PageRenderTime 56ms CodeModel.GetById 17ms RepoModel.GetById 0ms app.codeStats 0ms

/Lib/test/test_re.py

https://bitbucket.org/mirror/cpython/
Python | 1869 lines | 1817 code | 34 blank | 18 comment | 18 complexity | a90bba6abd28161c18061a8cb0794575 MD5 | raw file
Possible License(s): Unlicense, 0BSD, BSD-3-Clause

Large files files are truncated, but you can click here to view the full file

    

  1. from test.support import verbose, run_unittest, gc_collect, bigmemtest, _2G, \
    2. 

  2.         cpython_only, captured_stdout
    3. 

  3. import io
    4. 

  4. import locale
    5. 

  5. import re
    6. 

  6. from re import Scanner
    7. 

  7. import sre_compile
    8. 

  8. import sre_constants
    9. 

  9. import sys
    10. 

  10. import string
    11. 

  11. import traceback
    12. 

  12. import unittest
    13. 

  13. from weakref import proxy
    14. 

  14. 

  15. # Misc tests from Tim Peters' re.doc
    16. 

    17. 

  17. # WARNING: Don't change details in these tests if you don't know
    18. 

  18. # what you're doing. Some of these tests were carefully modeled to
    19. 

  19. # cover most of the code.
    20. 

  20. 

  21. class S(str):
    22. 

  22.     def __getitem__(self, index):
    23. 

  23.         return S(super().__getitem__(index))
    24. 

  24. 

  25. class B(bytes):
    26. 

  26.     def __getitem__(self, index):
    27. 

  27.         return B(super().__getitem__(index))
    28. 

  28. 

  29. class ReTests(unittest.TestCase):
    30. 

  30. 

  31.     def assertTypedEqual(self, actual, expect, msg=None):
    32. 

  32.         self.assertEqual(actual, expect, msg)
    33. 

  33.         def recurse(actual, expect):
    34. 

  34.             if isinstance(expect, (tuple, list)):
    35. 

  35.                 for x, y in zip(actual, expect):
    36. 

  36.                     recurse(x, y)
    37. 

  37.             else:
    38. 

  38.                 self.assertIs(type(actual), type(expect), msg)
    39. 

  39.         recurse(actual, expect)
    40. 

  40. 

  41.     def checkPatternError(self, pattern, errmsg, pos=None):
    42. 

  42.         with self.assertRaises(re.error) as cm:
    43. 

  43.             re.compile(pattern)
    44. 

  44.         with self.subTest(pattern=pattern):
    45. 

  45.             err = cm.exception
    46. 

  46.             self.assertEqual(err.msg, errmsg)
    47. 

  47.             if pos is not None:
    48. 

  48.                 self.assertEqual(err.pos, pos)
    49. 

  49. 

  50.     def checkTemplateError(self, pattern, repl, string, errmsg, pos=None):
    51. 

  51.         with self.assertRaises(re.error) as cm:
    52. 

  52.             re.sub(pattern, repl, string)
    53. 

  53.         with self.subTest(pattern=pattern, repl=repl):
    54. 

  54.             err = cm.exception
    55. 

  55.             self.assertEqual(err.msg, errmsg)
    56. 

  56.             if pos is not None:
    57. 

  57.                 self.assertEqual(err.pos, pos)
    58. 

  58. 

  59.     def test_keep_buffer(self):
    60. 

  60.         # See bug 14212
    61. 

  61.         b = bytearray(b'x')
    62. 

  62.         it = re.finditer(b'a', b)
    63. 

  63.         with self.assertRaises(BufferError):
    64. 

  64.             b.extend(b'x'*400)
    65. 

  65.         list(it)
    66. 

  66.         del it
    67. 

  67.         gc_collect()
    68. 

  68.         b.extend(b'x'*400)
    69. 

  69. 

  70.     def test_weakref(self):
    71. 

  71.         s = 'QabbbcR'
    72. 

  72.         x = re.compile('ab+c')
    73. 

  73.         y = proxy(x)
    74. 

  74.         self.assertEqual(x.findall('QabbbcR'), y.findall('QabbbcR'))
    75. 

  75. 

  76.     def test_search_star_plus(self):
    77. 

  77.         self.assertEqual(re.search('x*', 'axx').span(0), (0, 0))
    78. 

  78.         self.assertEqual(re.search('x*', 'axx').span(), (0, 0))
    79. 

  79.         self.assertEqual(re.search('x+', 'axx').span(0), (1, 3))
    80. 

  80.         self.assertEqual(re.search('x+', 'axx').span(), (1, 3))
    81. 

  81.         self.assertIsNone(re.search('x', 'aaa'))
    82. 

  82.         self.assertEqual(re.match('a*', 'xxx').span(0), (0, 0))
    83. 

  83.         self.assertEqual(re.match('a*', 'xxx').span(), (0, 0))
    84. 

  84.         self.assertEqual(re.match('x*', 'xxxa').span(0), (0, 3))
    85. 

  85.         self.assertEqual(re.match('x*', 'xxxa').span(), (0, 3))
    86. 

  86.         self.assertIsNone(re.match('a+', 'xxx'))
    87. 

  87. 

  88.     def bump_num(self, matchobj):
    89. 

  89.         int_value = int(matchobj.group(0))
    90. 

  90.         return str(int_value + 1)
    91. 

  91. 

  92.     def test_basic_re_sub(self):
    93. 

  93.         self.assertTypedEqual(re.sub('y', 'a', 'xyz'), 'xaz')
    94. 

  94.         self.assertTypedEqual(re.sub('y', S('a'), S('xyz')), 'xaz')
    95. 

  95.         self.assertTypedEqual(re.sub(b'y', b'a', b'xyz'), b'xaz')
    96. 

  96.         self.assertTypedEqual(re.sub(b'y', B(b'a'), B(b'xyz')), b'xaz')
    97. 

  97.         self.assertTypedEqual(re.sub(b'y', bytearray(b'a'), bytearray(b'xyz')), b'xaz')
    98. 

  98.         self.assertTypedEqual(re.sub(b'y', memoryview(b'a'), memoryview(b'xyz')), b'xaz')
    99. 

  99.         for y in ("\xe0", "\u0430", "\U0001d49c"):
    100. 

  100.             self.assertEqual(re.sub(y, 'a', 'x%sz' % y), 'xaz')
    101. 

  101. 

  102.         self.assertEqual(re.sub("(?i)b+", "x", "bbbb BBBB"), 'x x')
    103. 

  103.         self.assertEqual(re.sub(r'\d+', self.bump_num, '08.2 -2 23x99y'),
    104. 

  104.                          '9.3 -3 24x100y')
    105. 

  105.         self.assertEqual(re.sub(r'\d+', self.bump_num, '08.2 -2 23x99y', count=3),
    106. 

  106.                          '9.3 -3 23x99y')
    107. 

  107. 

  108.         self.assertEqual(re.sub('.', lambda m: r"\n", 'x'), '\\n')
    109. 

  109.         self.assertEqual(re.sub('.', r"\n", 'x'), '\n')
    110. 

  110. 

  111.         s = r"\1\1"
    112. 

  112.         self.assertEqual(re.sub('(.)', s, 'x'), 'xx')
    113. 

  113.         self.assertEqual(re.sub('(.)', re.escape(s), 'x'), s)
    114. 

  114.         self.assertEqual(re.sub('(.)', lambda m: s, 'x'), s)
    115. 

  115. 

  116.         self.assertEqual(re.sub('(?P<a>x)', '\g<a>\g<a>', 'xx'), 'xxxx')
    117. 

  117.         self.assertEqual(re.sub('(?P<a>x)', '\g<a>\g<1>', 'xx'), 'xxxx')
    118. 

  118.         self.assertEqual(re.sub('(?P<unk>x)', '\g<unk>\g<unk>', 'xx'), 'xxxx')
    119. 

  119.         self.assertEqual(re.sub('(?P<unk>x)', '\g<1>\g<1>', 'xx'), 'xxxx')
    120. 

  120. 

  121.         self.assertEqual(re.sub('a', r'\t\n\v\r\f\a\b', 'a'), '\t\n\v\r\f\a\b')
    122. 

  122.         self.assertEqual(re.sub('a', '\t\n\v\r\f\a\b', 'a'), '\t\n\v\r\f\a\b')
    123. 

  123.         self.assertEqual(re.sub('a', '\t\n\v\r\f\a\b', 'a'),
    124. 

  124.                          (chr(9)+chr(10)+chr(11)+chr(13)+chr(12)+chr(7)+chr(8)))
    125. 

  125.         for c in 'cdehijklmopqsuwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ':
    126. 

  126.             with self.subTest(c):
    127. 

  127.                 with self.assertRaises(re.error):
    128. 

  128.                     self.assertEqual(re.sub('a', '\\' + c, 'a'), '\\' + c)
    129. 

  129. 

  130.         self.assertEqual(re.sub('^\s*', 'X', 'test'), 'Xtest')
    131. 

  131. 

  132.     def test_bug_449964(self):
    133. 

  133.         # fails for group followed by other escape
    134. 

  134.         self.assertEqual(re.sub(r'(?P<unk>x)', '\g<1>\g<1>\\b', 'xx'),
    135. 

  135.                          'xx\bxx\b')
    136. 

  136. 

  137.     def test_bug_449000(self):
    138. 

  138.         # Test for sub() on escaped characters
    139. 

  139.         self.assertEqual(re.sub(r'\r\n', r'\n', 'abc\r\ndef\r\n'),
    140. 

  140.                          'abc\ndef\n')
    141. 

  141.         self.assertEqual(re.sub('\r\n', r'\n', 'abc\r\ndef\r\n'),
    142. 

  142.                          'abc\ndef\n')
    143. 

  143.         self.assertEqual(re.sub(r'\r\n', '\n', 'abc\r\ndef\r\n'),
    144. 

  144.                          'abc\ndef\n')
    145. 

  145.         self.assertEqual(re.sub('\r\n', '\n', 'abc\r\ndef\r\n'),
    146. 

  146.                          'abc\ndef\n')
    147. 

  147. 

  148.     def test_bug_1661(self):
    149. 

  149.         # Verify that flags do not get silently ignored with compiled patterns
    150. 

  150.         pattern = re.compile('.')
    151. 

  151.         self.assertRaises(ValueError, re.match, pattern, 'A', re.I)
    152. 

  152.         self.assertRaises(ValueError, re.search, pattern, 'A', re.I)
    153. 

  153.         self.assertRaises(ValueError, re.findall, pattern, 'A', re.I)
    154. 

  154.         self.assertRaises(ValueError, re.compile, pattern, re.I)
    155. 

  155. 

  156.     def test_bug_3629(self):
    157. 

  157.         # A regex that triggered a bug in the sre-code validator
    158. 

  158.         re.compile("(?P<quote>)(?(quote))")
    159. 

  159. 

  160.     def test_sub_template_numeric_escape(self):
    161. 

  161.         # bug 776311 and friends
    162. 

  162.         self.assertEqual(re.sub('x', r'\0', 'x'), '\0')
    163. 

  163.         self.assertEqual(re.sub('x', r'\000', 'x'), '\000')
    164. 

  164.         self.assertEqual(re.sub('x', r'\001', 'x'), '\001')
    165. 

  165.         self.assertEqual(re.sub('x', r'\008', 'x'), '\0' + '8')
    166. 

  166.         self.assertEqual(re.sub('x', r'\009', 'x'), '\0' + '9')
    167. 

  167.         self.assertEqual(re.sub('x', r'\111', 'x'), '\111')
    168. 

  168.         self.assertEqual(re.sub('x', r'\117', 'x'), '\117')
    169. 

  169.         self.assertEqual(re.sub('x', r'\377', 'x'), '\377')
    170. 

  170. 

  171.         self.assertEqual(re.sub('x', r'\1111', 'x'), '\1111')
    172. 

  172.         self.assertEqual(re.sub('x', r'\1111', 'x'), '\111' + '1')
    173. 

  173. 

  174.         self.assertEqual(re.sub('x', r'\00', 'x'), '\x00')
    175. 

  175.         self.assertEqual(re.sub('x', r'\07', 'x'), '\x07')
    176. 

  176.         self.assertEqual(re.sub('x', r'\08', 'x'), '\0' + '8')
    177. 

  177.         self.assertEqual(re.sub('x', r'\09', 'x'), '\0' + '9')
    178. 

  178.         self.assertEqual(re.sub('x', r'\0a', 'x'), '\0' + 'a')
    179. 

  179. 

  180.         self.checkTemplateError('x', r'\400', 'x',
    181. 

  181.                                 r'octal escape value \400 outside of '
    182. 

  182.                                 r'range 0-0o377', 0)
    183. 

  183.         self.checkTemplateError('x', r'\777', 'x',
    184. 

  184.                                 r'octal escape value \777 outside of '
    185. 

  185.                                 r'range 0-0o377', 0)
    186. 

  186. 

  187.         self.checkTemplateError('x', r'\1', 'x', 'invalid group reference')
    188. 

  188.         self.checkTemplateError('x', r'\8', 'x', 'invalid group reference')
    189. 

  189.         self.checkTemplateError('x', r'\9', 'x', 'invalid group reference')
    190. 

  190.         self.checkTemplateError('x', r'\11', 'x', 'invalid group reference')
    191. 

  191.         self.checkTemplateError('x', r'\18', 'x', 'invalid group reference')
    192. 

  192.         self.checkTemplateError('x', r'\1a', 'x', 'invalid group reference')
    193. 

  193.         self.checkTemplateError('x', r'\90', 'x', 'invalid group reference')
    194. 

  194.         self.checkTemplateError('x', r'\99', 'x', 'invalid group reference')
    195. 

  195.         self.checkTemplateError('x', r'\118', 'x', 'invalid group reference') # r'\11' + '8'
    196. 

  196.         self.checkTemplateError('x', r'\11a', 'x', 'invalid group reference')
    197. 

  197.         self.checkTemplateError('x', r'\181', 'x', 'invalid group reference') # r'\18' + '1'
    198. 

  198.         self.checkTemplateError('x', r'\800', 'x', 'invalid group reference') # r'\80' + '0'
    199. 

  199. 

  200.         # in python2.3 (etc), these loop endlessly in sre_parser.py
    201. 

  201.         self.assertEqual(re.sub('(((((((((((x)))))))))))', r'\11', 'x'), 'x')
    202. 

  202.         self.assertEqual(re.sub('((((((((((y))))))))))(.)', r'\118', 'xyz'),
    203. 

  203.                          'xz8')
    204. 

  204.         self.assertEqual(re.sub('((((((((((y))))))))))(.)', r'\11a', 'xyz'),
    205. 

  205.                          'xza')
    206. 

  206. 

  207.     def test_qualified_re_sub(self):
    208. 

  208.         self.assertEqual(re.sub('a', 'b', 'aaaaa'), 'bbbbb')
    209. 

  209.         self.assertEqual(re.sub('a', 'b', 'aaaaa', count=1), 'baaaa')
    210. 

  210. 

  211.     def test_bug_114660(self):
    212. 

  212.         self.assertEqual(re.sub(r'(\S)\s+(\S)', r'\1 \2', 'hello  there'),
    213. 

  213.                          'hello there')
    214. 

  214. 

  215.     def test_bug_462270(self):
    216. 

  216.         # Test for empty sub() behaviour, see SF bug #462270
    217. 

  217.         self.assertEqual(re.sub('x*', '-', 'abxd'), '-a-b-d-')
    218. 

  218.         self.assertEqual(re.sub('x+', '-', 'abxd'), 'ab-d')
    219. 

  219. 

  220.     def test_symbolic_groups(self):
    221. 

  221.         re.compile('(?P<a>x)(?P=a)(?(a)y)')
    222. 

  222.         re.compile('(?P<a1>x)(?P=a1)(?(a1)y)')
    223. 

  223.         re.compile('(?P<a1>x)\1(?(1)y)')
    224. 

  224.         self.checkPatternError('(?P<a>)(?P<a>)',
    225. 

  225.                                "redefinition of group name 'a' as group 2; "
    226. 

  226.                                "was group 1")
    227. 

  227.         self.checkPatternError('(?P<a>(?P=a))',
    228. 

  228.                                "cannot refer to an open group", 10)
    229. 

  229.         self.checkPatternError('(?Pxy)', 'unknown extension ?Px')
    230. 

  230.         self.checkPatternError('(?P<a>)(?P=a', 'missing ), unterminated name', 11)
    231. 

  231.         self.checkPatternError('(?P=', 'missing group name', 4)
    232. 

  232.         self.checkPatternError('(?P=)', 'missing group name', 4)
    233. 

  233.         self.checkPatternError('(?P=1)', "bad character in group name '1'", 4)
    234. 

  234.         self.checkPatternError('(?P=a)', "unknown group name 'a'")
    235. 

  235.         self.checkPatternError('(?P=a1)', "unknown group name 'a1'")
    236. 

  236.         self.checkPatternError('(?P=a.)', "bad character in group name 'a.'", 4)
    237. 

  237.         self.checkPatternError('(?P<)', 'missing >, unterminated name', 4)
    238. 

  238.         self.checkPatternError('(?P<a', 'missing >, unterminated name', 4)
    239. 

  239.         self.checkPatternError('(?P<', 'missing group name', 4)
    240. 

  240.         self.checkPatternError('(?P<>)', 'missing group name', 4)
    241. 

  241.         self.checkPatternError(r'(?P<1>)', "bad character in group name '1'", 4)
    242. 

  242.         self.checkPatternError(r'(?P<a.>)', "bad character in group name 'a.'", 4)
    243. 

  243.         self.checkPatternError(r'(?(', 'missing group name', 3)
    244. 

  244.         self.checkPatternError(r'(?())', 'missing group name', 3)
    245. 

  245.         self.checkPatternError(r'(?(a))', "unknown group name 'a'", 3)
    246. 

  246.         self.checkPatternError(r'(?(-1))', "bad character in group name '-1'", 3)
    247. 

  247.         self.checkPatternError(r'(?(1a))', "bad character in group name '1a'", 3)
    248. 

  248.         self.checkPatternError(r'(?(a.))', "bad character in group name 'a.'", 3)
    249. 

  249.         # New valid/invalid identifiers in Python 3
    250. 

  250.         re.compile('(?P<Âľ>x)(?P=Âľ)(?(Âľ)y)')
    251. 

  251.         re.compile('(?P<𝔘𝔫𝔦𝔠𝔬𝔡𝔢>x)(?P=𝔘𝔫𝔦𝔠𝔬𝔡𝔢)(?(𝔘𝔫𝔦𝔠𝔬𝔡𝔢)y)')
    252. 

  252.         self.checkPatternError('(?P<Š>x)', "bad character in group name 'Š'", 4)
    253. 

  253.         # Support > 100 groups.
    254. 

  254.         pat = '|'.join('x(?P<a%d>%x)y' % (i, i) for i in range(1, 200 + 1))
    255. 

  255.         pat = '(?:%s)(?(200)z|t)' % pat
    256. 

  256.         self.assertEqual(re.match(pat, 'xc8yz').span(), (0, 5))
    257. 

  257. 

  258.     def test_symbolic_refs(self):
    259. 

  259.         self.checkTemplateError('(?P<a>x)', '\g<a', 'xx',
    260. 

  260.                                 'missing >, unterminated name', 3)
    261. 

  261.         self.checkTemplateError('(?P<a>x)', '\g<', 'xx',
    262. 

  262.                                 'missing group name', 3)
    263. 

  263.         self.checkTemplateError('(?P<a>x)', '\g', 'xx', 'missing <', 2)
    264. 

  264.         self.checkTemplateError('(?P<a>x)', '\g<a a>', 'xx',
    265. 

  265.                                 "bad character in group name 'a a'", 3)
    266. 

  266.         self.checkTemplateError('(?P<a>x)', '\g<>', 'xx',
    267. 

  267.                                 'missing group name', 3)
    268. 

  268.         self.checkTemplateError('(?P<a>x)', '\g<1a1>', 'xx',
    269. 

  269.                                 "bad character in group name '1a1'", 3)
    270. 

  270.         self.checkTemplateError('(?P<a>x)', r'\g<2>', 'xx',
    271. 

  271.                                 'invalid group reference')
    272. 

  272.         self.checkTemplateError('(?P<a>x)', r'\2', 'xx',
    273. 

  273.                                 'invalid group reference')
    274. 

  274.         with self.assertRaisesRegex(IndexError, "unknown group name 'ab'"):
    275. 

  275.             re.sub('(?P<a>x)', '\g<ab>', 'xx')
    276. 

  276.         self.assertEqual(re.sub('(?P<a>x)|(?P<b>y)', r'\g<b>', 'xx'), '')
    277. 

  277.         self.assertEqual(re.sub('(?P<a>x)|(?P<b>y)', r'\2', 'xx'), '')
    278. 

  278.         self.checkTemplateError('(?P<a>x)', '\g<-1>', 'xx',
    279. 

  279.                                 "bad character in group name '-1'", 3)
    280. 

  280.         # New valid/invalid identifiers in Python 3
    281. 

  281.         self.assertEqual(re.sub('(?P<Âľ>x)', r'\g<Âľ>', 'xx'), 'xx')
    282. 

  282.         self.assertEqual(re.sub('(?P<𝔘𝔫𝔦𝔠𝔬𝔡𝔢>x)', r'\g<𝔘𝔫𝔦𝔠𝔬𝔡𝔢>', 'xx'), 'xx')
    283. 

  283.         self.checkTemplateError('(?P<a>x)', '\g<Š>', 'xx',
    284. 

  284.                                 "bad character in group name 'Š'", 3)
    285. 

  285.         # Support > 100 groups.
    286. 

  286.         pat = '|'.join('x(?P<a%d>%x)y' % (i, i) for i in range(1, 200 + 1))
    287. 

  287.         self.assertEqual(re.sub(pat, '\g<200>', 'xc8yzxc8y'), 'c8zc8')
    288. 

  288. 

  289.     def test_re_subn(self):
    290. 

  290.         self.assertEqual(re.subn("(?i)b+", "x", "bbbb BBBB"), ('x x', 2))
    291. 

  291.         self.assertEqual(re.subn("b+", "x", "bbbb BBBB"), ('x BBBB', 1))
    292. 

  292.         self.assertEqual(re.subn("b+", "x", "xyz"), ('xyz', 0))
    293. 

  293.         self.assertEqual(re.subn("b*", "x", "xyz"), ('xxxyxzx', 4))
    294. 

  294.         self.assertEqual(re.subn("b*", "x", "xyz", count=2), ('xxxyz', 2))
    295. 

  295. 

  296.     def test_re_split(self):
    297. 

  297.         for string in ":a:b::c", S(":a:b::c"):
    298. 

  298.             self.assertTypedEqual(re.split(":", string),
    299. 

  299.                                   ['', 'a', 'b', '', 'c'])
    300. 

  300.             self.assertTypedEqual(re.split(":+", string),
    301. 

  301.                                   ['', 'a', 'b', 'c'])
    302. 

  302.             self.assertTypedEqual(re.split("(:+)", string),
    303. 

  303.                                   ['', ':', 'a', ':', 'b', '::', 'c'])
    304. 

  304.         for string in (b":a:b::c", B(b":a:b::c"), bytearray(b":a:b::c"),
    305. 

  305.                        memoryview(b":a:b::c")):
    306. 

  306.             self.assertTypedEqual(re.split(b":", string),
    307. 

  307.                                   [b'', b'a', b'b', b'', b'c'])
    308. 

  308.             self.assertTypedEqual(re.split(b":+", string),
    309. 

  309.                                   [b'', b'a', b'b', b'c'])
    310. 

  310.             self.assertTypedEqual(re.split(b"(:+)", string),
    311. 

  311.                                   [b'', b':', b'a', b':', b'b', b'::', b'c'])
    312. 

  312.         for a, b, c in ("\xe0\xdf\xe7", "\u0430\u0431\u0432",
    313. 

  313.                         "\U0001d49c\U0001d49e\U0001d4b5"):
    314. 

  314.             string = ":%s:%s::%s" % (a, b, c)
    315. 

  315.             self.assertEqual(re.split(":", string), ['', a, b, '', c])
    316. 

  316.             self.assertEqual(re.split(":+", string), ['', a, b, c])
    317. 

  317.             self.assertEqual(re.split("(:+)", string),
    318. 

  318.                              ['', ':', a, ':', b, '::', c])
    319. 

  319. 

  320.         self.assertEqual(re.split("(?::+)", ":a:b::c"), ['', 'a', 'b', 'c'])
    321. 

  321.         self.assertEqual(re.split("(:)+", ":a:b::c"),
    322. 

  322.                          ['', ':', 'a', ':', 'b', ':', 'c'])
    323. 

  323.         self.assertEqual(re.split("([b:]+)", ":a:b::c"),
    324. 

  324.                          ['', ':', 'a', ':b::', 'c'])
    325. 

  325.         self.assertEqual(re.split("(b)|(:+)", ":a:b::c"),
    326. 

  326.                          ['', None, ':', 'a', None, ':', '', 'b', None, '',
    327. 

  327.                           None, '::', 'c'])
    328. 

  328.         self.assertEqual(re.split("(?:b)|(?::+)", ":a:b::c"),
    329. 

  329.                          ['', 'a', '', '', 'c'])
    330. 

  330. 

  331.         for sep, expected in [
    332. 

  332.             (':*', ['', 'a', 'b', 'c']),
    333. 

  333.             ('(?::*)', ['', 'a', 'b', 'c']),
    334. 

  334.             ('(:*)', ['', ':', 'a', ':', 'b', '::', 'c']),
    335. 

  335.             ('(:)*', ['', ':', 'a', ':', 'b', ':', 'c']),
    336. 

  336.         ]:
    337. 

  337.             with self.subTest(sep=sep), self.assertWarns(FutureWarning):
    338. 

  338.                 self.assertTypedEqual(re.split(sep, ':a:b::c'), expected)
    339. 

  339. 

  340.         for sep, expected in [
    341. 

  341.             ('', [':a:b::c']),
    342. 

  342.             (r'\b', [':a:b::c']),
    343. 

  343.             (r'(?=:)', [':a:b::c']),
    344. 

  344.             (r'(?<=:)', [':a:b::c']),
    345. 

  345.         ]:
    346. 

  346.             with self.subTest(sep=sep), self.assertRaises(ValueError):
    347. 

  347.                 self.assertTypedEqual(re.split(sep, ':a:b::c'), expected)
    348. 

  348. 

  349.     def test_qualified_re_split(self):
    350. 

  350.         self.assertEqual(re.split(":", ":a:b::c", maxsplit=2), ['', 'a', 'b::c'])
    351. 

  351.         self.assertEqual(re.split(':', 'a:b:c:d', maxsplit=2), ['a', 'b', 'c:d'])
    352. 

  352.         self.assertEqual(re.split("(:)", ":a:b::c", maxsplit=2),
    353. 

  353.                          ['', ':', 'a', ':', 'b::c'])
    354. 

  354.         self.assertEqual(re.split("(:+)", ":a:b::c", maxsplit=2),
    355. 

  355.                          ['', ':', 'a', ':', 'b::c'])
    356. 

  356.         with self.assertWarns(FutureWarning):
    357. 

  357.             self.assertEqual(re.split("(:*)", ":a:b::c", maxsplit=2),
    358. 

  358.                              ['', ':', 'a', ':', 'b::c'])
    359. 

  359. 

  360.     def test_re_findall(self):
    361. 

  361.         self.assertEqual(re.findall(":+", "abc"), [])
    362. 

  362.         for string in "a:b::c:::d", S("a:b::c:::d"):
    363. 

  363.             self.assertTypedEqual(re.findall(":+", string),
    364. 

  364.                                   [":", "::", ":::"])
    365. 

  365.             self.assertTypedEqual(re.findall("(:+)", string),
    366. 

  366.                                   [":", "::", ":::"])
    367. 

  367.             self.assertTypedEqual(re.findall("(:)(:*)", string),
    368. 

  368.                                   [(":", ""), (":", ":"), (":", "::")])
    369. 

  369.         for string in (b"a:b::c:::d", B(b"a:b::c:::d"), bytearray(b"a:b::c:::d"),
    370. 

  370.                        memoryview(b"a:b::c:::d")):
    371. 

  371.             self.assertTypedEqual(re.findall(b":+", string),
    372. 

  372.                                   [b":", b"::", b":::"])
    373. 

  373.             self.assertTypedEqual(re.findall(b"(:+)", string),
    374. 

  374.                                   [b":", b"::", b":::"])
    375. 

  375.             self.assertTypedEqual(re.findall(b"(:)(:*)", string),
    376. 

  376.                                   [(b":", b""), (b":", b":"), (b":", b"::")])
    377. 

  377.         for x in ("\xe0", "\u0430", "\U0001d49c"):
    378. 

  378.             xx = x * 2
    379. 

  379.             xxx = x * 3
    380. 

  380.             string = "a%sb%sc%sd" % (x, xx, xxx)
    381. 

  381.             self.assertEqual(re.findall("%s+" % x, string), [x, xx, xxx])
    382. 

  382.             self.assertEqual(re.findall("(%s+)" % x, string), [x, xx, xxx])
    383. 

  383.             self.assertEqual(re.findall("(%s)(%s*)" % (x, x), string),
    384. 

  384.                              [(x, ""), (x, x), (x, xx)])
    385. 

  385. 

  386.     def test_bug_117612(self):
    387. 

  387.         self.assertEqual(re.findall(r"(a|(b))", "aba"),
    388. 

  388.                          [("a", ""),("b", "b"),("a", "")])
    389. 

  389. 

  390.     def test_re_match(self):
    391. 

  391.         for string in 'a', S('a'):
    392. 

  392.             self.assertEqual(re.match('a', string).groups(), ())
    393. 

  393.             self.assertEqual(re.match('(a)', string).groups(), ('a',))
    394. 

  394.             self.assertEqual(re.match('(a)', string).group(0), 'a')
    395. 

  395.             self.assertEqual(re.match('(a)', string).group(1), 'a')
    396. 

  396.             self.assertEqual(re.match('(a)', string).group(1, 1), ('a', 'a'))
    397. 

  397.         for string in b'a', B(b'a'), bytearray(b'a'), memoryview(b'a'):
    398. 

  398.             self.assertEqual(re.match(b'a', string).groups(), ())
    399. 

  399.             self.assertEqual(re.match(b'(a)', string).groups(), (b'a',))
    400. 

  400.             self.assertEqual(re.match(b'(a)', string).group(0), b'a')
    401. 

  401.             self.assertEqual(re.match(b'(a)', string).group(1), b'a')
    402. 

  402.             self.assertEqual(re.match(b'(a)', string).group(1, 1), (b'a', b'a'))
    403. 

  403.         for a in ("\xe0", "\u0430", "\U0001d49c"):
    404. 

  404.             self.assertEqual(re.match(a, a).groups(), ())
    405. 

  405.             self.assertEqual(re.match('(%s)' % a, a).groups(), (a,))
    406. 

  406.             self.assertEqual(re.match('(%s)' % a, a).group(0), a)
    407. 

  407.             self.assertEqual(re.match('(%s)' % a, a).group(1), a)
    408. 

  408.             self.assertEqual(re.match('(%s)' % a, a).group(1, 1), (a, a))
    409. 

  409. 

  410.         pat = re.compile('((a)|(b))(c)?')
    411. 

  411.         self.assertEqual(pat.match('a').groups(), ('a', 'a', None, None))
    412. 

  412.         self.assertEqual(pat.match('b').groups(), ('b', None, 'b', None))
    413. 

  413.         self.assertEqual(pat.match('ac').groups(), ('a', 'a', None, 'c'))
    414. 

  414.         self.assertEqual(pat.match('bc').groups(), ('b', None, 'b', 'c'))
    415. 

  415.         self.assertEqual(pat.match('bc').groups(""), ('b', "", 'b', 'c'))
    416. 

  416. 

  417.         pat = re.compile('(?:(?P<a1>a)|(?P<b2>b))(?P<c3>c)?')
    418. 

  418.         self.assertEqual(pat.match('a').group(1, 2, 3), ('a', None, None))
    419. 

  419.         self.assertEqual(pat.match('b').group('a1', 'b2', 'c3'),
    420. 

  420.                          (None, 'b', None))
    421. 

  421.         self.assertEqual(pat.match('ac').group(1, 'b2', 3), ('a', None, 'c'))
    422. 

  422. 

  423.     def test_group(self):
    424. 

  424.         class Index:
    425. 

  425.             def __init__(self, value):
    426. 

  426.                 self.value = value
    427. 

  427.             def __index__(self):
    428. 

  428.                 return self.value
    429. 

  429.         # A single group
    430. 

  430.         m = re.match('(a)(b)', 'ab')
    431. 

  431.         self.assertEqual(m.group(), 'ab')
    432. 

  432.         self.assertEqual(m.group(0), 'ab')
    433. 

  433.         self.assertEqual(m.group(1), 'a')
    434. 

  434.         self.assertEqual(m.group(Index(1)), 'a')
    435. 

  435.         self.assertRaises(IndexError, m.group, -1)
    436. 

  436.         self.assertRaises(IndexError, m.group, 3)
    437. 

  437.         self.assertRaises(IndexError, m.group, 1<<1000)
    438. 

  438.         self.assertRaises(IndexError, m.group, Index(1<<1000))
    439. 

  439.         self.assertRaises(IndexError, m.group, 'x')
    440. 

  440.         # Multiple groups
    441. 

  441.         self.assertEqual(m.group(2, 1), ('b', 'a'))
    442. 

  442.         self.assertEqual(m.group(Index(2), Index(1)), ('b', 'a'))
    443. 

  443. 

  444.     def test_re_fullmatch(self):
    445. 

  445.         # Issue 16203: Proposal: add re.fullmatch() method.
    446. 

  446.         self.assertEqual(re.fullmatch(r"a", "a").span(), (0, 1))
    447. 

  447.         for string in "ab", S("ab"):
    448. 

  448.             self.assertEqual(re.fullmatch(r"a|ab", string).span(), (0, 2))
    449. 

  449.         for string in b"ab", B(b"ab"), bytearray(b"ab"), memoryview(b"ab"):
    450. 

  450.             self.assertEqual(re.fullmatch(br"a|ab", string).span(), (0, 2))
    451. 

  451.         for a, b in "\xe0\xdf", "\u0430\u0431", "\U0001d49c\U0001d49e":
    452. 

  452.             r = r"%s|%s" % (a, a + b)
    453. 

  453.             self.assertEqual(re.fullmatch(r, a + b).span(), (0, 2))
    454. 

  454.         self.assertEqual(re.fullmatch(r".*?$", "abc").span(), (0, 3))
    455. 

  455.         self.assertEqual(re.fullmatch(r".*?", "abc").span(), (0, 3))
    456. 

  456.         self.assertEqual(re.fullmatch(r"a.*?b", "ab").span(), (0, 2))
    457. 

  457.         self.assertEqual(re.fullmatch(r"a.*?b", "abb").span(), (0, 3))
    458. 

  458.         self.assertEqual(re.fullmatch(r"a.*?b", "axxb").span(), (0, 4))
    459. 

  459.         self.assertIsNone(re.fullmatch(r"a+", "ab"))
    460. 

  460.         self.assertIsNone(re.fullmatch(r"abc$", "abc\n"))
    461. 

  461.         self.assertIsNone(re.fullmatch(r"abc\Z", "abc\n"))
    462. 

  462.         self.assertIsNone(re.fullmatch(r"(?m)abc$", "abc\n"))
    463. 

  463.         self.assertEqual(re.fullmatch(r"ab(?=c)cd", "abcd").span(), (0, 4))
    464. 

  464.         self.assertEqual(re.fullmatch(r"ab(?<=b)cd", "abcd").span(), (0, 4))
    465. 

  465.         self.assertEqual(re.fullmatch(r"(?=a|ab)ab", "ab").span(), (0, 2))
    466. 

  466. 

  467.         self.assertEqual(
    468. 

  468.             re.compile(r"bc").fullmatch("abcd", pos=1, endpos=3).span(), (1, 3))
    469. 

  469.         self.assertEqual(
    470. 

  470.             re.compile(r".*?$").fullmatch("abcd", pos=1, endpos=3).span(), (1, 3))
    471. 

  471.         self.assertEqual(
    472. 

  472.             re.compile(r".*?").fullmatch("abcd", pos=1, endpos=3).span(), (1, 3))
    473. 

  473. 

  474.     def test_re_groupref_exists(self):
    475. 

  475.         self.assertEqual(re.match('^(\()?([^()]+)(?(1)\))$', '(a)').groups(),
    476. 

  476.                          ('(', 'a'))
    477. 

  477.         self.assertEqual(re.match('^(\()?([^()]+)(?(1)\))$', 'a').groups(),
    478. 

  478.                          (None, 'a'))
    479. 

  479.         self.assertIsNone(re.match('^(\()?([^()]+)(?(1)\))$', 'a)'))
    480. 

  480.         self.assertIsNone(re.match('^(\()?([^()]+)(?(1)\))$', '(a'))
    481. 

  481.         self.assertEqual(re.match('^(?:(a)|c)((?(1)b|d))$', 'ab').groups(),
    482. 

  482.                          ('a', 'b'))
    483. 

  483.         self.assertEqual(re.match('^(?:(a)|c)((?(1)b|d))$', 'cd').groups(),
    484. 

  484.                          (None, 'd'))
    485. 

  485.         self.assertEqual(re.match('^(?:(a)|c)((?(1)|d))$', 'cd').groups(),
    486. 

  486.                          (None, 'd'))
    487. 

  487.         self.assertEqual(re.match('^(?:(a)|c)((?(1)|d))$', 'a').groups(),
    488. 

  488.                          ('a', ''))
    489. 

  489. 

  490.         # Tests for bug #1177831: exercise groups other than the first group
    491. 

  491.         p = re.compile('(?P<g1>a)(?P<g2>b)?((?(g2)c|d))')
    492. 

  492.         self.assertEqual(p.match('abc').groups(),
    493. 

  493.                          ('a', 'b', 'c'))
    494. 

  494.         self.assertEqual(p.match('ad').groups(),
    495. 

  495.                          ('a', None, 'd'))
    496. 

  496.         self.assertIsNone(p.match('abd'))
    497. 

  497.         self.assertIsNone(p.match('ac'))
    498. 

  498. 

  499.         # Support > 100 groups.
    500. 

  500.         pat = '|'.join('x(?P<a%d>%x)y' % (i, i) for i in range(1, 200 + 1))
    501. 

  501.         pat = '(?:%s)(?(200)z)' % pat
    502. 

  502.         self.assertEqual(re.match(pat, 'xc8yz').span(), (0, 5))
    503. 

  503. 

  504.         self.checkPatternError(r'(?P<a>)(?(0))', 'bad group number', 10)
    505. 

  505.         self.checkPatternError(r'()(?(1)a|b',
    506. 

  506.                                'missing ), unterminated subpattern', 2)
    507. 

  507.         self.checkPatternError(r'()(?(1)a|b|c)',
    508. 

  508.                                'conditional backref with more than '
    509. 

  509.                                'two branches', 10)
    510. 

  510. 

  511.     def test_re_groupref_overflow(self):
    512. 

  512.         self.checkTemplateError('()', '\g<%s>' % sre_constants.MAXGROUPS, 'xx',
    513. 

  513.                                 'invalid group reference', 3)
    514. 

  514.         self.checkPatternError(r'(?P<a>)(?(%d))' % sre_constants.MAXGROUPS,
    515. 

  515.                                'invalid group reference', 10)
    516. 

  516. 

  517.     def test_re_groupref(self):
    518. 

  518.         self.assertEqual(re.match(r'^(\|)?([^()]+)\1$', '|a|').groups(),
    519. 

  519.                          ('|', 'a'))
    520. 

  520.         self.assertEqual(re.match(r'^(\|)?([^()]+)\1?$', 'a').groups(),
    521. 

  521.                          (None, 'a'))
    522. 

  522.         self.assertIsNone(re.match(r'^(\|)?([^()]+)\1$', 'a|'))
    523. 

  523.         self.assertIsNone(re.match(r'^(\|)?([^()]+)\1$', '|a'))
    524. 

  524.         self.assertEqual(re.match(r'^(?:(a)|c)(\1)$', 'aa').groups(),
    525. 

  525.                          ('a', 'a'))
    526. 

  526.         self.assertEqual(re.match(r'^(?:(a)|c)(\1)?$', 'c').groups(),
    527. 

  527.                          (None, None))
    528. 

  528. 

  529.         self.checkPatternError(r'(abc\1)', 'cannot refer to an open group', 4)
    530. 

  530. 

  531.     def test_groupdict(self):
    532. 

  532.         self.assertEqual(re.match('(?P<first>first) (?P<second>second)',
    533. 

  533.                                   'first second').groupdict(),
    534. 

  534.                          {'first':'first', 'second':'second'})
    535. 

  535. 

  536.     def test_expand(self):
    537. 

  537.         self.assertEqual(re.match("(?P<first>first) (?P<second>second)",
    538. 

  538.                                   "first second")
    539. 

  539.                                   .expand(r"\2 \1 \g<second> \g<first>"),
    540. 

  540.                          "second first second first")
    541. 

  541.         self.assertEqual(re.match("(?P<first>first)|(?P<second>second)",
    542. 

  542.                                   "first")
    543. 

  543.                                   .expand(r"\2 \g<second>"),
    544. 

  544.                          " ")
    545. 

  545. 

  546.     def test_repeat_minmax(self):
    547. 

  547.         self.assertIsNone(re.match("^(\w){1}$", "abc"))
    548. 

  548.         self.assertIsNone(re.match("^(\w){1}?$", "abc"))
    549. 

  549.         self.assertIsNone(re.match("^(\w){1,2}$", "abc"))
    550. 

  550.         self.assertIsNone(re.match("^(\w){1,2}?$", "abc"))
    551. 

  551. 

  552.         self.assertEqual(re.match("^(\w){3}$", "abc").group(1), "c")
    553. 

  553.         self.assertEqual(re.match("^(\w){1,3}$", "abc").group(1), "c")
    554. 

  554.         self.assertEqual(re.match("^(\w){1,4}$", "abc").group(1), "c")
    555. 

  555.         self.assertEqual(re.match("^(\w){3,4}?$", "abc").group(1), "c")
    556. 

  556.         self.assertEqual(re.match("^(\w){3}?$", "abc").group(1), "c")
    557. 

  557.         self.assertEqual(re.match("^(\w){1,3}?$", "abc").group(1), "c")
    558. 

  558.         self.assertEqual(re.match("^(\w){1,4}?$", "abc").group(1), "c")
    559. 

  559.         self.assertEqual(re.match("^(\w){3,4}?$", "abc").group(1), "c")
    560. 

  560. 

  561.         self.assertIsNone(re.match("^x{1}$", "xxx"))
    562. 

  562.         self.assertIsNone(re.match("^x{1}?$", "xxx"))
    563. 

  563.         self.assertIsNone(re.match("^x{1,2}$", "xxx"))
    564. 

  564.         self.assertIsNone(re.match("^x{1,2}?$", "xxx"))
    565. 

  565. 

  566.         self.assertTrue(re.match("^x{3}$", "xxx"))
    567. 

  567.         self.assertTrue(re.match("^x{1,3}$", "xxx"))
    568. 

  568.         self.assertTrue(re.match("^x{3,3}$", "xxx"))
    569. 

  569.         self.assertTrue(re.match("^x{1,4}$", "xxx"))
    570. 

  570.         self.assertTrue(re.match("^x{3,4}?$", "xxx"))
    571. 

  571.         self.assertTrue(re.match("^x{3}?$", "xxx"))
    572. 

  572.         self.assertTrue(re.match("^x{1,3}?$", "xxx"))
    573. 

  573.         self.assertTrue(re.match("^x{1,4}?$", "xxx"))
    574. 

  574.         self.assertTrue(re.match("^x{3,4}?$", "xxx"))
    575. 

  575. 

  576.         self.assertIsNone(re.match("^x{}$", "xxx"))
    577. 

  577.         self.assertTrue(re.match("^x{}$", "x{}"))
    578. 

  578. 

  579.         self.checkPatternError(r'x{2,1}',
    580. 

  580.                                'min repeat greater than max repeat', 2)
    581. 

  581. 

  582.     def test_getattr(self):
    583. 

  583.         self.assertEqual(re.compile("(?i)(a)(b)").pattern, "(?i)(a)(b)")
    584. 

  584.         self.assertEqual(re.compile("(?i)(a)(b)").flags, re.I | re.U)
    585. 

  585.         self.assertEqual(re.compile("(?i)(a)(b)").groups, 2)
    586. 

  586.         self.assertEqual(re.compile("(?i)(a)(b)").groupindex, {})
    587. 

  587.         self.assertEqual(re.compile("(?i)(?P<first>a)(?P<other>b)").groupindex,
    588. 

  588.                          {'first': 1, 'other': 2})
    589. 

  589. 

  590.         self.assertEqual(re.match("(a)", "a").pos, 0)
    591. 

  591.         self.assertEqual(re.match("(a)", "a").endpos, 1)
    592. 

  592.         self.assertEqual(re.match("(a)", "a").string, "a")
    593. 

  593.         self.assertEqual(re.match("(a)", "a").regs, ((0, 1), (0, 1)))
    594. 

  594.         self.assertTrue(re.match("(a)", "a").re)
    595. 

  595. 

  596.         # Issue 14260. groupindex should be non-modifiable mapping.
    597. 

  597.         p = re.compile(r'(?i)(?P<first>a)(?P<other>b)')
    598. 

  598.         self.assertEqual(sorted(p.groupindex), ['first', 'other'])
    599. 

  599.         self.assertEqual(p.groupindex['other'], 2)
    600. 

  600.         with self.assertRaises(TypeError):
    601. 

  601.             p.groupindex['other'] = 0
    602. 

  602.         self.assertEqual(p.groupindex['other'], 2)
    603. 

  603. 

  604.     def test_special_escapes(self):
    605. 

  605.         self.assertEqual(re.search(r"\b(b.)\b",
    606. 

  606.                                    "abcd abc bcd bx").group(1), "bx")
    607. 

  607.         self.assertEqual(re.search(r"\B(b.)\B",
    608. 

  608.                                    "abc bcd bc abxd").group(1), "bx")
    609. 

  609.         self.assertEqual(re.search(r"\b(b.)\b",
    610. 

  610.                                    "abcd abc bcd bx", re.ASCII).group(1), "bx")
    611. 

  611.         self.assertEqual(re.search(r"\B(b.)\B",
    612. 

  612.                                    "abc bcd bc abxd", re.ASCII).group(1), "bx")
    613. 

  613.         self.assertEqual(re.search(r"^abc$", "\nabc\n", re.M).group(0), "abc")
    614. 

  614.         self.assertEqual(re.search(r"^\Aabc\Z$", "abc", re.M).group(0), "abc")
    615. 

  615.         self.assertIsNone(re.search(r"^\Aabc\Z$", "\nabc\n", re.M))
    616. 

  616.         self.assertEqual(re.search(br"\b(b.)\b",
    617. 

  617.                                    b"abcd abc bcd bx").group(1), b"bx")
    618. 

  618.         self.assertEqual(re.search(br"\B(b.)\B",
    619. 

  619.                                    b"abc bcd bc abxd").group(1), b"bx")
    620. 

  620.         self.assertEqual(re.search(br"\b(b.)\b",
    621. 

  621.                                    b"abcd abc bcd bx", re.LOCALE).group(1), b"bx")
    622. 

  622.         self.assertEqual(re.search(br"\B(b.)\B",
    623. 

  623.                                    b"abc bcd bc abxd", re.LOCALE).group(1), b"bx")
    624. 

  624.         self.assertEqual(re.search(br"^abc$", b"\nabc\n", re.M).group(0), b"abc")
    625. 

  625.         self.assertEqual(re.search(br"^\Aabc\Z$", b"abc", re.M).group(0), b"abc")
    626. 

  626.         self.assertIsNone(re.search(br"^\Aabc\Z$", b"\nabc\n", re.M))
    627. 

  627.         self.assertEqual(re.search(r"\d\D\w\W\s\S",
    628. 

  628.                                    "1aa! a").group(0), "1aa! a")
    629. 

  629.         self.assertEqual(re.search(br"\d\D\w\W\s\S",
    630. 

  630.                                    b"1aa! a").group(0), b"1aa! a")
    631. 

  631.         self.assertEqual(re.search(r"\d\D\w\W\s\S",
    632. 

  632.                                    "1aa! a", re.ASCII).group(0), "1aa! a")
    633. 

  633.         self.assertEqual(re.search(br"\d\D\w\W\s\S",
    634. 

  634.                                    b"1aa! a", re.LOCALE).group(0), b"1aa! a")
    635. 

  635. 

  636.     def test_other_escapes(self):
    637. 

  637.         self.checkPatternError("\\", 'bad escape (end of pattern)', 0)
    638. 

  638.         self.assertEqual(re.match(r"\(", '(').group(), '(')
    639. 

  639.         self.assertIsNone(re.match(r"\(", ')'))
    640. 

  640.         self.assertEqual(re.match(r"\\", '\\').group(), '\\')
    641. 

  641.         self.assertEqual(re.match(r"[\]]", ']').group(), ']')
    642. 

  642.         self.assertIsNone(re.match(r"[\]]", '['))
    643. 

  643.         self.assertEqual(re.match(r"[a\-c]", '-').group(), '-')
    644. 

  644.         self.assertIsNone(re.match(r"[a\-c]", 'b'))
    645. 

  645.         self.assertEqual(re.match(r"[\^a]+", 'a^').group(), 'a^')
    646. 

  646.         self.assertIsNone(re.match(r"[\^a]+", 'b'))
    647. 

  647.         re.purge()  # for warnings
    648. 

  648.         for c in 'ceghijklmopqyzCEFGHIJKLMNOPQRTVXY':
    649. 

  649.             with self.subTest(c):
    650. 

  650.                 self.assertRaises(re.error, re.compile, '\\%c' % c)
    651. 

  651.         for c in 'ceghijklmopqyzABCEFGHIJKLMNOPQRTVXYZ':
    652. 

  652.             with self.subTest(c):
    653. 

  653.                 self.assertRaises(re.error, re.compile, '[\\%c]' % c)
    654. 

  654. 

  655.     def test_string_boundaries(self):
    656. 

  656.         # See http://bugs.python.org/issue10713
    657. 

  657.         self.assertEqual(re.search(r"\b(abc)\b", "abc").group(1),
    658. 

  658.                          "abc")
    659. 

  659.         # There's a word boundary at the start of a string.
    660. 

  660.         self.assertTrue(re.match(r"\b", "abc"))
    661. 

  661.         # A non-empty string includes a non-boundary zero-length match.
    662. 

  662.         self.assertTrue(re.search(r"\B", "abc"))
    663. 

  663.         # There is no non-boundary match at the start of a string.
    664. 

  664.         self.assertFalse(re.match(r"\B", "abc"))
    665. 

  665.         # However, an empty string contains no word boundaries, and also no
    666. 

  666.         # non-boundaries.
    667. 

  667.         self.assertIsNone(re.search(r"\B", ""))
    668. 

  668.         # This one is questionable and different from the perlre behaviour,
    669. 

  669.         # but describes current behavior.
    670. 

  670.         self.assertIsNone(re.search(r"\b", ""))
    671. 

  671.         # A single word-character string has two boundaries, but no
    672. 

  672.         # non-boundary gaps.
    673. 

  673.         self.assertEqual(len(re.findall(r"\b", "a")), 2)
    674. 

  674.         self.assertEqual(len(re.findall(r"\B", "a")), 0)
    675. 

  675.         # If there are no words, there are no boundaries
    676. 

  676.         self.assertEqual(len(re.findall(r"\b", " ")), 0)
    677. 

  677.         self.assertEqual(len(re.findall(r"\b", "   ")), 0)
    678. 

  678.         # Can match around the whitespace.
    679. 

  679.         self.assertEqual(len(re.findall(r"\B", " ")), 2)
    680. 

  680. 

  681.     def test_bigcharset(self):
    682. 

  682.         self.assertEqual(re.match("([\u2222\u2223])",
    683. 

  683.                                   "\u2222").group(1), "\u2222")
    684. 

  684.         r = '[%s]' % ''.join(map(chr, range(256, 2**16, 255)))
    685. 

  685.         self.assertEqual(re.match(r, "\uff01").group(), "\uff01")
    686. 

  686. 

  687.     def test_big_codesize(self):
    688. 

  688.         # Issue #1160
    689. 

  689.         r = re.compile('|'.join(('%d'%x for x in range(10000))))
    690. 

  690.         self.assertTrue(r.match('1000'))
    691. 

  691.         self.assertTrue(r.match('9999'))
    692. 

  692. 

  693.     def test_anyall(self):
    694. 

  694.         self.assertEqual(re.match("a.b", "a\nb", re.DOTALL).group(0),
    695. 

  695.                          "a\nb")
    696. 

  696.         self.assertEqual(re.match("a.*b", "a\n\nb", re.DOTALL).group(0),
    697. 

  697.                          "a\n\nb")
    698. 

  698. 

  699.     def test_lookahead(self):
    700. 

  700.         self.assertEqual(re.match("(a(?=\s[^a]))", "a b").group(1), "a")
    701. 

  701.         self.assertEqual(re.match("(a(?=\s[^a]*))", "a b").group(1), "a")
    702. 

  702.         self.assertEqual(re.match("(a(?=\s[abc]))", "a b").group(1), "a")
    703. 

  703.         self.assertEqual(re.match("(a(?=\s[abc]*))", "a bc").group(1), "a")
    704. 

  704.         self.assertEqual(re.match(r"(a)(?=\s\1)", "a a").group(1), "a")
    705. 

  705.         self.assertEqual(re.match(r"(a)(?=\s\1*)", "a aa").group(1), "a")
    706. 

  706.         self.assertEqual(re.match(r"(a)(?=\s(abc|a))", "a a").group(1), "a")
    707. 

  707. 

  708.         self.assertEqual(re.match(r"(a(?!\s[^a]))", "a a").group(1), "a")
    709. 

  709.         self.assertEqual(re.match(r"(a(?!\s[abc]))", "a d").group(1), "a")
    710. 

  710.         self.assertEqual(re.match(r"(a)(?!\s\1)", "a b").group(1), "a")
    711. 

  711.         self.assertEqual(re.match(r"(a)(?!\s(abc|a))", "a b").group(1), "a")
    712. 

  712. 

  713.         # Group reference.
    714. 

  714.         self.assertTrue(re.match(r'(a)b(?=\1)a', 'aba'))
    715. 

  715.         self.assertIsNone(re.match(r'(a)b(?=\1)c', 'abac'))
    716. 

  716.         # Conditional group reference.
    717. 

  717.         self.assertTrue(re.match(r'(?:(a)|(x))b(?=(?(2)x|c))c', 'abc'))
    718. 

  718.         self.assertIsNone(re.match(r'(?:(a)|(x))b(?=(?(2)c|x))c', 'abc'))
    719. 

  719.         self.assertTrue(re.match(r'(?:(a)|(x))b(?=(?(2)x|c))c', 'abc'))
    720. 

  720.         self.assertIsNone(re.match(r'(?:(a)|(x))b(?=(?(1)b|x))c', 'abc'))
    721. 

  721.         self.assertTrue(re.match(r'(?:(a)|(x))b(?=(?(1)c|x))c', 'abc'))
    722. 

  722.         # Group used before defined.
    723. 

  723.         self.assertTrue(re.match(r'(a)b(?=(?(2)x|c))(c)', 'abc'))
    724. 

  724.         self.assertIsNone(re.match(r'(a)b(?=(?(2)b|x))(c)', 'abc'))
    725. 

  725.         self.assertTrue(re.match(r'(a)b(?=(?(1)c|x))(c)', 'abc'))
    726. 

  726. 

  727.     def test_lookbehind(self):
    728. 

  728.         self.assertTrue(re.match(r'ab(?<=b)c', 'abc'))
    729. 

  729.         self.assertIsNone(re.match(r'ab(?<=c)c', 'abc'))
    730. 

  730.         self.assertIsNone(re.match(r'ab(?<!b)c', 'abc'))
    731. 

  731.         self.assertTrue(re.match(r'ab(?<!c)c', 'abc'))
    732. 

  732.         # Group reference.
    733. 

  733.         self.assertTrue(re.match(r'(a)a(?<=\1)c', 'aac'))
    734. 

  734.         self.assertIsNone(re.match(r'(a)b(?<=\1)a', 'abaa'))
    735. 

  735.         self.assertIsNone(re.match(r'(a)a(?<!\1)c', 'aac'))
    736. 

  736.         self.assertTrue(re.match(r'(a)b(?<!\1)a', 'abaa'))
    737. 

  737.         # Conditional group reference.
    738. 

  738.         self.assertIsNone(re.match(r'(?:(a)|(x))b(?<=(?(2)x|c))c', 'abc'))
    739. 

  739.         self.assertIsNone(re.match(r'(?:(a)|(x))b(?<=(?(2)b|x))c', 'abc'))
    740. 

  740.         self.assertTrue(re.match(r'(?:(a)|(x))b(?<=(?(2)x|b))c', 'abc'))
    741. 

  741.         self.assertIsNone(re.match(r'(?:(a)|(x))b(?<=(?(1)c|x))c', 'abc'))
    742. 

  742.         self.assertTrue(re.match(r'(?:(a)|(x))b(?<=(?(1)b|x))c', 'abc'))
    743. 

  743.         # Group used before defined.
    744. 

  744.         self.assertRaises(re.error, re.compile, r'(a)b(?<=(?(2)b|x))(c)')
    745. 

  745.         self.assertIsNone(re.match(r'(a)b(?<=(?(1)c|x))(c)', 'abc'))
    746. 

  746.         self.assertTrue(re.match(r'(a)b(?<=(?(1)b|x))(c)', 'abc'))
    747. 

  747.         # Group defined in the same lookbehind pattern
    748. 

  748.         self.assertRaises(re.error, re.compile, r'(a)b(?<=(.)\2)(c)')
    749. 

  749.         self.assertRaises(re.error, re.compile, r'(a)b(?<=(?P<a>.)(?P=a))(c)')
    750. 

  750.         self.assertRaises(re.error, re.compile, r'(a)b(?<=(a)(?(2)b|x))(c)')
    751. 

  751.         self.assertRaises(re.error, re.compile, r'(a)b(?<=(.)(?<=\2))(c)')
    752. 

  752. 

  753.     def test_ignore_case(self):
    754. 

  754.         self.assertEqual(re.match("abc", "ABC", re.I).group(0), "ABC")
    755. 

  755.         self.assertEqual(re.match(b"abc", b"ABC", re.I).group(0), b"ABC")
    756. 

  756.         self.assertEqual(re.match(r"(a\s[^a])", "a b", re.I).group(1), "a b")
    757. 

  757.         self.assertEqual(re.match(r"(a\s[^a]*)", "a bb", re.I).group(1), "a bb")
    758. 

  758.         self.assertEqual(re.match(r"(a\s[abc])", "a b", re.I).group(1), "a b")
    759. 

  759.         self.assertEqual(re.match(r"(a\s[abc]*)", "a bb", re.I).group(1), "a bb")
    760. 

  760.         self.assertEqual(re.match(r"((a)\s\2)", "a a", re.I).group(1), "a a")
    761. 

  761.         self.assertEqual(re.match(r"((a)\s\2*)", "a aa", re.I).group(1), "a aa")
    762. 

  762.         self.assertEqual(re.match(r"((a)\s(abc|a))", "a a", re.I).group(1), "a a")
    763. 

  763.         self.assertEqual(re.match(r"((a)\s(abc|a)*)", "a aa", re.I).group(1), "a aa")
    764. 

  764. 

  765.         assert '\u212a'.lower() == 'k' # 'K'
    766. 

  766.         self.assertTrue(re.match(r'K', '\u212a', re.I))
    767. 

  767.         self.assertTrue(re.match(r'k', '\u212a', re.I))
    768. 

  768.         self.assertTrue(re.match(r'\u212a', 'K', re.I))
    769. 

  769.         self.assertTrue(re.match(r'\u212a', 'k', re.I))
    770. 

  770.         assert '\u017f'.upper() == 'S' # 'Ĺż'
    771. 

  771.         self.assertTrue(re.match(r'S', '\u017f', re.I))
    772. 

  772.         self.assertTrue(re.match(r's', '\u017f', re.I))
    773. 

  773.         self.assertTrue(re.match(r'\u017f', 'S', re.I))
    774. 

  774.         self.assertTrue(re.match(r'\u017f', 's', re.I))
    775. 

  775.         assert '\ufb05'.upper() == '\ufb06'.upper() == 'ST' # 'ſt', 'st'
    776. 

  776.         self.assertTrue(re.match(r'\ufb05', '\ufb06', re.I))
    777. 

  777.         self.assertTrue(re.match(r'\ufb06', '\ufb05', re.I))
    778. 

  778. 

  779.     def test_ignore_case_set(self):
    780. 

  780.         self.assertTrue(re.match(r'[19A]', 'A', re.I))
    781. 

  781.         self.assertTrue(re.match(r'[19a]', 'a', re.I))
    782. 

  782.         self.assertTrue(re.match(r'[19a]', 'A', re.I))
    783. 

  783.         self.assertTrue(re.match(r'[19A]', 'a', re.I))
    784. 

  784.         self.assertTrue(re.match(br'[19A]', b'A', re.I))
    785. 

  785.         self.assertTrue(re.match(br'[19a]', b'a', re.I))
    786. 

  786.         self.assertTrue(re.match(br'[19a]', b'A', re.I))
    787. 

  787.         self.assertTrue(re.match(br'[19A]', b'a', re.I))
    788. 

  788.         assert '\u212a'.lower() == 'k' # 'K'
    789. 

  789.         self.assertTrue(re.match(r'[19K]', '\u212a', re.I))
    790. 

  790.         self.assertTrue(re.match(r'[19k]', '\u212a', re.I))
    791. 

  791.         self.assertTrue(re.match(r'[19\u212a]', 'K', re.I))
    792. 

  792.         self.assertTrue(re.match(r'[19\u212a]', 'k', re.I))
    793. 

  793.         assert '\u017f'.upper() == 'S' # 'Ĺż'
    794. 

  794.         self.assertTrue(re.match(r'[19S]', '\u017f', re.I))
    795. 

  795.         self.assertTrue(re.match(r'[19s]', '\u017f', re.I))
    796. 

  796.         self.assertTrue(re.match(r'[19\u017f]', 'S', re.I))
    797. 

  797.         self.assertTrue(re.match(r'[19\u017f]', 's', re.I))
    798. 

  798.         assert '\ufb05'.upper() == '\ufb06'.upper() == 'ST' # 'ſt', 'st'
    799. 

  799.         self.assertTrue(re.match(r'[19\ufb05]', '\ufb06', re.I))
    800. 

  800.         self.assertTrue(re.match(r'[19\ufb06]', '\ufb05', re.I))
    801. 

  801. 

  802.     def test_ignore_case_range(self):
    803. 

  803.         # Issues #3511, #17381.
    804. 

  804.         self.assertTrue(re.match(r'[9-a]', '_', re.I))
    805. 

  805.         self.assertIsNone(re.match(r'[9-A]', '_', re.I))
    806. 

  806.         self.assertTrue(re.match(br'[9-a]', b'_', re.I))
    807. 

  807.         self.assertIsNone(re.match(br'[9-A]', b'_', re.I))
    808. 

  808.         self.assertTrue(re.match(r'[\xc0-\xde]', '\xd7', re.I))
    809. 

  809.         self.assertIsNone(re.match(r'[\xc0-\xde]', '\xf7', re.I))
    810. 

  810.         self.assertTrue(re.match(r'[\xe0-\xfe]', '\xf7', re.I))
    811. 

  811.         self.assertIsNone(re.match(r'[\xe0-\xfe]', '\xd7', re.I))
    812. 

  812.         self.assertTrue(re.match(r'[\u0430-\u045f]', '\u0450', re.I))
    813. 

  813.         self.assertTrue(re.match(r'[\u0430-\u045f]', '\u0400', re.I))
    814. 

  814.         self.assertTrue(re.match(r'[\u0400-\u042f]', '\u0450', re.I))
    815. 

  815.         self.assertTrue(re.match(r'[\u0400-\u042f]', '\u0400', re.I))
    816. 

  816.         self.assertTrue(re.match(r'[\U00010428-\U0001044f]', '\U00010428', re.I))
    817. 

  817.         self.assertTrue(re.match(r'[\U00010428-\U0001044f]', '\U00010400', re.I))
    818. 

  818.         self.assertTrue(re.match(r'[\U00010400-\U00010427]', '\U00010428', re.I))
    819. 

  819.         self.assertTrue(re.match(r'[\U00010400-\U00010427]', '\U00010400', re.I))
    820. 

  820. 

  821.         assert '\u212a'.lower() == 'k' # 'K'
    822. 

  822.         self.assertTrue(re.match(r'[J-M]', '\u212a', re.I))
    823. 

  823.         self.assertTrue(re.match(r'[j-m]', '\u212a', re.I))
    824. 

  824.         self.assertTrue(re.match(r'[\u2129-\u212b]', 'K', re.I))
    825. 

  825.         self.assertTrue(re.match(r'[\u2129-\u212b]', 'k', re.I))
    826. 

  826.         assert '\u017f'.upper() == 'S' # 'Ĺż'
    827. 

  827.         self.assertTrue(re.match(r'[R-T]', '\u017f', re.I))
    828. 

  828.         self.assertTrue(re.match(r'[r-t]', '\u017f', re.I))
    829. 

  829.         self.assertTrue(re.match(r'[\u017e-\u0180]', 'S', re.I))
    830. 

  830.         self.assertTrue(re.match(r'[\u017e-\u0180]', 's', re.I))
    831. 

  831.         assert '\ufb05'.upper() == '\ufb06'.upper() == 'ST' # 'ſt', 'st'
    832. 

  832.         self.assertTrue(re.match(r'[\ufb04-\ufb05]', '\ufb06', re.I))
    833. 

  833.         self.assertTrue(re.match(r'[\ufb06-\ufb07]', '\ufb05', re.I))
    834. 

  834. 

  835.     def test_category(self):
    836. 

  836.         self.assertEqual(re.match(r"(\s)", " ").group(1), " ")
    837. 

  837. 

  838.     def test_getlower(self):
    839. 

  839.         import _sre
    840. 

  840.         self.assertEqual(_sre.getlower(ord('A'), 0), ord('a'))
    841. 

  841.         self.assertEqual(_sre.getlower(ord('A'), re.LOCALE), ord('a'))
    842. 

  842.         self.assertEqual(_sre.getlower(ord('A'), re.UNICODE), ord('a'))
    843. 

  843.         self.assertEqual(_sre.getlower(ord('A'), re.ASCII), ord('a'))
    844. 

  844. 

  845.         self.assertEqual(re.match("abc", "ABC", re.I).group(0), "ABC")
    846. 

  846.         self.assertEqual(re.match(b"abc", b"ABC", re.I).group(0), b"ABC")
    847. 

  847.         self.assertEqual(re.match("abc", "ABC", re.I|re.A).group(0), "ABC")
    848. 

  848.         self.assertEqual(re.match(b"abc", b"ABC", re.I|re.L).group(0), b"ABC")
    849. 

  849. 

  850.     def test_not_literal(self):
    851. 

  851.         self.assertEqual(re.search("\s([^a])", " b").group(1), "b")
    852. 

  852.         self.assertEqual(re.search("\s([^a]*)", " bb").group(1), "bb")
    853. 

  853. 

  854.     def test_search_coverage(self):
    855. 

  855.         self.assertEqual(re.search("\s(b)", " b").group(1), "b")
    856. 

  856.         self.assertEqual(re.search("a\s", "a ").group(0), "a ")
    857. 

  857. 

  858.     def assertMatch(self, pattern, text, match=None, span=None,
    859. 

  859.                     matcher=re.match):
    860. 

  860.         if match is None and span is None:
    861. 

  861.             # the pattern matches the whole text
    862. 

  862.             match = text
    863. 

  863.             span = (0, len(text))
    864. 

  864.         elif match is None or span is None:
    865. 

  865.             raise ValueError('If match is not None, span should be specified '
    866. 

  866.                              '(and vice versa).')
    867. 

  867.         m = matcher(pattern, text)
    868. 

  868.         self.assertTrue(m)
    869. 

  869.         self.assertEqual(m.group(), match)
    870. 

  870.         self.assertEqual(m.span(), span)
    871. 

  871. 

  872.     def test_re_escape(self):
    873. 

  873.         alnum_chars = string.ascii_letters + string.digits + '_'
    874. 

  874.         p = ''.join(chr(i) for i in range(256))
    875. 

  875.         for c in p:
    876. 

  876.             if c in alnum_chars:
    877. 

  877.                 self.assertEqual(re.escape(c), c)
    878. 

  878.             elif c == '\x00':
    879. 

  879.                 self.assertEqual(re.escape(c), '\\000')
    880. 

  880.             else:
    881. 

  881.                 self.assertEqual(re.escape(c), '\\' + c)
    882. 

  882.             self.assertMatch(re.escape(c), c)
    883. 

  883.         self.assertMatch(re.escape(p), p)
    884. 

  884. 

  885.     def test_re_escape_byte(self):
    886. 

  886.         alnum_chars = (string.ascii_letters + string.digits + '_').encode('ascii')
    887. 

  887.         p = bytes(range(256))
    888. 

  888.         for i in p:
    889. 

  889.             b = bytes([i])
    890. 

  890.             if b in alnum_chars:
    891. 

  891.                 self.assertEqual(re.escape(b), b)
    892. 

  892.             elif i == 0:
    893. 

  893.                 self.assertEqual(re.escape(b), b'\\000')
    894. 

  894.             else:
    895. 

  895.                 self.assertEqual(re.escape(b), b'\\' + b)
    896. 

  896.             self.assertMatch(re.escape(b), b)
    897. 

  897.         self.assertMatch(re.escape(p), p)
    898. 

  898. 

  899.     def test_re_escape_non_ascii(self):
    900. 

  900.         s = 'xxx\u2620\u2620\u2620xxx'
    901. 

  901.         s_escaped = re.escape(s)
    902. 

  902.         self.assertEqual(s_escaped, 'xxx\\\u2620\\\u2620\\\u2620xxx')
    903. 

  903.         self.assertMatch(s_escaped, s)
    904. 

  904.         self.assertMatch('.%s+.' % re.escape('\u2620'), s,
    905. 

  905.                          'x\u2620\u2620\u2620x', (2, 7), re.search)
    906. 

  906. 

  907.     def test_re_escape_non_ascii_bytes(self):
    908. 

  908.         b = 'y\u2620y\u2620y'.encode('utf-8')
    909. 

  909.         b_escaped = re.escape(b)
    910. 

  910.         self.assertEqual(b_escaped, b'y\\\xe2\\\x98\\\xa0y\\\xe2\\\x98\\\xa0y')
    911. 

  911.         self.assertMatch(b_escaped, b)
    912. 

  912.         res = re.findall(re.escape('\u2620'.encode('utf-8')), b)
    913. 

  913.         self.assertEqual(len(res), 2)
    914. 

  914. 

  915.     def test_pickling(self):
    916. 

  916.         import pickle
    917. 

  917.         oldpat = re.compile('a(?:b|(c|e){1,2}?|d)+?(.)', re.UNICODE)
    918. 

  918.         for proto in range(pickle.HIGHEST_PROTOCOL + 1):
    919. 

  919.             pickled = pickle.dumps(oldpat, proto)
    920. 

  920.             newpat = pickle.loads(pickled)
    921. 

  921.             self.assertEqual(newpat, oldpat)
    922. 

  922.         # current pickle expects the _compile() reconstructor in re module
    923. 

  923.         from re import _compile
    924. 

  924. 

  925.     def test_constants(self):
    926. 

  926.         self.assertEqual(re.I, re.IGNORECASE)
    927. 

  927.         self.assertEqual(re.L, re.LOCALE)
    928. 

  928.         self.assertEqual(re.M, re.MULTILINE)
    929. 

  929.         self.assertEqual(re.S, re.DOTALL)
    930. 

  930.         self.assertEqual(re.X, re.VERBOSE)
    931. 

  931. 

  932.     def test_flags(self):
    933. 

  933.         for flag in [re.I, re.M, re.X, re.S, re.A, re.U]:
    934. 

  934.             self.assertTrue(re.compile('^pattern$', flag))
    935. 

  935.         for flag in [re.I, re.M, re.X, re.S, re.A, re.L]:
    936. 

  936.             self.assertTrue(re.compile(b'^pattern$', flag))
    937. 

  937. 

  938.     def test_sre_character_literals(self):
    939. 

  939.         for i in [0, 8, 16, 32, 64, 127, 128, 255, 256, 0xFFFF, 0x10000, 0x10FFFF]:
    940. 

  940.             if i < 256:
    941. 

  941.                 self.assertTrue(re.match(r"\%03o" % i, chr(i)))
    942. 

  942.                 self.assertTrue(re.match(r"\%03o0" % i, chr(i)+"0"))
    943. 

  943.                 self.assertTrue(re.match(r"\%03o8" % i, chr(i)+"8"))
    944. 

  944.                 self.assertTrue(re.match(r"\x%02x" % i, chr(i)))
    945. 

  945.                 self.assertTrue(re.match(r"\x%02x0" % i, chr(i)+"0"))
    946. 

  946.                 self.assertTrue(re.match(r"\x%02xz" % i, chr(i)+"z"))
    947. 

  947.             if i < 0x10000:
    948. 

  948.                 self.assertTrue(re.match(r"\u%04x" % i, chr(i)))
    949. 

  949.                 self.assertTrue(re.match(r"\u%04x0" % i, chr(i)+"0"))
    950. 

  950.                 self.assertTrue(re.match(r"\u%04xz" % i, chr(i)+"z"))
    951. 

  951.             self.assertTrue(re.match(r"\U%08x" % i, chr(i)))
    952. 

  952.             self.assertTrue(re.match(r"\U%08x0" % i, chr(i)+"0"))
    953. 

  953.             self.assertTrue(re.match(r"\U%08xz" % i, chr(i)+"z"))
    954. 

  954.         self.assertTrue(re.match(r"\0", "\000"))
    955. 

  955.         self.assertTrue(re.match(r"\08", "\0008"))
    956. 

  956.         self.assertTrue(re.match(r"\01", "\001"))
    957. 

  957.         self.assertTrue(re.match(r"\018", "\0018"))
    958. 

  958.         self.checkPatternError(r"\567",
    959. 

  959.                                r'octal escape value \567 outside of '
    960. 

  960.                                r'range 0-0o377', 0)
    961. 

  961.         self.checkPatternError(r"\911", 'invalid group reference', 0)
    962. 

  962.         self.checkPatternError(r"\x1", r'incomplete escape \x1', 0)
    963. 

  963.         self.checkPatternError(r"\x1z", r'incomplete escape \x1', 0)
    964. 

  964.         self.checkPatternError(r"\u123", r'incomplete escape \u123', 0)
    965. 

  965.         self.checkPatternError(r"\u123z", r'incomplete escape \u123', 0)
    966. 

  966.         self.checkPatternError(r"\U0001234", r'incomplete escape \U0001234', 0)
    967. 

  967.         self.checkPatternError(r"\U0001234z", r'incomplete escape \U0001234', 0)
    968. 

  968.         self.checkPatternError(r"\U00110000", r'bad escape \U00110000', 0)
    969. 

  969. 

  970.     def test_sre_character_class_literals(self):
    971. 

  971.         for i in [0, 8, 16, 32, 64, 127, 128, 255, 256, 0xFFFF, 0x10000, 0x10FFFF]:
    972. 

  972.             if i < 256:
    973. 

  973.                 self.assertTrue(re.match(r"[\%o]" % i, chr(i)))
    974. 

  974.                 self.assertTrue(re.match(r"[\%o8]" % i, chr(i)))
    975. 

  975.                 self.assertTrue(re.match(r"[\%03o]" % i, chr(i)))
    976. 

  976.                 self.assertTrue(re.match(r"[\%03o0]" % i, chr(i)…
    977.

Large files files are truncated, but you can click here to view the full file