: statement+
+
+        statement: "r"
+                 | "c" /[a-z]/+
+
+        %ignore " "
+        """
+
+        program = """c b r"""
+
+        l = Lark(grammar, parser='earley', lexer=None)
+        l.parse(program)
 
 def _make_parser_test(LEXER, PARSER):
     def _Lark(grammar, **kwargs):