diff options
author | Benji York <benji@benjiyork.com> | 2011-09-30 16:14:42 -0400 |
---|---|---|
committer | Benji York <benji@benjiyork.com> | 2011-09-30 16:14:42 -0400 |
commit | f83a828ec5ff775cd313213385d4bce06efda8e6 (patch) | |
tree | 1ac3412cf51663aaf78c7f0d73369c262f3d825d /tests/test_perllexer.py | |
parent | 2a7deac5ab1e25650a5eb71b134e25bfb8f8be1d (diff) | |
download | pygments-f83a828ec5ff775cd313213385d4bce06efda8e6.tar.gz |
fix regular expressions that can exhibit runaway backtracking (bug 697)
Diffstat (limited to 'tests/test_perllexer.py')
-rw-r--r-- | tests/test_perllexer.py | 137 |
1 files changed, 137 insertions, 0 deletions
diff --git a/tests/test_perllexer.py b/tests/test_perllexer.py new file mode 100644 index 00000000..4f99af6b --- /dev/null +++ b/tests/test_perllexer.py @@ -0,0 +1,137 @@ +# -*- coding: utf-8 -*- +""" + Pygments regex lexer tests + ~~~~~~~~~~~~~~~~~~~~~~~~~~ + + :copyright: Copyright 2011 by the Pygments team, see AUTHORS. + :license: BSD, see LICENSE for details. +""" + +import time +import unittest + +from pygments.token import String +from pygments.lexers.agile import PerlLexer + + +class RunawayRegexTest(unittest.TestCase): + # A previous version of the Perl lexer would spend a great deal of + # time backtracking when given particular strings. These tests show that + # the runaway backtracking doesn't happen any more (at least for the given + # cases). + + lexer = PerlLexer() + + ### Test helpers. + + def assert_single_token(self, s, token): + """Show that a given string generates only one token.""" + tokens = list(self.lexer.get_tokens_unprocessed(s)) + self.assertEqual(len(tokens), 1, tokens) + self.assertEqual(s, tokens[0][2]) + self.assertEqual(token, tokens[0][1]) + + def assert_tokens(self, strings, expected_tokens): + """Show that a given string generates the expected tokens.""" + tokens = list(self.lexer.get_tokens_unprocessed(''.join(strings))) + self.assertEqual(len(tokens), len(expected_tokens), tokens) + for index, s in enumerate(strings): + self.assertEqual(s, tokens[index][2]) + self.assertEqual(expected_tokens[index], tokens[index][1]) + + def assert_fast_tokenization(self, s): + """Show that a given string is tokenized quickly.""" + start = time.time() + tokens = list(self.lexer.get_tokens_unprocessed(s)) + end = time.time() + # Isn't 10 seconds kind of a long time? Yes, but we don't want false + # positives when the tests are starved for CPU time. + if end-start > 10: + self.fail('tokenization took too long') + return tokens + + ### Strings. + + def test_single_quote_strings(self): + self.assert_single_token(r"'foo\tbar\\\'baz'", String) + self.assert_fast_tokenization("'" + '\\'*999) + + def test_double_quote_strings(self): + self.assert_single_token(r'"foo\tbar\\\"baz"', String) + self.assert_fast_tokenization('"' + '\\'*999) + + def test_backtick_strings(self): + self.assert_single_token(r'`foo\tbar\\\`baz`', String.Backtick) + self.assert_fast_tokenization('`' + '\\'*999) + + ### Regex matches with various delimiters. + + def test_match(self): + self.assert_single_token(r'/aa\tbb/', String.Regex) + self.assert_fast_tokenization('/' + '\\'*999) + + def test_match_with_slash(self): + self.assert_tokens(['m', '/\n\\t\\\\/'], [String.Regex, String.Regex]) + self.assert_fast_tokenization('m/xxx\n' + '\\'*999) + + def test_match_with_bang(self): + self.assert_tokens(['m', r'!aa\t\!bb!'], [String.Regex, String.Regex]) + self.assert_fast_tokenization('m!' + '\\'*999) + + def test_match_with_brace(self): + self.assert_tokens(['m', r'{aa\t\}bb}'], [String.Regex, String.Regex]) + self.assert_fast_tokenization('m{' + '\\'*999) + + def test_match_with_angle_brackets(self): + self.assert_tokens(['m', r'<aa\t\>bb>'], [String.Regex, String.Regex]) + self.assert_fast_tokenization('m<' + '\\'*999) + + def test_match_with_parenthesis(self): + self.assert_tokens(['m', r'(aa\t\)bb)'], [String.Regex, String.Regex]) + self.assert_fast_tokenization('m(' + '\\'*999) + + def test_match_with_at_sign(self): + self.assert_tokens(['m', r'@aa\t\@bb@'], [String.Regex, String.Regex]) + self.assert_fast_tokenization('m@' + '\\'*999) + + def test_match_with_percent_sign(self): + self.assert_tokens(['m', r'%aa\t\%bb%'], [String.Regex, String.Regex]) + self.assert_fast_tokenization('m%' + '\\'*999) + + def test_match_with_dollar_sign(self): + self.assert_tokens(['m', r'$aa\t\$bb$'], [String.Regex, String.Regex]) + self.assert_fast_tokenization('m$' + '\\'*999) + + ### Regex substitutions with various delimeters. + + def test_substitution_with_slash(self): + self.assert_single_token('s/aaa/bbb/g', String.Regex) + self.assert_fast_tokenization('s/foo/' + '\\'*999) + + def test_substitution_with_at_sign(self): + self.assert_single_token(r's@aaa@bbb@g', String.Regex) + self.assert_fast_tokenization('s@foo@' + '\\'*999) + + def test_substitution_with_percent_sign(self): + self.assert_single_token(r's%aaa%bbb%g', String.Regex) + self.assert_fast_tokenization('s%foo%' + '\\'*999) + + def test_substitution_with_brace(self): + self.assert_single_token(r's{aaa}', String.Regex) + self.assert_fast_tokenization('s{' + '\\'*999) + + def test_substitution_with_angle_bracket(self): + self.assert_single_token(r's<aaa>', String.Regex) + self.assert_fast_tokenization('s<' + '\\'*999) + + def test_substitution_with_angle_bracket(self): + self.assert_single_token(r's<aaa>', String.Regex) + self.assert_fast_tokenization('s<' + '\\'*999) + + def test_substitution_with_square_bracket(self): + self.assert_single_token(r's[aaa]', String.Regex) + self.assert_fast_tokenization('s[' + '\\'*999) + + def test_substitution_with_parenthesis(self): + self.assert_single_token(r's(aaa)', String.Regex) + self.assert_fast_tokenization('s(' + '\\'*999) |