Lib/test/string_tests.py

   1 """Common tests shared by test_string and test_userstring"""
   2
   3 import string
   4 from test.test_support import verify, vereq, verbose, TestFailed, have_unicode
   5
   6 transtable = '\000\001\002\003\004\005\006\007\010\011\012\013\014\015\016\017\020\021\022\023\024\025\026\027\030\031\032\033\034\035\036\037 !"#$%&\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`xyzdefghijklmnopqrstuvwxyz{|}~\177\200\201\202\203\204\205\206\207\210\211\212\213\214\215\216\217\220\221\222\223\224\225\226\227\230\231\232\233\234\235\236\237\240\241\242\243\244\245\246\247\250\251\252\253\254\255\256\257\260\261\262\263\264\265\266\267\270\271\272\273\274\275\276\277\300\301\302\303\304\305\306\307\310\311\312\313\314\315\316\317\320\321\322\323\324\325\326\327\330\331\332\333\334\335\336\337\340\341\342\343\344\345\346\347\350\351\352\353\354\355\356\357\360\361\362\363\364\365\366\367\370\371\372\373\374\375\376\377'
   7
   8 from UserList import UserList
   9
  10 class Sequence:
  11     def __init__(self): self.seq = 'wxyz'
  12     def __len__(self): return len(self.seq)
  13     def __getitem__(self, i): return self.seq[i]
  14
  15 class BadSeq1(Sequence):
  16     def __init__(self): self.seq = [7, 'hello', 123L]
  17
  18 class BadSeq2(Sequence):
  19     def __init__(self): self.seq = ['a', 'b', 'c']
  20     def __len__(self): return 8
  21
  22 def run_module_tests(test):
  23     """Run all tests that exercise a function in the string module"""
  24
  25     test('atoi', " 1 ", 1)
  26     test('atoi', " 1x", ValueError)
  27     test('atoi', " x1 ", ValueError)
  28     test('atol', "  1  ", 1L)
  29     test('atol', "  1x ", ValueError)
  30     test('atol', "  x1 ", ValueError)
  31     test('atof', "  1  ", 1.0)
  32     test('atof', "  1x ", ValueError)
  33     test('atof', "  x1 ", ValueError)
  34
  35     test('maketrans', 'abc', transtable, 'xyz')
  36     test('maketrans', 'abc', ValueError, 'xyzq')
  37
  38     # join now works with any sequence type
  39     test('join', ['a', 'b', 'c', 'd'], 'a b c d')
  40     test('join', ('a', 'b', 'c', 'd'), 'abcd', '')
  41     test('join', Sequence(), 'w x y z')
  42     test('join', 7, TypeError)
  43
  44     test('join', BadSeq1(), TypeError)
  45     test('join', BadSeq2(), 'a b c')
  46
  47     # try a few long ones
  48     print ":".join(['x' * 100] * 100)
  49     print ":".join(('x' * 100,) * 100)
  50
  51
  52 def run_method_tests(test):
  53     """Run all tests that exercise a method of a string object"""
  54
  55     test('capitalize', ' hello ', ' hello ')
  56     test('capitalize', 'hello ', 'Hello ')
  57     test('capitalize', 'aaaa', 'Aaaa')
  58     test('capitalize', 'AaAa', 'Aaaa')
  59
  60     test('count', 'aaa', 3, 'a')
  61     test('count', 'aaa', 0, 'b')
  62
  63     test('find', 'abcdefghiabc', 0, 'abc')
  64     test('find', 'abcdefghiabc', 9, 'abc', 1)
  65     test('find', 'abcdefghiabc', -1, 'def', 4)
  66     test('rfind', 'abcdefghiabc', 9, 'abc')
  67     test('lower', 'HeLLo', 'hello')
  68     test('lower', 'hello', 'hello')
  69     test('upper', 'HeLLo', 'HELLO')
  70     test('upper', 'HELLO', 'HELLO')
  71
  72     test('title', ' hello ', ' Hello ')
  73     test('title', 'hello ', 'Hello ')
  74     test('title', "fOrMaT thIs aS titLe String", 'Format This As Title String')
  75     test('title', "fOrMaT,thIs-aS*titLe;String", 'Format,This-As*Title;String')
  76     test('title', "getInt", 'Getint')
  77
  78     test('expandtabs', 'abc\rab\tdef\ng\thi', 'abc\rab      def\ng       hi')
  79     test('expandtabs', 'abc\rab\tdef\ng\thi', 'abc\rab      def\ng       hi', 8)
  80     test('expandtabs', 'abc\rab\tdef\ng\thi', 'abc\rab  def\ng   hi', 4)
  81     test('expandtabs', 'abc\r\nab\tdef\ng\thi', 'abc\r\nab  def\ng   hi', 4)
  82
  83     test('islower', 'a', 1)
  84     test('islower', 'A', 0)
  85     test('islower', '\n', 0)
  86     test('islower', 'abc', 1)
  87     test('islower', 'aBc', 0)
  88     test('islower', 'abc\n', 1)
  89
  90     test('isupper', 'a', 0)
  91     test('isupper', 'A', 1)
  92     test('isupper', '\n', 0)
  93     test('isupper', 'ABC', 1)
  94     test('isupper', 'AbC', 0)
  95     test('isupper', 'ABC\n', 1)
  96
  97     test('istitle', 'a', 0)
  98     test('istitle', 'A', 1)
  99     test('istitle', '\n', 0)
 100     test('istitle', 'A Titlecased Line', 1)
 101     test('istitle', 'A\nTitlecased Line', 1)
 102     test('istitle', 'A Titlecased, Line', 1)
 103     test('istitle', 'Not a capitalized String', 0)
 104     test('istitle', 'Not\ta Titlecase String', 0)
 105     test('istitle', 'Not--a Titlecase String', 0)
 106
 107     test('isalpha', 'a', 1)
 108     test('isalpha', 'A', 1)
 109     test('isalpha', '\n', 0)
 110     test('isalpha', 'abc', 1)
 111     test('isalpha', 'aBc123', 0)
 112     test('isalpha', 'abc\n', 0)
 113
 114     test('isalnum', 'a', 1)
 115     test('isalnum', 'A', 1)
 116     test('isalnum', '\n', 0)
 117     test('isalnum', '123abc456', 1)
 118     test('isalnum', 'a1b3c', 1)
 119     test('isalnum', 'aBc000 ', 0)
 120     test('isalnum', 'abc\n', 0)
 121
 122     # join now works with any sequence type
 123     test('join', ' ', 'a b c d', ['a', 'b', 'c', 'd'])
 124     test('join', '', 'abcd', ('a', 'b', 'c', 'd'))
 125     test('join', ' ', 'w x y z', Sequence())
 126     test('join', 'a', 'abc', ('abc',))
 127     test('join', 'a', 'z', UserList(['z']))
 128     if have_unicode:
 129         test('join', unicode('.'), unicode('a.b.c'), ['a', 'b', 'c'])
 130         test('join', '.', unicode('a.b.c'), [unicode('a'), 'b', 'c'])
 131         test('join', '.', unicode('a.b.c'), ['a', unicode('b'), 'c'])
 132         test('join', '.', unicode('a.b.c'), ['a', 'b', unicode('c')])
 133         test('join', '.', TypeError, ['a', unicode('b'), 3])
 134     for i in [5, 25, 125]:
 135         test('join', '-', ((('a' * i) + '-') * i)[:-1],
 136              ['a' * i] * i)
 137
 138     test('join', ' ', TypeError, BadSeq1())
 139     test('join', ' ', 'a b c', BadSeq2())
 140
 141     test('splitlines', "abc\ndef\n\rghi", ['abc', 'def', '', 'ghi'])
 142     test('splitlines', "abc\ndef\n\r\nghi", ['abc', 'def', '', 'ghi'])
 143     test('splitlines', "abc\ndef\r\nghi", ['abc', 'def', 'ghi'])
 144     test('splitlines', "abc\ndef\r\nghi\n", ['abc', 'def', 'ghi'])
 145     test('splitlines', "abc\ndef\r\nghi\n\r", ['abc', 'def', 'ghi', ''])
 146     test('splitlines', "\nabc\ndef\r\nghi\n\r", ['', 'abc', 'def', 'ghi', ''])
 147     test('splitlines', "\nabc\ndef\r\nghi\n\r", ['\n', 'abc\n', 'def\r\n', 'ghi\n', '\r'], 1)
 148
 149     test('split', 'this is the split function',
 150          ['this', 'is', 'the', 'split', 'function'])
 151     test('split', 'a|b|c|d', ['a', 'b', 'c', 'd'], '|')
 152     test('split', 'a|b|c|d', ['a', 'b', 'c|d'], '|', 2)
 153     test('split', 'a b c d', ['a', 'b c d'], None, 1)
 154     test('split', 'a b c d', ['a', 'b', 'c d'], None, 2)
 155     test('split', 'a b c d', ['a', 'b', 'c', 'd'], None, 3)
 156     test('split', 'a b c d', ['a', 'b', 'c', 'd'], None, 4)
 157     test('split', 'a b c d', ['a b c d'], None, 0)
 158     test('split', 'a  b  c  d', ['a', 'b', 'c  d'], None, 2)
 159     test('split', 'a b c d ', ['a', 'b', 'c', 'd'])
 160
 161     test('strip', '   hello   ', 'hello')
 162     test('lstrip', '   hello   ', 'hello   ')
 163     test('rstrip', '   hello   ', '   hello')
 164     test('strip', 'hello', 'hello')
 165
 166     # strip/lstrip/rstrip with None arg
 167     test('strip', '   hello   ', 'hello', None)
 168     test('lstrip', '   hello   ', 'hello   ', None)
 169     test('rstrip', '   hello   ', '   hello', None)
 170     test('strip', 'hello', 'hello', None)
 171
 172     # strip/lstrip/rstrip with str arg
 173     test('strip', 'xyzzyhelloxyzzy', 'hello', 'xyz')
 174     test('lstrip', 'xyzzyhelloxyzzy', 'helloxyzzy', 'xyz')
 175     test('rstrip', 'xyzzyhelloxyzzy', 'xyzzyhello', 'xyz')
 176     test('strip', 'hello', 'hello', 'xyz')
 177
 178     # strip/lstrip/rstrip with unicode arg
 179     if have_unicode:
 180         test('strip', 'xyzzyhelloxyzzy',
 181              unicode('hello', 'ascii'), unicode('xyz', 'ascii'))
 182         test('lstrip', 'xyzzyhelloxyzzy',
 183              unicode('helloxyzzy', 'ascii'), unicode('xyz', 'ascii'))
 184         test('rstrip', 'xyzzyhelloxyzzy',
 185              unicode('xyzzyhello', 'ascii'), unicode('xyz', 'ascii'))
 186         test('strip', 'hello',
 187              unicode('hello', 'ascii'), unicode('xyz', 'ascii'))
 188
 189     test('swapcase', 'HeLLo cOmpUteRs', 'hEllO CoMPuTErS')
 190     test('translate', 'xyzabcdef', 'xyzxyz', transtable, 'def')
 191
 192     table = string.maketrans('a', 'A')
 193     test('translate', 'abc', 'Abc', table)
 194     test('translate', 'xyz', 'xyz', table)
 195
 196     test('replace', 'one!two!three!', 'one@two!three!', '!', '@', 1)
 197     test('replace', 'one!two!three!', 'onetwothree', '!', '')
 198     test('replace', 'one!two!three!', 'one@two@three!', '!', '@', 2)
 199     test('replace', 'one!two!three!', 'one@two@three@', '!', '@', 3)
 200     test('replace', 'one!two!three!', 'one@two@three@', '!', '@', 4)
 201     test('replace', 'one!two!three!', 'one!two!three!', '!', '@', 0)
 202     test('replace', 'one!two!three!', 'one@two@three@', '!', '@')
 203     test('replace', 'one!two!three!', 'one!two!three!', 'x', '@')
 204     test('replace', 'one!two!three!', 'one!two!three!', 'x', '@', 2)
 205     test('replace', 'abc', '-a-b-c-', '', '-')
 206     test('replace', 'abc', '-a-b-c', '', '-', 3)
 207     test('replace', 'abc', 'abc', '', '-', 0)
 208     test('replace', '', '', '', '')
 209     # Next three for SF bug 422088: [OSF1 alpha] string.replace(); died with
 210     # MemoryError due to empty result (platform malloc issue when requesting
 211     # 0 bytes).
 212     test('replace', '123', '', '123', '')
 213     test('replace', '123123', '', '123', '')
 214     test('replace', '123x123', 'x', '123', '')
 215
 216     test('startswith', 'hello', 1, 'he')
 217     test('startswith', 'hello', 1, 'hello')
 218     test('startswith', 'hello', 0, 'hello world')
 219     test('startswith', 'hello', 1, '')
 220     test('startswith', 'hello', 0, 'ello')
 221     test('startswith', 'hello', 1, 'ello', 1)
 222     test('startswith', 'hello', 1, 'o', 4)
 223     test('startswith', 'hello', 0, 'o', 5)
 224     test('startswith', 'hello', 1, '', 5)
 225     test('startswith', 'hello', 0, 'lo', 6)
 226     test('startswith', 'helloworld', 1, 'lowo', 3)
 227     test('startswith', 'helloworld', 1, 'lowo', 3, 7)
 228     test('startswith', 'helloworld', 0, 'lowo', 3, 6)
 229
 230     # test negative indices in startswith
 231     test('startswith', 'hello', 1, 'he', 0, -1)
 232     test('startswith', 'hello', 1, 'he', -53, -1)
 233     test('startswith', 'hello', 0, 'hello', 0, -1)
 234     test('startswith', 'hello', 0, 'hello world', -1, -10)
 235     test('startswith', 'hello', 0, 'ello', -5)
 236     test('startswith', 'hello', 1, 'ello', -4)
 237     test('startswith', 'hello', 0, 'o', -2)
 238     test('startswith', 'hello', 1, 'o', -1)
 239     test('startswith', 'hello', 1, '', -3, -3)
 240     test('startswith', 'hello', 0, 'lo', -9)
 241
 242     test('endswith', 'hello', 1, 'lo')
 243     test('endswith', 'hello', 0, 'he')
 244     test('endswith', 'hello', 1, '')
 245     test('endswith', 'hello', 0, 'hello world')
 246     test('endswith', 'helloworld', 0, 'worl')
 247     test('endswith', 'helloworld', 1, 'worl', 3, 9)
 248     test('endswith', 'helloworld', 1, 'world', 3, 12)
 249     test('endswith', 'helloworld', 1, 'lowo', 1, 7)
 250     test('endswith', 'helloworld', 1, 'lowo', 2, 7)
 251     test('endswith', 'helloworld', 1, 'lowo', 3, 7)
 252     test('endswith', 'helloworld', 0, 'lowo', 4, 7)
 253     test('endswith', 'helloworld', 0, 'lowo', 3, 8)
 254     test('endswith', 'ab', 0, 'ab', 0, 1)
 255     test('endswith', 'ab', 0, 'ab', 0, 0)
 256
 257     # test negative indices in endswith
 258     test('endswith', 'hello', 1, 'lo', -2)
 259     test('endswith', 'hello', 0, 'he', -2)
 260     test('endswith', 'hello', 1, '', -3, -3)
 261     test('endswith', 'hello', 0, 'hello world', -10, -2)
 262     test('endswith', 'helloworld', 0, 'worl', -6)
 263     test('endswith', 'helloworld', 1, 'worl', -5, -1)
 264     test('endswith', 'helloworld', 1, 'worl', -5, 9)
 265     test('endswith', 'helloworld', 1, 'world', -7, 12)
 266     test('endswith', 'helloworld', 1, 'lowo', -99, -3)
 267     test('endswith', 'helloworld', 1, 'lowo', -8, -3)
 268     test('endswith', 'helloworld', 1, 'lowo', -7, -3)
 269     test('endswith', 'helloworld', 0, 'lowo', 3, -4)
 270     test('endswith', 'helloworld', 0, 'lowo', -8, -2)
 271
 272     test('zfill', '123', '123', 2)
 273     test('zfill', '123', '123', 3)
 274     test('zfill', '123', '0123', 4)
 275     test('zfill', '+123', '+123', 3)
 276     test('zfill', '+123', '+123', 4)
 277     test('zfill', '+123', '+0123', 5)
 278     test('zfill', '-123', '-123', 3)
 279     test('zfill', '-123', '-123', 4)
 280     test('zfill', '-123', '-0123', 5)
 281     test('zfill', '', '000', 3)
 282     test('zfill', '34', '34', 1)
 283     test('zfill', '34', '0034', 4)
 284
 285     # Encoding/decoding
 286     codecs = [('rot13', 'uryyb jbeyq'),
 287               ('base64', 'aGVsbG8gd29ybGQ=\n'),
 288               ('hex', '68656c6c6f20776f726c64'),
 289               ('uu', 'begin 666 <data>\n+:&5L;&\\@=V]R;&0 \n \nend\n')]
 290     for encoding, data in codecs:
 291         test('encode', 'hello world', data, encoding)
 292         test('decode', data, 'hello world', encoding)
 293     # zlib is optional, so we make the test optional too...
 294     try:
 295         import zlib
 296     except ImportError:
 297         pass
 298     else:
 299         data = 'x\x9c\xcbH\xcd\xc9\xc9W(\xcf/\xcaI\x01\x00\x1a\x0b\x04]'
 300         verify('hello world'.encode('zlib') == data)
 301         verify(data.decode('zlib') == 'hello world')
 302
 303 def test_exception(lhs, rhs, msg):
 304     try:
 305         lhs in rhs
 306     except TypeError:
 307         pass
 308     else:
 309         raise TestFailed, msg
 310
 311 def run_contains_tests(test):
 312     test('__contains__', '', True, '')         # vereq('' in '', True)
 313     test('__contains__', 'abc', True, '')      # vereq('' in 'abc', True)
 314     test('__contains__', 'abc', False, '\0')   # vereq('\0' in 'abc', False)
 315     test('__contains__', '\0abc', True, '\0')  # vereq('\0' in '\0abc', True)
 316     test('__contains__', 'abc\0', True, '\0')  # vereq('\0' in 'abc\0', True)
 317     test('__contains__', '\0abc', True, 'a')   # vereq('a' in '\0abc', True)
 318     test('__contains__', 'asdf', True, 'asdf') # vereq('asdf' in 'asdf', True)
 319     test('__contains__', 'asd', False, 'asdf') # vereq('asdf' in 'asd', False)
 320     test('__contains__', '', False, 'asdf')    # vereq('asdf' in '', False)
 321
 322 def run_inplace_tests(constructor):
 323     # Verify clearing of SF bug #592573
 324     s = t = constructor('abc')
 325     s += constructor('def')
 326     verify(s != t, 'in-place concatenate should create a new object')