annotate Lib/test/test_re.py @ 79675:c0a2294de787 2.7

Issue #16199: Backport 2to3 typo fix from 3.2.
author Chris Jerdonek <chris.jerdonek@gmail.com>
date Thu, 11 Oct 2012 16:15:46 -0700
parents cc48eef234dd
children 5b5c2c3b2268
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
63634
762d89faceaa Merged revisions 79539 via svnmerge from
Ezio Melotti <ezio.melotti@gmail.com>
parents: 58453
diff changeset
1 from test.test_support import verbose, run_unittest, import_module
5774
3ef2563e83ee test suite for re.py
Guido van Rossum <guido@python.org>
parents:
diff changeset
2 import re
37017
d68344d932c4 Rename sre.py -> re.py
Neal Norwitz <nnorwitz@gmail.com>
parents: 35591
diff changeset
3 from re import Scanner
68927
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
4 import sys
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
5 import string
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
6 import traceback
32160
aec9c98a47d2 Add weakref support to sockets and re pattern objects.
Raymond Hettinger <python@rcn.com>
parents: 32006
diff changeset
7 from weakref import proxy
5774
3ef2563e83ee test suite for re.py
Guido van Rossum <guido@python.org>
parents:
diff changeset
8
5821
78321a84d73e Added tests for sub, subn, and split.
Guido van Rossum <guido@python.org>
parents: 5805
diff changeset
9 # Misc tests from Tim Peters' re.doc
78321a84d73e Added tests for sub, subn, and split.
Guido van Rossum <guido@python.org>
parents: 5805
diff changeset
10
29742
68835bf101ec fixed typo in comment
Just van Rossum <just@letterror.com>
parents: 29548
diff changeset
11 # WARNING: Don't change details in these tests if you don't know
68503
ee5350f1d0c2 #11515: fix several typos. Patch by Piotr Kasprzyk.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68256
diff changeset
12 # what you're doing. Some of these tests were carefully modeled to
29548
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
13 # cover most of the code.
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
14
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
15 import unittest
7852
f34a658e7c15 AMK's latest
Guido van Rossum <guido@python.org>
parents: 7747
diff changeset
16
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
17 class ReTests(unittest.TestCase):
32160
aec9c98a47d2 Add weakref support to sockets and re pattern objects.
Raymond Hettinger <python@rcn.com>
parents: 32006
diff changeset
18
aec9c98a47d2 Add weakref support to sockets and re pattern objects.
Raymond Hettinger <python@rcn.com>
parents: 32006
diff changeset
19 def test_weakref(self):
aec9c98a47d2 Add weakref support to sockets and re pattern objects.
Raymond Hettinger <python@rcn.com>
parents: 32006
diff changeset
20 s = 'QabbbcR'
aec9c98a47d2 Add weakref support to sockets and re pattern objects.
Raymond Hettinger <python@rcn.com>
parents: 32006
diff changeset
21 x = re.compile('ab+c')
aec9c98a47d2 Add weakref support to sockets and re pattern objects.
Raymond Hettinger <python@rcn.com>
parents: 32006
diff changeset
22 y = proxy(x)
aec9c98a47d2 Add weakref support to sockets and re pattern objects.
Raymond Hettinger <python@rcn.com>
parents: 32006
diff changeset
23 self.assertEqual(x.findall('QabbbcR'), y.findall('QabbbcR'))
aec9c98a47d2 Add weakref support to sockets and re pattern objects.
Raymond Hettinger <python@rcn.com>
parents: 32006
diff changeset
24
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
25 def test_search_star_plus(self):
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
26 self.assertEqual(re.search('x*', 'axx').span(0), (0, 0))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
27 self.assertEqual(re.search('x*', 'axx').span(), (0, 0))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
28 self.assertEqual(re.search('x+', 'axx').span(0), (1, 3))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
29 self.assertEqual(re.search('x+', 'axx').span(), (1, 3))
28843
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
30 self.assertEqual(re.search('x', 'aaa'), None)
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
31 self.assertEqual(re.match('a*', 'xxx').span(0), (0, 0))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
32 self.assertEqual(re.match('a*', 'xxx').span(), (0, 0))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
33 self.assertEqual(re.match('x*', 'xxxa').span(0), (0, 3))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
34 self.assertEqual(re.match('x*', 'xxxa').span(), (0, 3))
28843
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
35 self.assertEqual(re.match('a+', 'xxx'), None)
7852
f34a658e7c15 AMK's latest
Guido van Rossum <guido@python.org>
parents: 7747
diff changeset
36
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
37 def bump_num(self, matchobj):
7747
aedd2d0dcb31 Mass check-in after untabifying all files that need it.
Guido van Rossum <guido@python.org>
parents: 7098
diff changeset
38 int_value = int(matchobj.group(0))
aedd2d0dcb31 Mass check-in after untabifying all files that need it.
Guido van Rossum <guido@python.org>
parents: 7098
diff changeset
39 return str(int_value + 1)
5821
78321a84d73e Added tests for sub, subn, and split.
Guido van Rossum <guido@python.org>
parents: 5805
diff changeset
40
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
41 def test_basic_re_sub(self):
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
42 self.assertEqual(re.sub("(?i)b+", "x", "bbbb BBBB"), 'x x')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
43 self.assertEqual(re.sub(r'\d+', self.bump_num, '08.2 -2 23x99y'),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
44 '9.3 -3 24x100y')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
45 self.assertEqual(re.sub(r'\d+', self.bump_num, '08.2 -2 23x99y', 3),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
46 '9.3 -3 23x99y')
6777
daa77d9af34c New re version from AMK
Guido van Rossum <guido@python.org>
parents: 6447
diff changeset
47
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
48 self.assertEqual(re.sub('.', lambda m: r"\n", 'x'), '\\n')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
49 self.assertEqual(re.sub('.', r"\n", 'x'), '\n')
6004
3f59b34eb90f 1.5a3 prerelease 1 from AMK
Guido van Rossum <guido@python.org>
parents: 5831
diff changeset
50
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
51 s = r"\1\1"
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
52 self.assertEqual(re.sub('(.)', s, 'x'), 'xx')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
53 self.assertEqual(re.sub('(.)', re.escape(s), 'x'), s)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
54 self.assertEqual(re.sub('(.)', lambda m: s, 'x'), s)
5821
78321a84d73e Added tests for sub, subn, and split.
Guido van Rossum <guido@python.org>
parents: 5805
diff changeset
55
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
56 self.assertEqual(re.sub('(?P<a>x)', '\g<a>\g<a>', 'xx'), 'xxxx')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
57 self.assertEqual(re.sub('(?P<a>x)', '\g<a>\g<1>', 'xx'), 'xxxx')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
58 self.assertEqual(re.sub('(?P<unk>x)', '\g<unk>\g<unk>', 'xx'), 'xxxx')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
59 self.assertEqual(re.sub('(?P<unk>x)', '\g<1>\g<1>', 'xx'), 'xxxx')
7852
f34a658e7c15 AMK's latest
Guido van Rossum <guido@python.org>
parents: 7747
diff changeset
60
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
61 self.assertEqual(re.sub('a',r'\t\n\v\r\f\a\b\B\Z\a\A\w\W\s\S\d\D','a'),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
62 '\t\n\v\r\f\a\b\\B\\Z\a\\A\\w\\W\\s\\S\\d\\D')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
63 self.assertEqual(re.sub('a', '\t\n\v\r\f\a', 'a'), '\t\n\v\r\f\a')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
64 self.assertEqual(re.sub('a', '\t\n\v\r\f\a', 'a'),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
65 (chr(9)+chr(10)+chr(11)+chr(13)+chr(12)+chr(7)))
7098
1676fad7b63e Added tests for qualified sub and split
Guido van Rossum <guido@python.org>
parents: 7002
diff changeset
66
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
67 self.assertEqual(re.sub('^\s*', 'X', 'test'), 'Xtest')
5831
985b7bf4755b Some new tests by Jeffrey
Guido van Rossum <guido@python.org>
parents: 5821
diff changeset
68
28826
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
69 def test_bug_449964(self):
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
70 # fails for group followed by other escape
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
71 self.assertEqual(re.sub(r'(?P<unk>x)', '\g<1>\g<1>\\b', 'xx'),
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
72 'xx\bxx\b')
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
73
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
74 def test_bug_449000(self):
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
75 # Test for sub() on escaped characters
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
76 self.assertEqual(re.sub(r'\r\n', r'\n', 'abc\r\ndef\r\n'),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
77 'abc\ndef\n')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
78 self.assertEqual(re.sub('\r\n', r'\n', 'abc\r\ndef\r\n'),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
79 'abc\ndef\n')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
80 self.assertEqual(re.sub(r'\r\n', '\n', 'abc\r\ndef\r\n'),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
81 'abc\ndef\n')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
82 self.assertEqual(re.sub('\r\n', '\n', 'abc\r\ndef\r\n'),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
83 'abc\ndef\n')
5831
985b7bf4755b Some new tests by Jeffrey
Guido van Rossum <guido@python.org>
parents: 5821
diff changeset
84
43381
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
85 def test_bug_1140(self):
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
86 # re.sub(x, y, u'') should return u'', not '', and
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
87 # re.sub(x, y, '') should return '', not u''.
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
88 # Also:
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
89 # re.sub(x, y, unicode(x)) should return unicode(y), and
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
90 # re.sub(x, y, str(x)) should return
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
91 # str(y) if isinstance(y, str) else unicode(y).
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
92 for x in 'x', u'x':
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
93 for y in 'y', u'y':
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
94 z = re.sub(x, y, u'')
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
95 self.assertEqual(z, u'')
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
96 self.assertEqual(type(z), unicode)
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
97 #
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
98 z = re.sub(x, y, '')
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
99 self.assertEqual(z, '')
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
100 self.assertEqual(type(z), str)
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
101 #
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
102 z = re.sub(x, y, unicode(x))
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
103 self.assertEqual(z, y)
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
104 self.assertEqual(type(z), unicode)
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
105 #
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
106 z = re.sub(x, y, str(x))
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
107 self.assertEqual(z, y)
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
108 self.assertEqual(type(z), type(y))
9cc132c8461e issue1140: Guido's patch from revision 58098 (2.6) applied to 2.5.
Sean Reifscheider <jafo@tummy.com>
parents: 41533
diff changeset
109
44354
b1347c68d456 Fix issue 1661: Flags argument silently ignored in re functions with compiled regexes.
Raymond Hettinger <python@rcn.com>
parents: 43326
diff changeset
110 def test_bug_1661(self):
b1347c68d456 Fix issue 1661: Flags argument silently ignored in re functions with compiled regexes.
Raymond Hettinger <python@rcn.com>
parents: 43326
diff changeset
111 # Verify that flags do not get silently ignored with compiled patterns
b1347c68d456 Fix issue 1661: Flags argument silently ignored in re functions with compiled regexes.
Raymond Hettinger <python@rcn.com>
parents: 43326
diff changeset
112 pattern = re.compile('.')
b1347c68d456 Fix issue 1661: Flags argument silently ignored in re functions with compiled regexes.
Raymond Hettinger <python@rcn.com>
parents: 43326
diff changeset
113 self.assertRaises(ValueError, re.match, pattern, 'A', re.I)
b1347c68d456 Fix issue 1661: Flags argument silently ignored in re functions with compiled regexes.
Raymond Hettinger <python@rcn.com>
parents: 43326
diff changeset
114 self.assertRaises(ValueError, re.search, pattern, 'A', re.I)
b1347c68d456 Fix issue 1661: Flags argument silently ignored in re functions with compiled regexes.
Raymond Hettinger <python@rcn.com>
parents: 43326
diff changeset
115 self.assertRaises(ValueError, re.findall, pattern, 'A', re.I)
b1347c68d456 Fix issue 1661: Flags argument silently ignored in re functions with compiled regexes.
Raymond Hettinger <python@rcn.com>
parents: 43326
diff changeset
116 self.assertRaises(ValueError, re.compile, pattern, re.I)
b1347c68d456 Fix issue 1661: Flags argument silently ignored in re functions with compiled regexes.
Raymond Hettinger <python@rcn.com>
parents: 43326
diff changeset
117
49126
95ac698358ab - Issue #3629: Fix sre "bytecode" validator for an end case.
Guido van Rossum <guido@python.org>
parents: 49099
diff changeset
118 def test_bug_3629(self):
95ac698358ab - Issue #3629: Fix sre "bytecode" validator for an end case.
Guido van Rossum <guido@python.org>
parents: 49099
diff changeset
119 # A regex that triggered a bug in the sre-code validator
95ac698358ab - Issue #3629: Fix sre "bytecode" validator for an end case.
Guido van Rossum <guido@python.org>
parents: 49099
diff changeset
120 re.compile("(?P<quote>)(?(quote))")
95ac698358ab - Issue #3629: Fix sre "bytecode" validator for an end case.
Guido van Rossum <guido@python.org>
parents: 49099
diff changeset
121
33397
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
122 def test_sub_template_numeric_escape(self):
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
123 # bug 776311 and friends
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
124 self.assertEqual(re.sub('x', r'\0', 'x'), '\0')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
125 self.assertEqual(re.sub('x', r'\000', 'x'), '\000')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
126 self.assertEqual(re.sub('x', r'\001', 'x'), '\001')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
127 self.assertEqual(re.sub('x', r'\008', 'x'), '\0' + '8')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
128 self.assertEqual(re.sub('x', r'\009', 'x'), '\0' + '9')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
129 self.assertEqual(re.sub('x', r'\111', 'x'), '\111')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
130 self.assertEqual(re.sub('x', r'\117', 'x'), '\117')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
131
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
132 self.assertEqual(re.sub('x', r'\1111', 'x'), '\1111')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
133 self.assertEqual(re.sub('x', r'\1111', 'x'), '\111' + '1')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
134
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
135 self.assertEqual(re.sub('x', r'\00', 'x'), '\x00')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
136 self.assertEqual(re.sub('x', r'\07', 'x'), '\x07')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
137 self.assertEqual(re.sub('x', r'\08', 'x'), '\0' + '8')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
138 self.assertEqual(re.sub('x', r'\09', 'x'), '\0' + '9')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
139 self.assertEqual(re.sub('x', r'\0a', 'x'), '\0' + 'a')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
140
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
141 self.assertEqual(re.sub('x', r'\400', 'x'), '\0')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
142 self.assertEqual(re.sub('x', r'\777', 'x'), '\377')
33463
28ef5fe0b8b5 Whitespace normalization.
Tim Peters <tim.peters@gmail.com>
parents: 33398
diff changeset
143
33397
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
144 self.assertRaises(re.error, re.sub, 'x', r'\1', 'x')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
145 self.assertRaises(re.error, re.sub, 'x', r'\8', 'x')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
146 self.assertRaises(re.error, re.sub, 'x', r'\9', 'x')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
147 self.assertRaises(re.error, re.sub, 'x', r'\11', 'x')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
148 self.assertRaises(re.error, re.sub, 'x', r'\18', 'x')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
149 self.assertRaises(re.error, re.sub, 'x', r'\1a', 'x')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
150 self.assertRaises(re.error, re.sub, 'x', r'\90', 'x')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
151 self.assertRaises(re.error, re.sub, 'x', r'\99', 'x')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
152 self.assertRaises(re.error, re.sub, 'x', r'\118', 'x') # r'\11' + '8'
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
153 self.assertRaises(re.error, re.sub, 'x', r'\11a', 'x')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
154 self.assertRaises(re.error, re.sub, 'x', r'\181', 'x') # r'\18' + '1'
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
155 self.assertRaises(re.error, re.sub, 'x', r'\800', 'x') # r'\80' + '0'
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
156
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
157 # in python2.3 (etc), these loop endlessly in sre_parser.py
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
158 self.assertEqual(re.sub('(((((((((((x)))))))))))', r'\11', 'x'), 'x')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
159 self.assertEqual(re.sub('((((((((((y))))))))))(.)', r'\118', 'xyz'),
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
160 'xz8')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
161 self.assertEqual(re.sub('((((((((((y))))))))))(.)', r'\11a', 'xyz'),
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
162 'xza')
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
163
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
164 def test_qualified_re_sub(self):
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
165 self.assertEqual(re.sub('a', 'b', 'aaaaa'), 'bbbbb')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
166 self.assertEqual(re.sub('a', 'b', 'aaaaa', 1), 'baaaa')
5831
985b7bf4755b Some new tests by Jeffrey
Guido van Rossum <guido@python.org>
parents: 5821
diff changeset
167
28826
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
168 def test_bug_114660(self):
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
169 self.assertEqual(re.sub(r'(\S)\s+(\S)', r'\1 \2', 'hello there'),
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
170 'hello there')
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
171
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
172 def test_bug_462270(self):
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
173 # Test for empty sub() behaviour, see SF bug #462270
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
174 self.assertEqual(re.sub('x*', '-', 'abxd'), '-a-b-d-')
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
175 self.assertEqual(re.sub('x+', '-', 'abxd'), 'ab-d')
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
176
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
177 def test_symbolic_refs(self):
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
178 self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<a', 'xx')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
179 self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<', 'xx')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
180 self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g', 'xx')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
181 self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<a a>', 'xx')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
182 self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<1a1>', 'xx')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
183 self.assertRaises(IndexError, re.sub, '(?P<a>x)', '\g<ab>', 'xx')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
184 self.assertRaises(re.error, re.sub, '(?P<a>x)|(?P<b>y)', '\g<b>', 'xx')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
185 self.assertRaises(re.error, re.sub, '(?P<a>x)|(?P<b>y)', '\\2', 'xx')
33397
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
186 self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<-1>', 'xx')
5831
985b7bf4755b Some new tests by Jeffrey
Guido van Rossum <guido@python.org>
parents: 5821
diff changeset
187
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
188 def test_re_subn(self):
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
189 self.assertEqual(re.subn("(?i)b+", "x", "bbbb BBBB"), ('x x', 2))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
190 self.assertEqual(re.subn("b+", "x", "bbbb BBBB"), ('x BBBB', 1))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
191 self.assertEqual(re.subn("b+", "x", "xyz"), ('xyz', 0))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
192 self.assertEqual(re.subn("b*", "x", "xyz"), ('xxxyxzx', 4))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
193 self.assertEqual(re.subn("b*", "x", "xyz", 2), ('xxxyz', 2))
5831
985b7bf4755b Some new tests by Jeffrey
Guido van Rossum <guido@python.org>
parents: 5821
diff changeset
194
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
195 def test_re_split(self):
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
196 self.assertEqual(re.split(":", ":a:b::c"), ['', 'a', 'b', '', 'c'])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
197 self.assertEqual(re.split(":*", ":a:b::c"), ['', 'a', 'b', 'c'])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
198 self.assertEqual(re.split("(:*)", ":a:b::c"),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
199 ['', ':', 'a', ':', 'b', '::', 'c'])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
200 self.assertEqual(re.split("(?::*)", ":a:b::c"), ['', 'a', 'b', 'c'])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
201 self.assertEqual(re.split("(:)*", ":a:b::c"),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
202 ['', ':', 'a', ':', 'b', ':', 'c'])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
203 self.assertEqual(re.split("([b:]+)", ":a:b::c"),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
204 ['', ':', 'a', ':b::', 'c'])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
205 self.assertEqual(re.split("(b)|(:+)", ":a:b::c"),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
206 ['', None, ':', 'a', None, ':', '', 'b', None, '',
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
207 None, '::', 'c'])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
208 self.assertEqual(re.split("(?:b)|(?::+)", ":a:b::c"),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
209 ['', 'a', '', '', 'c'])
5831
985b7bf4755b Some new tests by Jeffrey
Guido van Rossum <guido@python.org>
parents: 5821
diff changeset
210
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
211 def test_qualified_re_split(self):
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
212 self.assertEqual(re.split(":", ":a:b::c", 2), ['', 'a', 'b::c'])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
213 self.assertEqual(re.split(':', 'a:b:c:d', 2), ['a', 'b', 'c:d'])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
214 self.assertEqual(re.split("(:)", ":a:b::c", 2),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
215 ['', ':', 'a', ':', 'b::c'])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
216 self.assertEqual(re.split("(:*)", ":a:b::c", 2),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
217 ['', ':', 'a', ':', 'b::c'])
13895
c3320d24d7de -- whitespace cleanup (more tests to be added in the next commit)
Fredrik Lundh <fredrik@pythonware.com>
parents: 13821
diff changeset
218
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
219 def test_re_findall(self):
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
220 self.assertEqual(re.findall(":+", "abc"), [])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
221 self.assertEqual(re.findall(":+", "a:b::c:::d"), [":", "::", ":::"])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
222 self.assertEqual(re.findall("(:+)", "a:b::c:::d"), [":", "::", ":::"])
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
223 self.assertEqual(re.findall("(:)(:*)", "a:b::c:::d"), [(":", ""),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
224 (":", ":"),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
225 (":", "::")])
5821
78321a84d73e Added tests for sub, subn, and split.
Guido van Rossum <guido@python.org>
parents: 5805
diff changeset
226
28843
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
227 def test_bug_117612(self):
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
228 self.assertEqual(re.findall(r"(a|(b))", "aba"),
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
229 [("a", ""),("b", "b"),("a", "")])
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
230
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
231 def test_re_match(self):
28843
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
232 self.assertEqual(re.match('a', 'a').groups(), ())
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
233 self.assertEqual(re.match('(a)', 'a').groups(), ('a',))
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
234 self.assertEqual(re.match(r'(a)', 'a').group(0), 'a')
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
235 self.assertEqual(re.match(r'(a)', 'a').group(1), 'a')
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
236 self.assertEqual(re.match(r'(a)', 'a').group(1, 1), ('a', 'a'))
7098
1676fad7b63e Added tests for qualified sub and split
Guido van Rossum <guido@python.org>
parents: 7002
diff changeset
237
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
238 pat = re.compile('((a)|(b))(c)?')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
239 self.assertEqual(pat.match('a').groups(), ('a', 'a', None, None))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
240 self.assertEqual(pat.match('b').groups(), ('b', None, 'b', None))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
241 self.assertEqual(pat.match('ac').groups(), ('a', 'a', None, 'c'))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
242 self.assertEqual(pat.match('bc').groups(), ('b', None, 'b', 'c'))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
243 self.assertEqual(pat.match('bc').groups(""), ('b', "", 'b', 'c'))
8679
ba744462e958 Added tests for findall().
Guido van Rossum <guido@python.org>
parents: 8551
diff changeset
244
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
245 # A single group
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
246 m = re.match('(a)', 'a')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
247 self.assertEqual(m.group(0), 'a')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
248 self.assertEqual(m.group(0), 'a')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
249 self.assertEqual(m.group(1), 'a')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
250 self.assertEqual(m.group(1, 1), ('a', 'a'))
8679
ba744462e958 Added tests for findall().
Guido van Rossum <guido@python.org>
parents: 8551
diff changeset
251
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
252 pat = re.compile('(?:(?P<a1>a)|(?P<b2>b))(?P<c3>c)?')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
253 self.assertEqual(pat.match('a').group(1, 2, 3), ('a', None, None))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
254 self.assertEqual(pat.match('b').group('a1', 'b2', 'c3'),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
255 (None, 'b', None))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
256 self.assertEqual(pat.match('ac').group(1, 'b2', 3), ('a', None, 'c'))
7852
f34a658e7c15 AMK's latest
Guido van Rossum <guido@python.org>
parents: 7747
diff changeset
257
29548
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
258 def test_re_groupref_exists(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
259 self.assertEqual(re.match('^(\()?([^()]+)(?(1)\))$', '(a)').groups(),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
260 ('(', 'a'))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
261 self.assertEqual(re.match('^(\()?([^()]+)(?(1)\))$', 'a').groups(),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
262 (None, 'a'))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
263 self.assertEqual(re.match('^(\()?([^()]+)(?(1)\))$', 'a)'), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
264 self.assertEqual(re.match('^(\()?([^()]+)(?(1)\))$', '(a'), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
265 self.assertEqual(re.match('^(?:(a)|c)((?(1)b|d))$', 'ab').groups(),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
266 ('a', 'b'))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
267 self.assertEqual(re.match('^(?:(a)|c)((?(1)b|d))$', 'cd').groups(),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
268 (None, 'd'))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
269 self.assertEqual(re.match('^(?:(a)|c)((?(1)|d))$', 'cd').groups(),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
270 (None, 'd'))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
271 self.assertEqual(re.match('^(?:(a)|c)((?(1)|d))$', 'a').groups(),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
272 ('a', ''))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
273
35002
1ca7ae48a931 fix exceptions.TabError. _Please_ run test suite before checking in.
Anthony Baxter <anthonybaxter@gmail.com>
parents: 34993
diff changeset
274 # Tests for bug #1177831: exercise groups other than the first group
1ca7ae48a931 fix exceptions.TabError. _Please_ run test suite before checking in.
Anthony Baxter <anthonybaxter@gmail.com>
parents: 34993
diff changeset
275 p = re.compile('(?P<g1>a)(?P<g2>b)?((?(g2)c|d))')
1ca7ae48a931 fix exceptions.TabError. _Please_ run test suite before checking in.
Anthony Baxter <anthonybaxter@gmail.com>
parents: 34993
diff changeset
276 self.assertEqual(p.match('abc').groups(),
1ca7ae48a931 fix exceptions.TabError. _Please_ run test suite before checking in.
Anthony Baxter <anthonybaxter@gmail.com>
parents: 34993
diff changeset
277 ('a', 'b', 'c'))
1ca7ae48a931 fix exceptions.TabError. _Please_ run test suite before checking in.
Anthony Baxter <anthonybaxter@gmail.com>
parents: 34993
diff changeset
278 self.assertEqual(p.match('ad').groups(),
1ca7ae48a931 fix exceptions.TabError. _Please_ run test suite before checking in.
Anthony Baxter <anthonybaxter@gmail.com>
parents: 34993
diff changeset
279 ('a', None, 'd'))
1ca7ae48a931 fix exceptions.TabError. _Please_ run test suite before checking in.
Anthony Baxter <anthonybaxter@gmail.com>
parents: 34993
diff changeset
280 self.assertEqual(p.match('abd'), None)
1ca7ae48a931 fix exceptions.TabError. _Please_ run test suite before checking in.
Anthony Baxter <anthonybaxter@gmail.com>
parents: 34993
diff changeset
281 self.assertEqual(p.match('ac'), None)
1ca7ae48a931 fix exceptions.TabError. _Please_ run test suite before checking in.
Anthony Baxter <anthonybaxter@gmail.com>
parents: 34993
diff changeset
282
34993
a2b2227773e7 [Bug #1177831] Fix (?(id)yes|no) for a group other than the first one, and add a test case
Andrew M. Kuchling <amk@amk.ca>
parents: 33463
diff changeset
283
29548
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
284 def test_re_groupref(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
285 self.assertEqual(re.match(r'^(\|)?([^()]+)\1$', '|a|').groups(),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
286 ('|', 'a'))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
287 self.assertEqual(re.match(r'^(\|)?([^()]+)\1?$', 'a').groups(),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
288 (None, 'a'))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
289 self.assertEqual(re.match(r'^(\|)?([^()]+)\1$', 'a|'), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
290 self.assertEqual(re.match(r'^(\|)?([^()]+)\1$', '|a'), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
291 self.assertEqual(re.match(r'^(?:(a)|c)(\1)$', 'aa').groups(),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
292 ('a', 'a'))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
293 self.assertEqual(re.match(r'^(?:(a)|c)(\1)?$', 'c').groups(),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
294 (None, None))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
295
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
296 def test_groupdict(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
297 self.assertEqual(re.match('(?P<first>first) (?P<second>second)',
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
298 'first second').groupdict(),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
299 {'first':'first', 'second':'second'})
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
300
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
301 def test_expand(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
302 self.assertEqual(re.match("(?P<first>first) (?P<second>second)",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
303 "first second")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
304 .expand(r"\2 \1 \g<second> \g<first>"),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
305 "second first second first")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
306
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
307 def test_repeat_minmax(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
308 self.assertEqual(re.match("^(\w){1}$", "abc"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
309 self.assertEqual(re.match("^(\w){1}?$", "abc"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
310 self.assertEqual(re.match("^(\w){1,2}$", "abc"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
311 self.assertEqual(re.match("^(\w){1,2}?$", "abc"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
312
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
313 self.assertEqual(re.match("^(\w){3}$", "abc").group(1), "c")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
314 self.assertEqual(re.match("^(\w){1,3}$", "abc").group(1), "c")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
315 self.assertEqual(re.match("^(\w){1,4}$", "abc").group(1), "c")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
316 self.assertEqual(re.match("^(\w){3,4}?$", "abc").group(1), "c")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
317 self.assertEqual(re.match("^(\w){3}?$", "abc").group(1), "c")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
318 self.assertEqual(re.match("^(\w){1,3}?$", "abc").group(1), "c")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
319 self.assertEqual(re.match("^(\w){1,4}?$", "abc").group(1), "c")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
320 self.assertEqual(re.match("^(\w){3,4}?$", "abc").group(1), "c")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
321
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
322 self.assertEqual(re.match("^x{1}$", "xxx"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
323 self.assertEqual(re.match("^x{1}?$", "xxx"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
324 self.assertEqual(re.match("^x{1,2}$", "xxx"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
325 self.assertEqual(re.match("^x{1,2}?$", "xxx"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
326
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
327 self.assertNotEqual(re.match("^x{3}$", "xxx"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
328 self.assertNotEqual(re.match("^x{1,3}$", "xxx"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
329 self.assertNotEqual(re.match("^x{1,4}$", "xxx"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
330 self.assertNotEqual(re.match("^x{3,4}?$", "xxx"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
331 self.assertNotEqual(re.match("^x{3}?$", "xxx"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
332 self.assertNotEqual(re.match("^x{1,3}?$", "xxx"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
333 self.assertNotEqual(re.match("^x{1,4}?$", "xxx"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
334 self.assertNotEqual(re.match("^x{3,4}?$", "xxx"), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
335
35591
9a2e20318c9f Bug #1202493: Fixing SRE parser to handle '{}' as perl does, rather than
Georg Brandl <georg@python.org>
parents: 35002
diff changeset
336 self.assertEqual(re.match("^x{}$", "xxx"), None)
9a2e20318c9f Bug #1202493: Fixing SRE parser to handle '{}' as perl does, rather than
Georg Brandl <georg@python.org>
parents: 35002
diff changeset
337 self.assertNotEqual(re.match("^x{}$", "x{}"), None)
9a2e20318c9f Bug #1202493: Fixing SRE parser to handle '{}' as perl does, rather than
Georg Brandl <georg@python.org>
parents: 35002
diff changeset
338
29548
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
339 def test_getattr(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
340 self.assertEqual(re.match("(a)", "a").pos, 0)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
341 self.assertEqual(re.match("(a)", "a").endpos, 1)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
342 self.assertEqual(re.match("(a)", "a").string, "a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
343 self.assertEqual(re.match("(a)", "a").regs, ((0, 1), (0, 1)))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
344 self.assertNotEqual(re.match("(a)", "a").re, None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
345
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
346 def test_special_escapes(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
347 self.assertEqual(re.search(r"\b(b.)\b",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
348 "abcd abc bcd bx").group(1), "bx")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
349 self.assertEqual(re.search(r"\B(b.)\B",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
350 "abc bcd bc abxd").group(1), "bx")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
351 self.assertEqual(re.search(r"\b(b.)\b",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
352 "abcd abc bcd bx", re.LOCALE).group(1), "bx")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
353 self.assertEqual(re.search(r"\B(b.)\B",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
354 "abc bcd bc abxd", re.LOCALE).group(1), "bx")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
355 self.assertEqual(re.search(r"\b(b.)\b",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
356 "abcd abc bcd bx", re.UNICODE).group(1), "bx")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
357 self.assertEqual(re.search(r"\B(b.)\B",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
358 "abc bcd bc abxd", re.UNICODE).group(1), "bx")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
359 self.assertEqual(re.search(r"^abc$", "\nabc\n", re.M).group(0), "abc")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
360 self.assertEqual(re.search(r"^\Aabc\Z$", "abc", re.M).group(0), "abc")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
361 self.assertEqual(re.search(r"^\Aabc\Z$", "\nabc\n", re.M), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
362 self.assertEqual(re.search(r"\b(b.)\b",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
363 u"abcd abc bcd bx").group(1), "bx")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
364 self.assertEqual(re.search(r"\B(b.)\B",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
365 u"abc bcd bc abxd").group(1), "bx")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
366 self.assertEqual(re.search(r"^abc$", u"\nabc\n", re.M).group(0), "abc")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
367 self.assertEqual(re.search(r"^\Aabc\Z$", u"abc", re.M).group(0), "abc")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
368 self.assertEqual(re.search(r"^\Aabc\Z$", u"\nabc\n", re.M), None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
369 self.assertEqual(re.search(r"\d\D\w\W\s\S",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
370 "1aa! a").group(0), "1aa! a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
371 self.assertEqual(re.search(r"\d\D\w\W\s\S",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
372 "1aa! a", re.LOCALE).group(0), "1aa! a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
373 self.assertEqual(re.search(r"\d\D\w\W\s\S",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
374 "1aa! a", re.UNICODE).group(0), "1aa! a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
375
75339
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
376 def test_string_boundaries(self):
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
377 # See http://bugs.python.org/issue10713
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
378 self.assertEqual(re.search(r"\b(abc)\b", "abc").group(1),
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
379 "abc")
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
380 # There's a word boundary at the start of a string.
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
381 self.assertTrue(re.match(r"\b", "abc"))
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
382 # A non-empty string includes a non-boundary zero-length match.
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
383 self.assertTrue(re.search(r"\B", "abc"))
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
384 # There is no non-boundary match at the start of a string.
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
385 self.assertFalse(re.match(r"\B", "abc"))
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
386 # However, an empty string contains no word boundaries, and also no
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
387 # non-boundaries.
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
388 self.assertEqual(re.search(r"\B", ""), None)
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
389 # This one is questionable and different from the perlre behaviour,
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
390 # but describes current behavior.
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
391 self.assertEqual(re.search(r"\b", ""), None)
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
392 # A single word-character string has two boundaries, but no
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
393 # non-boundary gaps.
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
394 self.assertEqual(len(re.findall(r"\b", "a")), 2)
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
395 self.assertEqual(len(re.findall(r"\B", "a")), 0)
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
396 # If there are no words, there are no boundaries
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
397 self.assertEqual(len(re.findall(r"\b", " ")), 0)
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
398 self.assertEqual(len(re.findall(r"\b", " ")), 0)
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
399 # Can match around the whitespace.
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
400 self.assertEqual(len(re.findall(r"\B", " ")), 2)
fc89e09ca2fc #10713: Improve documentation for \b and \B and add a few tests. Initial patch and tests by Martin Pool.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68927
diff changeset
401
29548
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
402 def test_bigcharset(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
403 self.assertEqual(re.match(u"([\u2222\u2223])",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
404 u"\u2222").group(1), u"\u2222")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
405 self.assertEqual(re.match(u"([\u2222\u2223])",
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
406 u"\u2222", re.UNICODE).group(1), u"\u2222")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
407
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
408 def test_anyall(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
409 self.assertEqual(re.match("a.b", "a\nb", re.DOTALL).group(0),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
410 "a\nb")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
411 self.assertEqual(re.match("a.*b", "a\n\nb", re.DOTALL).group(0),
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
412 "a\n\nb")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
413
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
414 def test_non_consuming(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
415 self.assertEqual(re.match("(a(?=\s[^a]))", "a b").group(1), "a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
416 self.assertEqual(re.match("(a(?=\s[^a]*))", "a b").group(1), "a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
417 self.assertEqual(re.match("(a(?=\s[abc]))", "a b").group(1), "a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
418 self.assertEqual(re.match("(a(?=\s[abc]*))", "a bc").group(1), "a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
419 self.assertEqual(re.match(r"(a)(?=\s\1)", "a a").group(1), "a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
420 self.assertEqual(re.match(r"(a)(?=\s\1*)", "a aa").group(1), "a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
421 self.assertEqual(re.match(r"(a)(?=\s(abc|a))", "a a").group(1), "a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
422
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
423 self.assertEqual(re.match(r"(a(?!\s[^a]))", "a a").group(1), "a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
424 self.assertEqual(re.match(r"(a(?!\s[abc]))", "a d").group(1), "a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
425 self.assertEqual(re.match(r"(a)(?!\s\1)", "a b").group(1), "a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
426 self.assertEqual(re.match(r"(a)(?!\s(abc|a))", "a b").group(1), "a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
427
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
428 def test_ignore_case(self):
48855
70edac890f2b #3654: fix duplicate test method name. Review by Benjamin P.
Georg Brandl <georg@python.org>
parents: 44624
diff changeset
429 self.assertEqual(re.match("abc", "ABC", re.I).group(0), "ABC")
70edac890f2b #3654: fix duplicate test method name. Review by Benjamin P.
Georg Brandl <georg@python.org>
parents: 44624
diff changeset
430 self.assertEqual(re.match("abc", u"ABC", re.I).group(0), "ABC")
29548
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
431 self.assertEqual(re.match(r"(a\s[^a])", "a b", re.I).group(1), "a b")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
432 self.assertEqual(re.match(r"(a\s[^a]*)", "a bb", re.I).group(1), "a bb")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
433 self.assertEqual(re.match(r"(a\s[abc])", "a b", re.I).group(1), "a b")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
434 self.assertEqual(re.match(r"(a\s[abc]*)", "a bb", re.I).group(1), "a bb")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
435 self.assertEqual(re.match(r"((a)\s\2)", "a a", re.I).group(1), "a a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
436 self.assertEqual(re.match(r"((a)\s\2*)", "a aa", re.I).group(1), "a aa")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
437 self.assertEqual(re.match(r"((a)\s(abc|a))", "a a", re.I).group(1), "a a")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
438 self.assertEqual(re.match(r"((a)\s(abc|a)*)", "a aa", re.I).group(1), "a aa")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
439
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
440 def test_category(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
441 self.assertEqual(re.match(r"(\s)", " ").group(1), " ")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
442
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
443 def test_getlower(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
444 import _sre
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
445 self.assertEqual(_sre.getlower(ord('A'), 0), ord('a'))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
446 self.assertEqual(_sre.getlower(ord('A'), re.LOCALE), ord('a'))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
447 self.assertEqual(_sre.getlower(ord('A'), re.UNICODE), ord('a'))
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
448
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
449 self.assertEqual(re.match("abc", "ABC", re.I).group(0), "ABC")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
450 self.assertEqual(re.match("abc", u"ABC", re.I).group(0), "ABC")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
451
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
452 def test_not_literal(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
453 self.assertEqual(re.search("\s([^a])", " b").group(1), "b")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
454 self.assertEqual(re.search("\s([^a]*)", " bb").group(1), "bb")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
455
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
456 def test_search_coverage(self):
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
457 self.assertEqual(re.search("\s(b)", " b").group(1), "b")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
458 self.assertEqual(re.search("a\s", "a ").group(0), "a ")
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
459
68927
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
460 def assertMatch(self, pattern, text, match=None, span=None,
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
461 matcher=re.match):
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
462 if match is None and span is None:
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
463 # the pattern matches the whole text
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
464 match = text
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
465 span = (0, len(text))
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
466 elif match is None or span is None:
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
467 raise ValueError('If match is not None, span should be specified '
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
468 '(and vice versa).')
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
469 m = matcher(pattern, text)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
470 self.assertTrue(m)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
471 self.assertEqual(m.group(), match)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
472 self.assertEqual(m.span(), span)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
473
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
474 def test_re_escape(self):
68927
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
475 alnum_chars = string.ascii_letters + string.digits
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
476 p = u''.join(unichr(i) for i in range(256))
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
477 for c in p:
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
478 if c in alnum_chars:
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
479 self.assertEqual(re.escape(c), c)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
480 elif c == u'\x00':
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
481 self.assertEqual(re.escape(c), u'\\000')
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
482 else:
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
483 self.assertEqual(re.escape(c), u'\\' + c)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
484 self.assertMatch(re.escape(c), c)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
485 self.assertMatch(re.escape(p), p)
7852
f34a658e7c15 AMK's latest
Guido van Rossum <guido@python.org>
parents: 7747
diff changeset
486
68927
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
487 def test_re_escape_byte(self):
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
488 alnum_chars = (string.ascii_letters + string.digits).encode('ascii')
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
489 p = ''.join(chr(i) for i in range(256))
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
490 for b in p:
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
491 if b in alnum_chars:
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
492 self.assertEqual(re.escape(b), b)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
493 elif b == b'\x00':
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
494 self.assertEqual(re.escape(b), b'\\000')
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
495 else:
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
496 self.assertEqual(re.escape(b), b'\\' + b)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
497 self.assertMatch(re.escape(b), b)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
498 self.assertMatch(re.escape(p), p)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
499
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
500 def test_re_escape_non_ascii(self):
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
501 s = u'xxx\u2620\u2620\u2620xxx'
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
502 s_escaped = re.escape(s)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
503 self.assertEqual(s_escaped, u'xxx\\\u2620\\\u2620\\\u2620xxx')
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
504 self.assertMatch(s_escaped, s)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
505 self.assertMatch(u'.%s+.' % re.escape(u'\u2620'), s,
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
506 u'x\u2620\u2620\u2620x', (2, 7), re.search)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
507
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
508 def test_re_escape_non_ascii_bytes(self):
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
509 b = u'y\u2620y\u2620y'.encode('utf-8')
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
510 b_escaped = re.escape(b)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
511 self.assertEqual(b_escaped, b'y\\\xe2\\\x98\\\xa0y\\\xe2\\\x98\\\xa0y')
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
512 self.assertMatch(b_escaped, b)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
513 res = re.findall(re.escape(u'\u2620'.encode('utf-8')), b)
d52b1faa7b11 #2650: Refactor re.escape and its tests.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 68503
diff changeset
514 self.assertEqual(len(res), 2)
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
515
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
516 def test_pickling(self):
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
517 import pickle
28840
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
518 self.pickle_test(pickle)
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
519 import cPickle
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
520 self.pickle_test(cPickle)
41272
408dfc7e3018 Bug #1675967: re patterns pickled with older Python versions can
Žiga Seilnacht <ziga.seilnacht@gmail.com>
parents: 37017
diff changeset
521 # old pickles expect the _compile() reconstructor in sre module
63634
762d89faceaa Merged revisions 79539 via svnmerge from
Ezio Melotti <ezio.melotti@gmail.com>
parents: 58453
diff changeset
522 import_module("sre", deprecated=True)
762d89faceaa Merged revisions 79539 via svnmerge from
Ezio Melotti <ezio.melotti@gmail.com>
parents: 58453
diff changeset
523 from sre import _compile
28840
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
524
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
525 def pickle_test(self, pickle):
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
526 oldpat = re.compile('a(?:b|(c|e){1,2}?|d)+?(.)')
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
527 s = pickle.dumps(oldpat)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
528 newpat = pickle.loads(s)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
529 self.assertEqual(oldpat, newpat)
7852
f34a658e7c15 AMK's latest
Guido van Rossum <guido@python.org>
parents: 7747
diff changeset
530
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
531 def test_constants(self):
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
532 self.assertEqual(re.I, re.IGNORECASE)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
533 self.assertEqual(re.L, re.LOCALE)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
534 self.assertEqual(re.M, re.MULTILINE)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
535 self.assertEqual(re.S, re.DOTALL)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
536 self.assertEqual(re.X, re.VERBOSE)
7852
f34a658e7c15 AMK's latest
Guido van Rossum <guido@python.org>
parents: 7747
diff changeset
537
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
538 def test_flags(self):
28840
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
539 for flag in [re.I, re.M, re.X, re.S, re.L]:
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
540 self.assertNotEqual(re.compile('^pattern$', flag), None)
6777
daa77d9af34c New re version from AMK
Guido van Rossum <guido@python.org>
parents: 6447
diff changeset
541
28823
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
542 def test_sre_character_literals(self):
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
543 for i in [0, 8, 16, 32, 64, 127, 128, 255]:
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
544 self.assertNotEqual(re.match(r"\%03o" % i, chr(i)), None)
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
545 self.assertNotEqual(re.match(r"\%03o0" % i, chr(i)+"0"), None)
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
546 self.assertNotEqual(re.match(r"\%03o8" % i, chr(i)+"8"), None)
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
547 self.assertNotEqual(re.match(r"\x%02x" % i, chr(i)), None)
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
548 self.assertNotEqual(re.match(r"\x%02x0" % i, chr(i)+"0"), None)
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
549 self.assertNotEqual(re.match(r"\x%02xz" % i, chr(i)+"z"), None)
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
550 self.assertRaises(re.error, re.match, "\911", "")
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
551
33397
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
552 def test_sre_character_class_literals(self):
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
553 for i in [0, 8, 16, 32, 64, 127, 128, 255]:
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
554 self.assertNotEqual(re.match(r"[\%03o]" % i, chr(i)), None)
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
555 self.assertNotEqual(re.match(r"[\%03o0]" % i, chr(i)), None)
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
556 self.assertNotEqual(re.match(r"[\%03o8]" % i, chr(i)), None)
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
557 self.assertNotEqual(re.match(r"[\x%02x]" % i, chr(i)), None)
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
558 self.assertNotEqual(re.match(r"[\x%02x0]" % i, chr(i)), None)
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
559 self.assertNotEqual(re.match(r"[\x%02xz]" % i, chr(i)), None)
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
560 self.assertRaises(re.error, re.match, "[\911]", "")
98c8017e8535 Applying modified version of patch #1018386, which fixes
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 32160
diff changeset
561
28823
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
562 def test_bug_113254(self):
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
563 self.assertEqual(re.match(r'(a)|(b)', 'b').start(1), -1)
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
564 self.assertEqual(re.match(r'(a)|(b)', 'b').end(1), -1)
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
565 self.assertEqual(re.match(r'(a)|(b)', 'b').span(1), (-1, -1))
7296cd3f45c7 copy a few tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28811
diff changeset
566
28826
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
567 def test_bug_527371(self):
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
568 # bug described in patches 527371/672491
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
569 self.assertEqual(re.match(r'(a)?a','a').lastindex, None)
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
570 self.assertEqual(re.match(r'(a)(b)?b','ab').lastindex, 1)
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
571 self.assertEqual(re.match(r'(?P<a>a)(?P<b>b)?b','ab').lastgroup, 'a')
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
572 self.assertEqual(re.match("(?P<a>a(b))", "ab").lastgroup, 'a')
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
573 self.assertEqual(re.match("((a))", "a").lastindex, 1)
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
574
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
575 def test_bug_545855(self):
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
576 # bug 545855 -- This pattern failed to cause a compile error as it
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
577 # should, instead provoking a TypeError.
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
578 self.assertRaises(re.error, re.compile, 'foo[a-')
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
579
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
580 def test_bug_418626(self):
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
581 # bugs 418626 at al. -- Testing Greg Chapman's addition of op code
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
582 # SRE_OP_MIN_REPEAT_ONE for eliminating recursion on simple uses of
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
583 # pattern '*?' on a long string.
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
584 self.assertEqual(re.match('.*?c', 10000*'ab'+'cd').end(0), 20001)
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
585 self.assertEqual(re.match('.*?cd', 5000*'ab'+'c'+5000*'ab'+'cde').end(0),
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
586 20003)
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
587 self.assertEqual(re.match('.*?cd', 20000*'abc'+'de').end(0), 60001)
30703
8282aa1f9343 Implemented non-recursive SRE matching.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 29748
diff changeset
588 # non-simple '*?' still used to hit the recursion limit, before the
31389
aca1ce1b4b61 Whitespace normalization.
Tim Peters <tim.peters@gmail.com>
parents: 30703
diff changeset
589 # non-recursive scheme was implemented.
30703
8282aa1f9343 Implemented non-recursive SRE matching.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 29748
diff changeset
590 self.assertEqual(re.search('(a|b)*?c', 10000*'ab'+'cd').end(0), 20001)
28826
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
591
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
592 def test_bug_612074(self):
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
593 pat=u"["+re.escape(u"\u2039")+u"]"
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
594 self.assertEqual(re.compile(pat) and 1, 1)
9c785ab1eff0 more tests from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28823
diff changeset
595
28840
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
596 def test_stack_overflow(self):
30703
8282aa1f9343 Implemented non-recursive SRE matching.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 29748
diff changeset
597 # nasty cases that used to overflow the straightforward recursive
28840
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
598 # implementation of repeated groups.
30703
8282aa1f9343 Implemented non-recursive SRE matching.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 29748
diff changeset
599 self.assertEqual(re.match('(x)*', 50000*'x').group(1), 'x')
8282aa1f9343 Implemented non-recursive SRE matching.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 29748
diff changeset
600 self.assertEqual(re.match('(x)*y', 50000*'x'+'y').group(1), 'x')
8282aa1f9343 Implemented non-recursive SRE matching.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 29748
diff changeset
601 self.assertEqual(re.match('(x)*?y', 50000*'x'+'y').group(1), 'x')
28840
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
602
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
603 def test_scanner(self):
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
604 def s_ident(scanner, token): return token
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
605 def s_operator(scanner, token): return "op%s" % token
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
606 def s_float(scanner, token): return float(token)
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
607 def s_int(scanner, token): return int(token)
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
608
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
609 scanner = Scanner([
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
610 (r"[a-zA-Z_]\w*", s_ident),
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
611 (r"\d+\.\d*", s_float),
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
612 (r"\d+", s_int),
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
613 (r"=|\+|-|\*|/", s_operator),
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
614 (r"\s+", None),
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
615 ])
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
616
29548
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
617 self.assertNotEqual(scanner.scanner.scanner("").pattern, None)
8780a49d8ef2 Many new tests, based on gcov's coverage information.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28917
diff changeset
618
28840
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
619 self.assertEqual(scanner.scan("sum = 3*foo + 312.50 + bar"),
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
620 (['sum', 'op=', 3, 'op*', 'foo', 'op+', 312.5,
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
621 'op+', 'bar'], ''))
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
622
28843
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
623 def test_bug_448951(self):
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
624 # bug 448951 (similar to 429357, but with single char match)
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
625 # (Also test greedy matches.)
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
626 for op in '','?','*':
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
627 self.assertEqual(re.match(r'((.%s):)?z'%op, 'z').groups(),
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
628 (None, None))
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
629 self.assertEqual(re.match(r'((.%s):)?z'%op, 'a:z').groups(),
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
630 ('a:', 'a'))
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
631
28864
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
632 def test_bug_725106(self):
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
633 # capturing groups in alternatives in repeats
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
634 self.assertEqual(re.match('^((a)|b)*', 'abc').groups(),
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
635 ('b', 'a'))
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
636 self.assertEqual(re.match('^(([ab])|c)*', 'abc').groups(),
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
637 ('c', 'b'))
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
638 self.assertEqual(re.match('^((d)|[ab])*', 'abc').groups(),
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
639 ('b', None))
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
640 self.assertEqual(re.match('^((a)c|[ab])*', 'abc').groups(),
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
641 ('b', None))
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
642 self.assertEqual(re.match('^((a)|b)*?c', 'abc').groups(),
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
643 ('b', 'a'))
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
644 self.assertEqual(re.match('^(([ab])|c)*?d', 'abcd').groups(),
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
645 ('c', 'b'))
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
646 self.assertEqual(re.match('^((d)|[ab])*?c', 'abc').groups(),
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
647 ('b', None))
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
648 self.assertEqual(re.match('^((a)c|[ab])*?c', 'abc').groups(),
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
649 ('b', None))
16374052cb5d Applied patch #725106, by Greg Chapman, fixing capturing groups
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28843
diff changeset
650
28865
53ff23bb3248 Fix for part of the problem mentioned in #725149 by Greg Chapman.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28864
diff changeset
651 def test_bug_725149(self):
53ff23bb3248 Fix for part of the problem mentioned in #725149 by Greg Chapman.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28864
diff changeset
652 # mark_stack_base restoring before restoring marks
53ff23bb3248 Fix for part of the problem mentioned in #725149 by Greg Chapman.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28864
diff changeset
653 self.assertEqual(re.match('(a)(?:(?=(b)*)c)*', 'abb').groups(),
53ff23bb3248 Fix for part of the problem mentioned in #725149 by Greg Chapman.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28864
diff changeset
654 ('a', None))
53ff23bb3248 Fix for part of the problem mentioned in #725149 by Greg Chapman.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28864
diff changeset
655 self.assertEqual(re.match('(a)((?!(b)*))*', 'abb').groups(),
53ff23bb3248 Fix for part of the problem mentioned in #725149 by Greg Chapman.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28864
diff changeset
656 ('a', None, None))
53ff23bb3248 Fix for part of the problem mentioned in #725149 by Greg Chapman.
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 28864
diff changeset
657
29748
5af89037ceba Fix and test for bug #764548:
Just van Rossum <just@letterror.com>
parents: 29742
diff changeset
658 def test_bug_764548(self):
5af89037ceba Fix and test for bug #764548:
Just van Rossum <just@letterror.com>
parents: 29742
diff changeset
659 # bug 764548, re.compile() barfs on str/unicode subclasses
5af89037ceba Fix and test for bug #764548:
Just van Rossum <just@letterror.com>
parents: 29742
diff changeset
660 try:
5af89037ceba Fix and test for bug #764548:
Just van Rossum <just@letterror.com>
parents: 29742
diff changeset
661 unicode
5af89037ceba Fix and test for bug #764548:
Just van Rossum <just@letterror.com>
parents: 29742
diff changeset
662 except NameError:
5af89037ceba Fix and test for bug #764548:
Just van Rossum <just@letterror.com>
parents: 29742
diff changeset
663 return # no problem if we have no unicode
5af89037ceba Fix and test for bug #764548:
Just van Rossum <just@letterror.com>
parents: 29742
diff changeset
664 class my_unicode(unicode): pass
5af89037ceba Fix and test for bug #764548:
Just van Rossum <just@letterror.com>
parents: 29742
diff changeset
665 pat = re.compile(my_unicode("abc"))
5af89037ceba Fix and test for bug #764548:
Just van Rossum <just@letterror.com>
parents: 29742
diff changeset
666 self.assertEqual(pat.match("xyz"), None)
5af89037ceba Fix and test for bug #764548:
Just van Rossum <just@letterror.com>
parents: 29742
diff changeset
667
28843
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
668 def test_finditer(self):
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
669 iter = re.finditer(r":+", "a:b::c:::d")
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
670 self.assertEqual([item.group(0) for item in iter],
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
671 [":", "::", ":::"])
86511b96d980 final bit of tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28840
diff changeset
672
31935
01ea03e7b265 SF #926075: Fixed the bug that returns a wrong pattern object for
Hye-Shik Chang <hyeshik@gmail.com>
parents: 31389
diff changeset
673 def test_bug_926075(self):
01ea03e7b265 SF #926075: Fixed the bug that returns a wrong pattern object for
Hye-Shik Chang <hyeshik@gmail.com>
parents: 31389
diff changeset
674 try:
01ea03e7b265 SF #926075: Fixed the bug that returns a wrong pattern object for
Hye-Shik Chang <hyeshik@gmail.com>
parents: 31389
diff changeset
675 unicode
01ea03e7b265 SF #926075: Fixed the bug that returns a wrong pattern object for
Hye-Shik Chang <hyeshik@gmail.com>
parents: 31389
diff changeset
676 except NameError:
01ea03e7b265 SF #926075: Fixed the bug that returns a wrong pattern object for
Hye-Shik Chang <hyeshik@gmail.com>
parents: 31389
diff changeset
677 return # no problem if we have no unicode
55002
6768b2350c57 convert usage of fail* to assert*
Benjamin Peterson <benjamin@python.org>
parents: 49126
diff changeset
678 self.assertTrue(re.compile('bug_926075') is not
31935
01ea03e7b265 SF #926075: Fixed the bug that returns a wrong pattern object for
Hye-Shik Chang <hyeshik@gmail.com>
parents: 31389
diff changeset
679 re.compile(eval("u'bug_926075'")))
01ea03e7b265 SF #926075: Fixed the bug that returns a wrong pattern object for
Hye-Shik Chang <hyeshik@gmail.com>
parents: 31389
diff changeset
680
32006
e2a0ef87d4d4 Fix _sre.CODESIZE on 64-bit machines in UCS-4 mode. Fixes #931848.
Martin v. Löwis <martin@v.loewis.de>
parents: 31935
diff changeset
681 def test_bug_931848(self):
e2a0ef87d4d4 Fix _sre.CODESIZE on 64-bit machines in UCS-4 mode. Fixes #931848.
Martin v. Löwis <martin@v.loewis.de>
parents: 31935
diff changeset
682 try:
e2a0ef87d4d4 Fix _sre.CODESIZE on 64-bit machines in UCS-4 mode. Fixes #931848.
Martin v. Löwis <martin@v.loewis.de>
parents: 31935
diff changeset
683 unicode
e2a0ef87d4d4 Fix _sre.CODESIZE on 64-bit machines in UCS-4 mode. Fixes #931848.
Martin v. Löwis <martin@v.loewis.de>
parents: 31935
diff changeset
684 except NameError:
e2a0ef87d4d4 Fix _sre.CODESIZE on 64-bit machines in UCS-4 mode. Fixes #931848.
Martin v. Löwis <martin@v.loewis.de>
parents: 31935
diff changeset
685 pass
e2a0ef87d4d4 Fix _sre.CODESIZE on 64-bit machines in UCS-4 mode. Fixes #931848.
Martin v. Löwis <martin@v.loewis.de>
parents: 31935
diff changeset
686 pattern = eval('u"[\u002E\u3002\uFF0E\uFF61]"')
e2a0ef87d4d4 Fix _sre.CODESIZE on 64-bit machines in UCS-4 mode. Fixes #931848.
Martin v. Löwis <martin@v.loewis.de>
parents: 31935
diff changeset
687 self.assertEqual(re.compile(pattern).split("a.b.c"),
e2a0ef87d4d4 Fix _sre.CODESIZE on 64-bit machines in UCS-4 mode. Fixes #931848.
Martin v. Löwis <martin@v.loewis.de>
parents: 31935
diff changeset
688 ['a','b','c'])
e2a0ef87d4d4 Fix _sre.CODESIZE on 64-bit machines in UCS-4 mode. Fixes #931848.
Martin v. Löwis <martin@v.loewis.de>
parents: 31935
diff changeset
689
33398
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
690 def test_bug_581080(self):
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
691 iter = re.finditer(r"\s", "a b")
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
692 self.assertEqual(iter.next().span(), (1,2))
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
693 self.assertRaises(StopIteration, iter.next)
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
694
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
695 scanner = re.compile(r"\s").scanner("a b")
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
696 self.assertEqual(scanner.search().span(), (1, 2))
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
697 self.assertEqual(scanner.search(), None)
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
698
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
699 def test_bug_817234(self):
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
700 iter = re.finditer(r".*", "asdf")
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
701 self.assertEqual(iter.next().span(), (0, 4))
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
702 self.assertEqual(iter.next().span(), (4, 4))
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
703 self.assertRaises(StopIteration, iter.next)
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
704
55480
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
705 def test_bug_6561(self):
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
706 # '\d' should match characters in Unicode category 'Nd'
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
707 # (Number, Decimal Digit), but not those in 'Nl' (Number,
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
708 # Letter) or 'No' (Number, Other).
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
709 decimal_digits = [
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
710 u'\u0037', # '\N{DIGIT SEVEN}', category 'Nd'
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
711 u'\u0e58', # '\N{THAI DIGIT SIX}', category 'Nd'
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
712 u'\uff10', # '\N{FULLWIDTH DIGIT ZERO}', category 'Nd'
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
713 ]
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
714 for x in decimal_digits:
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
715 self.assertEqual(re.match('^\d$', x, re.UNICODE).group(0), x)
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
716
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
717 not_decimal_digits = [
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
718 u'\u2165', # '\N{ROMAN NUMERAL SIX}', category 'Nl'
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
719 u'\u3039', # '\N{HANGZHOU NUMERAL TWENTY}', category 'Nl'
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
720 u'\u2082', # '\N{SUBSCRIPT TWO}', category 'No'
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
721 u'\u32b4', # '\N{CIRCLED NUMBER THIRTY NINE}', category 'No'
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
722 ]
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
723 for x in not_decimal_digits:
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
724 self.assertIsNone(re.match('^\d$', x, re.UNICODE))
56b2da93ad9c Issue #6561: '\d' regular expression should not match characters of
Mark Dickinson <dickinsm@gmail.com>
parents: 55002
diff changeset
725
41360
6a3dd1c4a7df Array module's buffer interface can now handle empty arrays.
Raymond Hettinger <python@rcn.com>
parents: 41272
diff changeset
726 def test_empty_array(self):
6a3dd1c4a7df Array module's buffer interface can now handle empty arrays.
Raymond Hettinger <python@rcn.com>
parents: 41272
diff changeset
727 # SF buf 1647541
6a3dd1c4a7df Array module's buffer interface can now handle empty arrays.
Raymond Hettinger <python@rcn.com>
parents: 41272
diff changeset
728 import array
6a3dd1c4a7df Array module's buffer interface can now handle empty arrays.
Raymond Hettinger <python@rcn.com>
parents: 41272
diff changeset
729 for typecode in 'cbBuhHiIlLfd':
6a3dd1c4a7df Array module's buffer interface can now handle empty arrays.
Raymond Hettinger <python@rcn.com>
parents: 41272
diff changeset
730 a = array.array(typecode)
6a3dd1c4a7df Array module's buffer interface can now handle empty arrays.
Raymond Hettinger <python@rcn.com>
parents: 41272
diff changeset
731 self.assertEqual(re.compile("bla").match(a), None)
41533
7ded787e6311 Whitespace normalization
Neal Norwitz <nnorwitz@gmail.com>
parents: 41360
diff changeset
732 self.assertEqual(re.compile("").match(a).groups(), ())
33398
d1f8c632b605 Fixing bug #817234, which made SRE get into an infinite loop on
Gustavo Niemeyer <gustavo@niemeyer.net>
parents: 33397
diff changeset
733
44435
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
734 def test_inline_flags(self):
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
735 # Bug #1700
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
736 upper_char = unichr(0x1ea0) # Latin Capital Letter A with Dot Bellow
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
737 lower_char = unichr(0x1ea1) # Latin Small Letter A with Dot Bellow
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
738
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
739 p = re.compile(upper_char, re.I | re.U)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
740 q = p.match(lower_char)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
741 self.assertNotEqual(q, None)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
742
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
743 p = re.compile(lower_char, re.I | re.U)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
744 q = p.match(upper_char)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
745 self.assertNotEqual(q, None)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
746
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
747 p = re.compile('(?i)' + upper_char, re.U)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
748 q = p.match(lower_char)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
749 self.assertNotEqual(q, None)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
750
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
751 p = re.compile('(?i)' + lower_char, re.U)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
752 q = p.match(upper_char)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
753 self.assertNotEqual(q, None)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
754
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
755 p = re.compile('(?iu)' + upper_char)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
756 q = p.match(lower_char)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
757 self.assertNotEqual(q, None)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
758
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
759 p = re.compile('(?iu)' + lower_char)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
760 q = p.match(upper_char)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
761 self.assertNotEqual(q, None)
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
762
44624
e753e31a0c8f Closing issue1761.
Amaury Forgeot d'Arc <amauryfa@gmail.com>
parents: 44436
diff changeset
763 def test_dollar_matches_twice(self):
e753e31a0c8f Closing issue1761.
Amaury Forgeot d'Arc <amauryfa@gmail.com>
parents: 44436
diff changeset
764 "$ matches the end of string, and just before the terminating \n"
e753e31a0c8f Closing issue1761.
Amaury Forgeot d'Arc <amauryfa@gmail.com>
parents: 44436
diff changeset
765 pattern = re.compile('$')
e753e31a0c8f Closing issue1761.
Amaury Forgeot d'Arc <amauryfa@gmail.com>
parents: 44436
diff changeset
766 self.assertEqual(pattern.sub('#', 'a\nb\n'), 'a\nb#\n#')
e753e31a0c8f Closing issue1761.
Amaury Forgeot d'Arc <amauryfa@gmail.com>
parents: 44436
diff changeset
767 self.assertEqual(pattern.sub('#', 'a\nb\nc'), 'a\nb\nc#')
e753e31a0c8f Closing issue1761.
Amaury Forgeot d'Arc <amauryfa@gmail.com>
parents: 44436
diff changeset
768 self.assertEqual(pattern.sub('#', '\n'), '#\n#')
e753e31a0c8f Closing issue1761.
Amaury Forgeot d'Arc <amauryfa@gmail.com>
parents: 44436
diff changeset
769
e753e31a0c8f Closing issue1761.
Amaury Forgeot d'Arc <amauryfa@gmail.com>
parents: 44436
diff changeset
770 pattern = re.compile('$', re.MULTILINE)
e753e31a0c8f Closing issue1761.
Amaury Forgeot d'Arc <amauryfa@gmail.com>
parents: 44436
diff changeset
771 self.assertEqual(pattern.sub('#', 'a\nb\n' ), 'a#\nb#\n#' )
e753e31a0c8f Closing issue1761.
Amaury Forgeot d'Arc <amauryfa@gmail.com>
parents: 44436
diff changeset
772 self.assertEqual(pattern.sub('#', 'a\nb\nc'), 'a#\nb#\nc#')
e753e31a0c8f Closing issue1761.
Amaury Forgeot d'Arc <amauryfa@gmail.com>
parents: 44436
diff changeset
773 self.assertEqual(pattern.sub('#', '\n'), '#\n#')
e753e31a0c8f Closing issue1761.
Amaury Forgeot d'Arc <amauryfa@gmail.com>
parents: 44436
diff changeset
774
58326
35469e79ff80 Merged revisions 77499 via svnmerge from
Antoine Pitrou <solipsis@pitrou.net>
parents: 49126
diff changeset
775 def test_dealloc(self):
35469e79ff80 Merged revisions 77499 via svnmerge from
Antoine Pitrou <solipsis@pitrou.net>
parents: 49126
diff changeset
776 # issue 3299: check for segfault in debug build
35469e79ff80 Merged revisions 77499 via svnmerge from
Antoine Pitrou <solipsis@pitrou.net>
parents: 49126
diff changeset
777 import _sre
58453
bb1c33efff93 Merged revisions 77706 via svnmerge from
Ezio Melotti <ezio.melotti@gmail.com>
parents: 58326
diff changeset
778 # the overflow limit is different on wide and narrow builds and it
bb1c33efff93 Merged revisions 77706 via svnmerge from
Ezio Melotti <ezio.melotti@gmail.com>
parents: 58326
diff changeset
779 # depends on the definition of SRE_CODE (see sre.h).
bb1c33efff93 Merged revisions 77706 via svnmerge from
Ezio Melotti <ezio.melotti@gmail.com>
parents: 58326
diff changeset
780 # 2**128 should be big enough to overflow on both. For smaller values
bb1c33efff93 Merged revisions 77706 via svnmerge from
Ezio Melotti <ezio.melotti@gmail.com>
parents: 58326
diff changeset
781 # a RuntimeError is raised instead of OverflowError.
bb1c33efff93 Merged revisions 77706 via svnmerge from
Ezio Melotti <ezio.melotti@gmail.com>
parents: 58326
diff changeset
782 long_overflow = 2**128
58326
35469e79ff80 Merged revisions 77499 via svnmerge from
Antoine Pitrou <solipsis@pitrou.net>
parents: 49126
diff changeset
783 self.assertRaises(TypeError, re.finditer, "a", {})
35469e79ff80 Merged revisions 77499 via svnmerge from
Antoine Pitrou <solipsis@pitrou.net>
parents: 49126
diff changeset
784 self.assertRaises(OverflowError, _sre.compile, "abc", 0, [long_overflow])
44435
18a1bc49a696 Issue #1700, reported by Nguyen Quan Son, fix by Fredruk Lundh:
Guido van Rossum <guido@python.org>
parents: 43381
diff changeset
785
75563
cc48eef234dd #14179: add tests for re.compile. Patch by Florian Mladitsch.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 75339
diff changeset
786 def test_compile(self):
cc48eef234dd #14179: add tests for re.compile. Patch by Florian Mladitsch.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 75339
diff changeset
787 # Test return value when given string and pattern as parameter
cc48eef234dd #14179: add tests for re.compile. Patch by Florian Mladitsch.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 75339
diff changeset
788 pattern = re.compile('random pattern')
cc48eef234dd #14179: add tests for re.compile. Patch by Florian Mladitsch.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 75339
diff changeset
789 self.assertIsInstance(pattern, re._pattern_type)
cc48eef234dd #14179: add tests for re.compile. Patch by Florian Mladitsch.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 75339
diff changeset
790 same_pattern = re.compile(pattern)
cc48eef234dd #14179: add tests for re.compile. Patch by Florian Mladitsch.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 75339
diff changeset
791 self.assertIsInstance(same_pattern, re._pattern_type)
cc48eef234dd #14179: add tests for re.compile. Patch by Florian Mladitsch.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 75339
diff changeset
792 self.assertIs(same_pattern, pattern)
cc48eef234dd #14179: add tests for re.compile. Patch by Florian Mladitsch.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 75339
diff changeset
793 # Test behaviour when not given a string or pattern as parameter
cc48eef234dd #14179: add tests for re.compile. Patch by Florian Mladitsch.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 75339
diff changeset
794 self.assertRaises(TypeError, re.compile, 0)
cc48eef234dd #14179: add tests for re.compile. Patch by Florian Mladitsch.
Ezio Melotti <ezio.melotti@gmail.com>
parents: 75339
diff changeset
795
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
796 def run_re_tests():
58817
70e115c34f97 Remove unused imports in test modules.
Georg Brandl <georg@python.org>
parents: 58452
diff changeset
797 from test.re_tests import tests, SUCCEED, FAIL, SYNTAX_ERROR
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
798 if verbose:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
799 print 'Running re_tests test suite'
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
800 else:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
801 # To save time, only run the first and last 10 tests
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
802 #tests = tests[:10] + tests[-10:]
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
803 pass
7852
f34a658e7c15 AMK's latest
Guido van Rossum <guido@python.org>
parents: 7747
diff changeset
804
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
805 for t in tests:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
806 sys.stdout.flush()
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
807 pattern = s = outcome = repl = expected = None
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
808 if len(t) == 5:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
809 pattern, s, outcome, repl, expected = t
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
810 elif len(t) == 3:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
811 pattern, s, outcome = t
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
812 else:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
813 raise ValueError, ('Test tuples should have 3 or 5 fields', t)
5774
3ef2563e83ee test suite for re.py
Guido van Rossum <guido@python.org>
parents:
diff changeset
814
7747
aedd2d0dcb31 Mass check-in after untabifying all files that need it.
Guido van Rossum <guido@python.org>
parents: 7098
diff changeset
815 try:
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
816 obj = re.compile(pattern)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
817 except re.error:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
818 if outcome == SYNTAX_ERROR: pass # Expected a syntax error
7747
aedd2d0dcb31 Mass check-in after untabifying all files that need it.
Guido van Rossum <guido@python.org>
parents: 7098
diff changeset
819 else:
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
820 print '=== Syntax error:', t
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
821 except KeyboardInterrupt: raise KeyboardInterrupt
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
822 except:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
823 print '*** Unexpected error ***', t
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
824 if verbose:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
825 traceback.print_exc(file=sys.stdout)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
826 else:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
827 try:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
828 result = obj.search(s)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
829 except re.error, msg:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
830 print '=== Unexpected exception', t, repr(msg)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
831 if outcome == SYNTAX_ERROR:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
832 # This should have been a syntax error; forget it.
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
833 pass
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
834 elif outcome == FAIL:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
835 if result is None: pass # No match, as expected
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
836 else: print '=== Succeeded incorrectly', t
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
837 elif outcome == SUCCEED:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
838 if result is not None:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
839 # Matched, as expected, so now we compute the
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
840 # result string and compare it to our expected result.
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
841 start, end = result.span(0)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
842 vardict={'found': result.group(0),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
843 'groups': result.group(),
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
844 'flags': result.re.flags}
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
845 for i in range(1, 100):
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
846 try:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
847 gi = result.group(i)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
848 # Special hack because else the string concat fails:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
849 if gi is None:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
850 gi = "None"
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
851 except IndexError:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
852 gi = "Error"
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
853 vardict['g%d' % i] = gi
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
854 for i in result.re.groupindex.keys():
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
855 try:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
856 gi = result.group(i)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
857 if gi is None:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
858 gi = "None"
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
859 except IndexError:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
860 gi = "Error"
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
861 vardict[i] = gi
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
862 repl = eval(repl, vardict)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
863 if repl != expected:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
864 print '=== grouping error', t,
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
865 print repr(repl) + ' should be ' + repr(expected)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
866 else:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
867 print '=== Failed incorrectly', t
6004
3f59b34eb90f 1.5a3 prerelease 1 from AMK
Guido van Rossum <guido@python.org>
parents: 5831
diff changeset
868
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
869 # Try the match on a unicode string, and check that it
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
870 # still succeeds.
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
871 try:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
872 result = obj.search(unicode(s, "latin-1"))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
873 if result is None:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
874 print '=== Fails on unicode match', t
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
875 except NameError:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
876 continue # 1.5.2
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
877 except TypeError:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
878 continue # unicode test case
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
879
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
880 # Try the match on a unicode pattern, and check that it
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
881 # still succeeds.
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
882 obj=re.compile(unicode(pattern, "latin-1"))
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
883 result = obj.search(s)
17145
8368404373ec SRE 2.1b1: don't do unicode tests under 1.5.2, or on unicode
Fredrik Lundh <fredrik@pythonware.com>
parents: 16518
diff changeset
884 if result is None:
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
885 print '=== Fails on unicode pattern match', t
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
886
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
887 # Try the match with the search area limited to the extent
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
888 # of the match and see if it still succeeds. \B will
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
889 # break (because it won't match at the end or start of a
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
890 # string), so we'll ignore patterns that feature it.
13897
0430a1309596 -- enabled some temporarily disabled RE tests
Fredrik Lundh <fredrik@pythonware.com>
parents: 13895
diff changeset
891
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
892 if pattern[:2] != '\\B' and pattern[-2:] != '\\B' \
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
893 and result is not None:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
894 obj = re.compile(pattern)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
895 result = obj.search(s, result.start(0), result.end(0) + 1)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
896 if result is None:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
897 print '=== Failed on range-limited match', t
13897
0430a1309596 -- enabled some temporarily disabled RE tests
Fredrik Lundh <fredrik@pythonware.com>
parents: 13895
diff changeset
898
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
899 # Try the match with IGNORECASE enabled, and check that it
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
900 # still succeeds.
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
901 obj = re.compile(pattern, re.IGNORECASE)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
902 result = obj.search(s)
15540
9d43129c8c55 Update the code to better reflect recommended style:
Fred Drake <fdrake@acm.org>
parents: 15305
diff changeset
903 if result is None:
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
904 print '=== Fails on case-insensitive match', t
6777
daa77d9af34c New re version from AMK
Guido van Rossum <guido@python.org>
parents: 6447
diff changeset
905
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
906 # Try the match with LOCALE enabled, and check that it
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
907 # still succeeds.
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
908 obj = re.compile(pattern, re.LOCALE)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
909 result = obj.search(s)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
910 if result is None:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
911 print '=== Fails on locale-sensitive match', t
6777
daa77d9af34c New re version from AMK
Guido van Rossum <guido@python.org>
parents: 6447
diff changeset
912
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
913 # Try the match with UNICODE locale enabled, and check
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
914 # that it still succeeds.
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
915 obj = re.compile(pattern, re.UNICODE)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
916 result = obj.search(s)
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
917 if result is None:
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
918 print '=== Fails on unicode-sensitive match', t
13897
0430a1309596 -- enabled some temporarily disabled RE tests
Fredrik Lundh <fredrik@pythonware.com>
parents: 13895
diff changeset
919
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
920 def test_main():
28917
1741af48fd9f Combine the functionality of test_support.run_unittest()
Walter Dörwald <walter@livinglogic.de>
parents: 28865
diff changeset
921 run_unittest(ReTests)
28840
490af00fb183 more tests converted from test_sre
Skip Montanaro <skip@pobox.com>
parents: 28826
diff changeset
922 run_re_tests()
28800
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
923
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
924 if __name__ == "__main__":
72db7575f970 first cut at unittest version of re tests
Skip Montanaro <skip@pobox.com>
parents: 24442
diff changeset
925 test_main()