summaryrefslogtreecommitdiffstats
path: root/Lib/test/test_bigmem.py
blob: d4fc6eb478510c495cc21bf181f6eaf717bb83f6 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
from test import test_support
from test.test_support import bigmemtest, _1G, _2G

import unittest
import operator
import string
import sys

# Bigmem testing houserules:
#
#  - Try not to allocate too many large objects. It's okay to rely on
#    refcounting semantics, but don't forget that 's = create_largestring()'
#    doesn't release the old 's' (if it exists) until well after its new
#    value has been created. Use 'del s' before the create_largestring call.
#
#  - Do *not* compare large objects using assertEquals or similar. It's a
#    lengty operation and the errormessage will be utterly useless due to
#    its size. To make sure whether a result has the right contents, better
#    to use the strip or count methods, or compare meaningful slices.
#
#  - Don't forget to test for large indices, offsets and results and such,
#    in addition to large sizes.
#
#  - When repeating an object (say, a substring, or a small list) to create
#    a large object, make the subobject of a length that is not a power of
#    2. That way, int-wrapping problems are more easily detected.
#
#  - While the bigmemtest decorator speaks of 'minsize', all tests will
#    actually be called with a much smaller number too, in the normal
#    test run (5Kb currently.) This is so the tests themselves get frequent
#    testing. Consequently, always make all large allocations based on the
#    passed-in 'size', and don't rely on the size being very large. Also,
#    memuse-per-size should remain sane (less than a few thousand); if your
#    test uses more, adjust 'size' upward, instead.

class StrTest(unittest.TestCase):
    @bigmemtest(minsize=_2G, memuse=2)
    def test_capitalize(self, size):
        SUBSTR = ' abc def ghi'
        s = '-' * size + SUBSTR
        caps = s.capitalize()
        self.assertEquals(caps[-len(SUBSTR):],
                         SUBSTR.capitalize())
        self.assertEquals(caps.lstrip('-'), SUBSTR)

    @bigmemtest(minsize=_2G + 10, memuse=1)
    def test_center(self, size):
        SUBSTR = ' abc def ghi'
        s = SUBSTR.center(size)
        self.assertEquals(len(s), size)
        lpadsize = rpadsize = (len(s) - len(SUBSTR)) // 2
        if len(s) % 2:
            lpadsize += 1
        self.assertEquals(s[lpadsize:-rpadsize], SUBSTR)
        self.assertEquals(s.strip(), SUBSTR.strip())

    @bigmemtest(minsize=_2G, memuse=2)
    def test_count(self, size):
        SUBSTR = ' abc def ghi'
        s = '.' * size + SUBSTR
        self.assertEquals(s.count('.'), size)
        s += '.'
        self.assertEquals(s.count('.'), size + 1)
        self.assertEquals(s.count(' '), 3)
        self.assertEquals(s.count('i'), 1)
        self.assertEquals(s.count('j'), 0)

    @bigmemtest(minsize=_2G + 2, memuse=3)
    def test_decode(self, size):
        s = '.' * size
        self.assertEquals(len(s.decode('utf-8')), size)

    @bigmemtest(minsize=_2G + 2, memuse=3)
    def test_encode(self, size):
        s = u'.' * size
        self.assertEquals(len(s.encode('utf-8')), size)

    @bigmemtest(minsize=_2G, memuse=2)
    def test_endswith(self, size):
        SUBSTR = ' abc def ghi'
        s = '-' * size + SUBSTR
        self.failUnless(s.endswith(SUBSTR))
        self.failUnless(s.endswith(s))
        s2 = '...' + s
        self.failUnless(s2.endswith(s))
        self.failIf(s.endswith('a' + SUBSTR))
        self.failIf(SUBSTR.endswith(s))

    @bigmemtest(minsize=_2G + 10, memuse=2)
    def test_expandtabs(self, size):
        s = '-' * size
        tabsize = 8
        self.assertEquals(s.expandtabs(), s)
        del s
        slen, remainder = divmod(size, tabsize)
        s = '       \t' * slen
        s = s.expandtabs(tabsize)
        self.assertEquals(len(s), size - remainder)
        self.assertEquals(len(s.strip(' ')), 0)

    @bigmemtest(minsize=_2G, memuse=2)
    def test_find(self, size):
        SUBSTR = ' abc def ghi'
        sublen = len(SUBSTR)
        s = ''.join([SUBSTR, '-' * size, SUBSTR])
        self.assertEquals(s.find(' '), 0)
        self.assertEquals(s.find(SUBSTR), 0)
        self.assertEquals(s.find(' ', sublen), sublen + size)
        self.assertEquals(s.find(SUBSTR, len(SUBSTR)), sublen + size)
        self.assertEquals(s.find('i'), SUBSTR.find('i'))
        self.assertEquals(s.find('i', sublen),
                         sublen + size + SUBSTR.find('i'))
        self.assertEquals(s.find('i', size),
                         sublen + size + SUBSTR.find('i'))
        self.assertEquals(s.find('j'), -1)

    @bigmemtest(minsize=_2G, memuse=2)
    def test_index(self, size):
        SUBSTR = ' abc def ghi'
        sublen = len(SUBSTR)
        s = ''.join([SUBSTR, '-' * size, SUBSTR])
        self.assertEquals(s.index(' '), 0)
        self.assertEquals(s.index(SUBSTR), 0)
        self.assertEquals(s.index(' ', sublen), sublen + size)
        self.assertEquals(s.index(SUBSTR, sublen), sublen + size)
        self.assertEquals(s.index('i'), SUBSTR.index('i'))
        self.assertEquals(s.index('i', sublen),
                         sublen + size + SUBSTR.index('i'))
        self.assertEquals(s.index('i', size),
                         sublen + size + SUBSTR.index('i'))
        self.assertRaises(ValueError, s.index, 'j')

    @bigmemtest(minsize=_2G, memuse=2)
    def test_isalnum(self, size):
        SUBSTR = '123456'
        s = 'a' * size + SUBSTR
        self.failUnless(s.isalnum())
        s += '.'
        self.failIf(s.isalnum())

    @bigmemtest(minsize=_2G, memuse=2)
    def test_isalpha(self, size):
        SUBSTR = 'zzzzzzz'
        s = 'a' * size + SUBSTR
        self.failUnless(s.isalpha())
        s += '.'
        self.failIf(s.isalpha())

    @bigmemtest(minsize=_2G, memuse=2)
    def test_isdigit(self, size):
        SUBSTR = '123456'
        s = '9' * size + SUBSTR
        self.failUnless(s.isdigit())
        s += 'z'
        self.failIf(s.isdigit())

    @bigmemtest(minsize=_2G, memuse=2)
    def test_islower(self, size):
        chars = ''.join([ chr(c) for c in range(255) if not chr(c).isupper() ])
        repeats = size // len(chars) + 2
        s = chars * repeats
        self.failUnless(s.islower())
        s += 'A'
        self.failIf(s.islower())

    @bigmemtest(minsize=_2G, memuse=2)
    def test_isspace(self, size):
        whitespace = ' \f\n\r\t\v'
        repeats = size // len(whitespace) + 2
        s = whitespace * repeats
        self.failUnless(s.isspace())
        s += 'j'
        self.failIf(s.isspace())

    @bigmemtest(minsize=_2G, memuse=2)
    def test_istitle(self, size):
        SUBSTR = '123456'
        s = ''.join(['A', 'a' * size, SUBSTR])
        self.failUnless(s.istitle())
        s += 'A'
        self.failUnless(s.istitle())
        s += 'aA'
        self.failIf(s.istitle())

    @bigmemtest(minsize=_2G, memuse=2)
    def test_isupper(self, size):
        chars = ''.join([ chr(c) for c in range(255) if not chr(c).islower() ])
        repeats = size // len(chars) + 2
        s = chars * repeats
        self.failUnless(s.isupper())
        s += 'a'
        self.failIf(s.isupper())

    @bigmemtest(minsize=_2G, memuse=2)
    def test_join(self, size):
        s = 'A' * size
        x = s.join(['aaaaa', 'bbbbb'])
        self.assertEquals(x.count('a'), 5)
        self.assertEquals(x.count('b'), 5)
        self.failUnless(x.startswith('aaaaaA'))
        self.failUnless(x.endswith('Abbbbb'))

    @bigmemtest(minsize=_2G + 10, memuse=1)
    def test_ljust(self, size):
        SUBSTR = ' abc def ghi'
        s = SUBSTR.ljust(size)
        self.failUnless(s.startswith(SUBSTR + '  '))
        self.assertEquals(len(s), size)
        self.assertEquals(s.strip(), SUBSTR.strip())

    @bigmemtest(minsize=_2G + 10, memuse=2)
    def test_lower(self, size):
        s = 'A' * size
        s = s.lower()
        self.assertEquals(len(s), size)
        self.assertEquals(s.count('a'), size)

    @bigmemtest(minsize=_2G + 10, memuse=1)
    def test_lstrip(self, size):
        SUBSTR = 'abc def ghi'
        s = SUBSTR.rjust(size)
        self.assertEquals(len(s), size)
        self.assertEquals(s.lstrip(), SUBSTR.lstrip())
        del s
        s = SUBSTR.ljust(size)
        self.assertEquals(len(s), size)
        stripped = s.lstrip()
        self.failUnless(stripped is s)

    @bigmemtest(minsize=_2G + 10, memuse=2)
    def test_replace(self, size):
        replacement = 'a'
        s = ' ' * size
        s = s.replace(' ', replacement)
        self.assertEquals(len(s), size)
        self.assertEquals(s.count(replacement), size)
        s = s.replace(replacement, ' ', size - 4)
        self.assertEquals(len(s), size)
        self.assertEquals(s.count(replacement), 4)
        self.assertEquals(s[-10:], '      aaaa')

    @bigmemtest(minsize=_2G, memuse=2)
    def test_rfind(self, size):
        SUBSTR = ' abc def ghi'
        sublen = len(SUBSTR)
        s = ''.join([SUBSTR, '-' * size, SUBSTR])
        self.assertEquals(s.rfind(' '), sublen + size + SUBSTR.rfind(' '))
        self.assertEquals(s.rfind(SUBSTR), sublen + size)
        self.assertEquals(s.rfind(' ', 0, size), SUBSTR.rfind(' '))
        self.assertEquals(s.rfind(SUBSTR, 0, sublen + size), 0)
        self.assertEquals(s.rfind('i'), sublen + size + SUBSTR.rfind('i'))
        self.assertEquals(s.rfind('i', 0, sublen), SUBSTR.rfind('i'))
        self.assertEquals(s.rfind('i', 0, sublen + size),
                          SUBSTR.rfind('i'))
        self.assertEquals(s.rfind('j'), -1)

    @bigmemtest(minsize=_2G, memuse=2)
    def test_rindex(self, size):
        SUBSTR = ' abc def ghi'
        sublen = len(SUBSTR)
        s = ''.join([SUBSTR, '-' * size, SUBSTR])
        self.assertEquals(s.rindex(' '),
                          sublen + size + SUBSTR.rindex(' '))
        self.assertEquals(s.rindex(SUBSTR), sublen + size)
        self.assertEquals(s.rindex(' ', 0, sublen + size - 1),
                          SUBSTR.rindex(' '))
        self.assertEquals(s.rindex(SUBSTR, 0, sublen + size), 0)
        self.assertEquals(s.rindex('i'),
                          sublen + size + SUBSTR.rindex('i'))
        self.assertEquals(s.rindex('i', 0, sublen), SUBSTR.rindex('i'))
        self.assertEquals(s.rindex('i', 0, sublen + size),
                          SUBSTR.rindex('i'))
        self.assertRaises(ValueError, s.rindex, 'j')

    @bigmemtest(minsize=_2G + 10, memuse=1)
    def test_rjust(self, size):
        SUBSTR = ' abc def ghi'
        s = SUBSTR.ljust(size)
        self.failUnless(s.startswith(SUBSTR + '  '))
        self.assertEquals(len(s), size)
        self.assertEquals(s.strip(), SUBSTR.strip())

    @bigmemtest(minsize=_2G + 10, memuse=1)
    def test_rstrip(self, size):
        SUBSTR = ' abc def ghi'
        s = SUBSTR.ljust(size)
        self.assertEquals(len(s), size)
        self.assertEquals(s.rstrip(), SUBSTR.rstrip())
        del s
        s = SUBSTR.rjust(size)
        self.assertEquals(len(s), size)
        stripped = s.rstrip()
        self.failUnless(stripped is s)

    # The test takes about size bytes to build a string, and then about
    # sqrt(size) substrings of sqrt(size) in size and a list to
    # hold sqrt(size) items. It's close but just over 2x size.
    @bigmemtest(minsize=_2G, memuse=2.1)
    def test_split_small(self, size):
        # Crudely calculate an estimate so that the result of s.split won't
        # take up an inordinate amount of memory
        chunksize = int(size ** 0.5 + 2)
        SUBSTR = 'a' + ' ' * chunksize
        s = SUBSTR * chunksize
        l = s.split()
        self.assertEquals(len(l), chunksize)
        self.assertEquals(set(l), set(['a']))
        del l
        l = s.split('a')
        self.assertEquals(len(l), chunksize + 1)
        self.assertEquals(set(l), set(['', ' ' * chunksize]))

    # Allocates a string of twice size (and briefly two) and a list of
    # size.  Because of internal affairs, the s.split() call produces a
    # list of size times the same one-character string, so we only
    # suffer for the list size. (Otherwise, it'd cost another 48 times
    # size in bytes!) Nevertheless, a list of size takes
    # 8*size bytes.
    @bigmemtest(minsize=_2G + 5, memuse=10)
    def test_split_large(self, size):
        s = ' a' * size + ' '
        l = s.split()
        self.assertEquals(len(l), size)
        self.assertEquals(set(l), set(['a']))
        del l
        l = s.split('a')
        self.assertEquals(len(l), size + 1)
        self.assertEquals(set(l), set([' ']))

    @bigmemtest(minsize=_2G, memuse=2.1)
    def test_splitlines(self, size):
        # Crudely calculate an estimate so that the result of s.split won't
        # take up an inordinate amount of memory
        chunksize = int(size ** 0.5 + 2) // 2
        SUBSTR = ' ' * chunksize + '\n' + ' ' * chunksize + '\r\n'
        s = SUBSTR * chunksize
        l = s.splitlines()
        self.assertEquals(len(l), chunksize * 2)
        self.assertEquals(set(l), set([' ' * chunksize]))

    @bigmemtest(minsize=_2G, memuse=2)
    def test_startswith(self, size):
        SUBSTR = ' abc def ghi'
        s = '-' * size + SUBSTR
        self.failUnless(s.startswith(s))
        self.failUnless(s.startswith('-' * size))
        self.failIf(s.startswith(SUBSTR))

    @bigmemtest(minsize=_2G, memuse=1)
    def test_strip(self, size):
        SUBSTR = '   abc def ghi   '
        s = SUBSTR.rjust(size)
        self.assertEquals(len(s), size)
        self.assertEquals(s.strip(), SUBSTR.strip())
        del s
        s = SUBSTR.ljust(size)
        self.assertEquals(len(s), size)
        self.assertEquals(s.strip(), SUBSTR.strip())

    @bigmemtest(minsize=_2G, memuse=2)
    def test_swapcase(self, size):
        SUBSTR = "aBcDeFG12.'\xa9\x00"
        sublen = len(SUBSTR)
        repeats = size // sublen + 2
        s = SUBSTR * repeats
        s = s.swapcase()
        self.assertEquals(len(s), sublen * repeats)
        self.assertEquals(s[:sublen * 3], SUBSTR.swapcase() * 3)
        self.assertEquals(s[-sublen * 3:], SUBSTR.swapcase() * 3)

    @bigmemtest(minsize=_2G, memuse=2)
    def test_title(self, size):
        SUBSTR = 'SpaaHAaaAaham'
        s = SUBSTR * (size // len(SUBSTR) + 2)
        s = s.title()
        self.failUnless(s.startswith((SUBSTR * 3).title()))
        self.failUnless(s.endswith(SUBSTR.lower() * 3))

    @bigmemtest(minsize=_2G, memuse=2)
    def test_translate(self, size):
        trans = string.maketrans('.aZ', '-!$')
        SUBSTR = 'aZz.z.Aaz.'
        sublen = len(SUBSTR)
        repeats = size // sublen + 2
        s = SUBSTR * repeats
        s = s.translate(trans)
        self.assertEquals(len(s), repeats * sublen)
        self.assertEquals(s[:sublen], SUBSTR.translate(trans))
        self.assertEquals(s[-sublen:], SUBSTR.translate(trans))
        self.assertEquals(s.count('.'), 0)
        self.assertEquals(s.count('!'), repeats * 2)
        self.assertEquals(s.count('z'), repeats * 3)

    @bigmemtest(minsize=_2G + 5, memuse=2)
    def test_upper(self, size):
        s = 'a' * size
        s = s.upper()
        self.assertEquals(len(s), size)
        self.assertEquals(s.count('A'), size)

    @bigmemtest(minsize=_2G + 20, memuse=1)
    def test_zfill(self, size):
        SUBSTR = '-568324723598234'
        s = SUBSTR.zfill(size)
        self.failUnless(s.endswith('0' + SUBSTR[1:]))
        self.failUnless(s.startswith('-0'))
        self.assertEquals(len(s), size)
        self.assertEquals(s.count('0'), size - len(SUBSTR))

    @bigmemtest(minsize=_2G + 10, memuse=2)
    def test_format(self, size):
        s = '-' * size
        sf = '%s' % (s,)
        self.failUnless(s == sf)
        del sf
        sf = '..%s..' % (s,)
        self.assertEquals(len(sf), len(s) + 4)
        self.failUnless(sf.startswith('..-'))
        self.failUnless(sf.endswith('-..'))
        del s, sf

        size //= 2
        edge = '-' * size
        s = ''.join([edge, '%s', edge])
        del edge
        s = s % '...'
        self.assertEquals(len(s), size * 2 + 3)
        self.assertEquals(s.count('.'), 3)
        self.assertEquals(s.count('-'), size * 2)

    @bigmemtest(minsize=_2G + 10, memuse=2)
    def test_repr_small(self, size):
        s = '-' * size
        s = repr(s)
        self.assertEquals(len(s), size + 2)
        self.assertEquals(s[0], "'")
        self.assertEquals(s[-1], "'")
        self.assertEquals(s.count('-'), size)
        del s
        # repr() will create a string four times as large as this 'binary
        # string', but we don't want to allocate much more than twice
        # size in total.  (We do extra testing in test_repr_large())
        size = size // 5 * 2
        s = '\x00' * size
        s = repr(s)
        self.assertEquals(len(s), size * 4 + 2)
        self.assertEquals(s[0], "'")
        self.assertEquals(s[-1], "'")
        self.assertEquals(s.count('\\'), size)
        self.assertEquals(s.count('0'), size * 2)

    @bigmemtest(minsize=_2G + 10, memuse=5)
    def test_repr_large(self, size):
        s = '\x00' * size
        s = repr(s)
        self.assertEquals(len(s), size * 4 + 2)
        self.assertEquals(s[0], "'")
        self.assertEquals(s[-1], "'")
        self.assertEquals(s.count('\\'), size)
        self.assertEquals(s.count('0'), size * 2)

    # This test is meaningful even with size < 2G, as long as the
    # doubled string is > 2G (but it tests more if both are > 2G :)
    @bigmemtest(minsize=_1G + 2, memuse=3)
    def test_concat(self, size):
        s = '.' * size
        self.assertEquals(len(s), size)
        s = s + s
        self.assertEquals(len(s), size * 2)
        self.assertEquals(s.count('.'), size * 2)

    # This test is meaningful even with size < 2G, as long as the
    # repeated string is > 2G (but it tests more if both are > 2G :)
    @bigmemtest(minsize=_1G + 2, memuse=3)
    def test_repeat(self, size):
        s = '.' * size
        self.assertEquals(len(s), size)
        s = s * 2
        self.assertEquals(len(s), size * 2)
        self.assertEquals(s.count('.'), size * 2)

    @bigmemtest(minsize=_2G + 20, memuse=1)
    def test_slice_and_getitem(self, size):
        SUBSTR = '0123456789'
        sublen = len(SUBSTR)
        s = SUBSTR * (size // sublen)
        stepsize = len(s) // 100
        stepsize = stepsize - (stepsize % sublen)
        for i in range(0, len(s) - stepsize, stepsize):
            self.assertEquals(s[i], SUBSTR[0])
            self.assertEquals(s[i:i + sublen], SUBSTR)
            self.assertEquals(s[i:i + sublen:2], SUBSTR[::2])
            if i > 0:
                self.assertEquals(s[i + sublen - 1:i - 1:-3],
                                  SUBSTR[sublen::-3])
        # Make sure we do some slicing and indexing near the end of the
        # string, too.
        self.assertEquals(s[len(s) - 1], SUBSTR[-1])
        self.assertEquals(s[-1], SUBSTR[-1])
        self.assertEquals(s[len(s) - 10], SUBSTR[0])
        self.assertEquals(s[-sublen], SUBSTR[0])
        self.assertEquals(s[len(s):], '')
        self.assertEquals(s[len(s) - 1:], SUBSTR[-1])
        self.assertEquals(s[-1:], SUBSTR[-1])
        self.assertEquals(s[len(s) - sublen:], SUBSTR)
        self.assertEquals(s[-sublen:], SUBSTR)
        self.assertEquals(len(s[:]), len(s))
        self.assertEquals(len(s[:len(s) - 5]), len(s) - 5)
        self.assertEquals(len(s[5:-5]), len(s) - 10)

        self.assertRaises(IndexError, operator.getitem, s, len(s))
        self.assertRaises(IndexError, operator.getitem, s, len(s) + 1)
        self.assertRaises(IndexError, operator.getitem, s, len(s) + 1<<31)

    @bigmemtest(minsize=_2G, memuse=2)
    def test_contains(self, size):
        SUBSTR = '0123456789'
        edge = '-' * (size // 2)
        s = ''.join([edge, SUBSTR, edge])
        del edge
        self.failUnless(SUBSTR in s)
        self.failIf(SUBSTR * 2 in s)
        self.failUnless('-' in s)
        self.failIf('a' in s)
        s += 'a'
        self.failUnless('a' in s)

    @bigmemtest(minsize=_2G + 10, memuse=2)
    def test_compare(self, size):
        s1 = '-' * size
        s2 = '-' * size
        self.failUnless(s1 == s2)
        del s2
        s2 = s1 + 'a'
        self.failIf(s1 == s2)
        del s2
        s2 = '.' * size
        self.failIf(s1 == s2)

    @bigmemtest(minsize=_2G + 10, memuse=1)
    def test_hash(self, size):
        # Not sure if we can do any meaningful tests here...  Even if we
        # start relying on the exact algorithm used, the result will be
        # different depending on the size of the C 'long int'.  Even this
        # test is dodgy (there's no *guarantee* that the two things should
        # have a different hash, even if they, in the current
        # implementation, almost always do.)
        s = '\x00' * size
        h1 = hash(s)
        del s
        s = '\x00' * (size + 1)
        self.failIf(h1 == hash(s))

class TupleTest(unittest.TestCase):

    # Tuples have a small, fixed-sized head and an array of pointers to
    # data.  Since we're testing 64-bit addressing, we can assume that the
    # pointers are 8 bytes, and that thus that the tuples take up 8 bytes
    # per size.

    # As a side-effect of testing long tuples, these tests happen to test
    # having more than 2<<31 references to any given object. Hence the
    # use of different types of objects as contents in different tests.

    @bigmemtest(minsize=_2G + 2, memuse=16)
    def test_compare(self, size):
        t1 = (u'',) * size
        t2 = (u'',) * size
        self.failUnless(t1 == t2)
        del t2
        t2 = (u'',) * (size + 1)
        self.failIf(t1 == t2)
        del t2
        t2 = (1,) * size
        self.failIf(t1 == t2)

    # Test concatenating into a single tuple of more than 2G in length,
    # and concatenating a tuple of more than 2G in length separately, so
    # the smaller test still gets run even if there isn't memory for the
    # larger test (but we still let the tester know the larger test is
    # skipped, in verbose mode.)
    def basic_concat_test(self, size):
        t = ((),) * size
        self.assertEquals(len(t), size)
        t = t + t
        self.assertEquals(len(t), size * 2)

    @bigmemtest(minsize=_2G // 2 + 2, memuse=24)
    def test_concat_small(self, size):
        return self.basic_concat_test(size)

    @bigmemtest(minsize=_2G + 2, memuse=24)
    def test_concat_large(self, size):
        return self.basic_concat_test(size)

    @bigmemtest(minsize=_2G // 5 + 10, memuse=8 * 5)
    def test_contains(self, size):
        t = (1, 2, 3, 4, 5) * size
        self.assertEquals(len(t), size * 5)
        self.failUnless(5 in t)
        self.failIf((1, 2, 3, 4, 5) in t)
        self.failIf(0 in t)

    @bigmemtest(minsize=_2G + 10, memuse=8)
    def test_hash(self, size):
        t1 = (0,) * size
        h1 = hash(t1)
        del t1
        t2 = (0,) * (size + 1)
        self.failIf(h1 == hash(t2))

    @bigmemtest(minsize=_2G + 10, memuse=8)
    def test_index_and_slice(self, size):
        t = (None,) * size
        self.assertEquals(len(t), size)
        self.assertEquals(t[-1], None)
        self.assertEquals(t[5], None)
        self.assertEquals(t[size - 1], None)
        self.assertRaises(IndexError, operator.getitem, t, size)
        self.assertEquals(t[:5], (None,) * 5)
        self.assertEquals(t[-5:], (None,) * 5)
        self.assertEquals(t[20:25], (None,) * 5)
        self.assertEquals(t[-25:-20], (None,) * 5)
        self.assertEquals(t[size - 5:], (None,) * 5)
        self.assertEquals(t[size - 5:size], (None,) * 5)
        self.assertEquals(t[size - 6:size - 2], (None,) * 4)
        self.assertEquals(t[size:size], ())
        self.assertEquals(t[size:size+5], ())

    # Like test_concat, split in two.
    def basic_test_repeat(self, size):
        t = ('',) * size
        self.assertEquals(len(t), size)
        t = t * 2
        self.assertEquals(len(t), size * 2)

    @bigmemtest(minsize=_2G // 2 + 2, memuse=24)
    def test_repeat_small(self, size):
        return self.basic_test_repeat(size)

    @bigmemtest(minsize=_2G + 2, memuse=24)
    def test_repeat_large(self, size):
        return self.basic_test_repeat(size)

    # Like test_concat, split in two.
    def basic_test_repr(self, size):
        t = (0,) * size
        s = repr(t)
        # The repr of a tuple of 0's is exactly three times the tuple length.
        self.assertEquals(len(s), size * 3)
        self.assertEquals(s[:5], '(0, 0')
        self.assertEquals(s[-5:], '0, 0)')
        self.assertEquals(s.count('0'), size)

    @bigmemtest(minsize=_2G // 3 + 2, memuse=8 + 3)
    def test_repr_small(self, size):
        return self.basic_test_repr(size)

    @bigmemtest(minsize=_2G + 2, memuse=8 + 3)
    def test_repr_large(self, size):
        return self.basic_test_repr(size)

class ListTest(unittest.TestCase):

    # Like tuples, lists have a small, fixed-sized head and an array of
    # pointers to data, so 8 bytes per size. Also like tuples, we make the
    # lists hold references to various objects to test their refcount
    # limits.

    @bigmemtest(minsize=_2G + 2, memuse=16)
    def test_compare(self, size):
        l1 = [u''] * size
        l2 = [u''] * size
        self.failUnless(l1 == l2)
        del l2
        l2 = [u''] * (size + 1)
        self.failIf(l1 == l2)
        del l2
        l2 = [2] * size
        self.failIf(l1 == l2)

    # Test concatenating into a single list of more than 2G in length,
    # and concatenating a list of more than 2G in length separately, so
    # the smaller test still gets run even if there isn't memory for the
    # larger test (but we still let the tester know the larger test is
    # skipped, in verbose mode.)
    def basic_test_concat(self, size):
        l = [[]] * size
        self.assertEquals(len(l), size)
        l = l + l
        self.assertEquals(len(l), size * 2)

    @bigmemtest(minsize=_2G // 2 + 2, memuse=24)
    def test_concat_small(self, size):
        return self.basic_test_concat(size)

    @bigmemtest(minsize=_2G + 2, memuse=24)
    def test_concat_large(self, size):
        return self.basic_test_concat(size)

    def basic_test_inplace_concat(self, size):
        l = [sys.stdout] * size
        l += l
        self.assertEquals(len(l), size * 2)
        self.failUnless(l[0] is l[-1])
        self.failUnless(l[size - 1] is l[size + 1])

    @bigmemtest(minsize=_2G // 2 + 2, memuse=24)
    def test_inplace_concat_small(self, size):
        return self.basic_test_inplace_concat(size)

    @bigmemtest(minsize=_2G + 2, memuse=24)
    def test_inplace_concat_large(self, size):
        return self.basic_test_inplace_concat(size)

    @bigmemtest(minsize=_2G // 5 + 10, memuse=8 * 5)
    def test_contains(self, size):
        l = [1, 2, 3, 4, 5] * size
        self.assertEquals(len(l), size * 5)
        self.failUnless(5 in l)
        self.failIf([1, 2, 3, 4, 5] in l)
        self.failIf(0 in l)

    @bigmemtest(minsize=_2G + 10, memuse=8)
    def test_hash(self, size):
        l = [0] * size
        self.failUnlessRaises(TypeError, hash, l)

    @bigmemtest(minsize=_2G + 10, memuse=8)
    def test_index_and_slice(self, size):
        l = [None] * size
        self.assertEquals(len(l), size)
        self.assertEquals(l[-1], None)
        self.assertEquals(l[5], None)
        self.assertEquals(l[size - 1], None)
        self.assertRaises(IndexError, operator.getitem, l, size)
        self.assertEquals(l[:5], [None] * 5)
        self.assertEquals(l[-5:], [None] * 5)
        self.assertEquals(l[20:25], [None] * 5)
        self.assertEquals(l[-25:-20], [None] * 5)
        self.assertEquals(l[size - 5:], [None] * 5)
        self.assertEquals(l[size - 5:size], [None] * 5)
        self.assertEquals(l[size - 6:size - 2], [None] * 4)
        self.assertEquals(l[size:size], [])
        self.assertEquals(l[size:size+5], [])

        l[size - 2] = 5
        self.assertEquals(len(l), size)
        self.assertEquals(l[-3:], [None, 5, None])
        self.assertEquals(l.count(5), 1)
        self.assertRaises(IndexError, operator.setitem, l, size, 6)
        self.assertEquals(len(l), size)

        l[size - 7:] = [1, 2, 3, 4, 5]
        size -= 2
        self.assertEquals(len(l), size)
        self.assertEquals(l[-7:], [None, None, 1, 2, 3, 4, 5])

        l[:7] = [1, 2, 3, 4, 5]
        size -= 2
        self.assertEquals(len(l), size)
        self.assertEquals(l[:7], [1, 2, 3, 4, 5, None, None])

        del l[size - 1]
        size -= 1
        self.assertEquals(len(l), size)
        self.assertEquals(l[-1], 4)

        del l[-2:]
        size -= 2
        self.assertEquals(len(l), size)
        self.assertEquals(l[-1], 2)

        del l[0]
        size -= 1
        self.assertEquals(len(l), size)
        self.assertEquals(l[0], 2)

        del l[:2]
        size -= 2
        self.assertEquals(len(l), size)
        self.assertEquals(l[0], 4)

    # Like test_concat, split in two.
    def basic_test_repeat(self, size):
        l = [] * size
        self.failIf(l)
        l = [''] * size
        self.assertEquals(len(l), size)
        l = l * 2
        self.assertEquals(len(l), size * 2)

    @bigmemtest(minsize=_2G // 2 + 2, memuse=24)
    def test_repeat_small(self, size):
        return self.basic_test_repeat(size)

    @bigmemtest(minsize=_2G + 2, memuse=24)
    def test_repeat_large(self, size):
        return self.basic_test_repeat(size)

    def basic_test_inplace_repeat(self, size):
        l = ['']
        l *= size
        self.assertEquals(len(l), size)
        self.failUnless(l[0] is l[-1])
        del l

        l = [''] * size
        l *= 2
        self.assertEquals(len(l), size * 2)
        self.failUnless(l[size - 1] is l[-1])

    @bigmemtest(minsize=_2G // 2 + 2, memuse=16)
    def test_inplace_repeat_small(self, size):
        return self.basic_test_inplace_repeat(size)

    @bigmemtest(minsize=_2G + 2, memuse=16)
    def test_inplace_repeat_large(self, size):
        return self.basic_test_inplace_repeat(size)

    def basic_test_repr(self, size):
        l = [0] * size
        s = repr(l)
        # The repr of a list of 0's is exactly three times the list length.
        self.assertEquals(len(s), size * 3)
        self.assertEquals(s[:5], '[0, 0')
        self.assertEquals(s[-5:], '0, 0]')
        self.assertEquals(s.count('0'), size)

    @bigmemtest(minsize=_2G // 3 + 2, memuse=8 + 3)
    def test_repr_small(self, size):
        return self.basic_test_repr(size)

    @bigmemtest(minsize=_2G + 2, memuse=8 + 3)
    def test_repr_large(self, size):
        return self.basic_test_repr(size)

    # list overallocates ~1/8th of the total size (on first expansion) so
    # the single list.append call puts memuse at 9 bytes per size.
    @bigmemtest(minsize=_2G, memuse=9)
    def test_append(self, size):
        l = [object()] * size
        l.append(object())
        self.assertEquals(len(l), size+1)
        self.failUnless(l[-3] is l[-2])
        self.failIf(l[-2] is l[-1])

    @bigmemtest(minsize=_2G // 5 + 2, memuse=8 * 5)
    def test_count(self, size):
        l = [1, 2, 3, 4, 5] * size
        self.assertEquals(l.count(1), size)
        self.assertEquals(l.count("1"), 0)

    def basic_test_extend(self, size):
        l = [file] * size
        l.extend(l)
        self.assertEquals(len(l), size * 2)
        self.failUnless(l[0] is l[-1])
        self.failUnless(l[size - 1] is l[size + 1])

    @bigmemtest(minsize=_2G // 2 + 2, memuse=16)
    def test_extend_small(self, size):
        return self.basic_test_extend(size)

    @bigmemtest(minsize=_2G + 2, memuse=16)
    def test_extend_large(self, size):
        return self.basic_test_extend(size)

    @bigmemtest(minsize=_2G // 5 + 2, memuse=8 * 5)
    def test_index(self, size):
        l = [1L, 2L, 3L, 4L, 5L] * size
        size *= 5
        self.assertEquals(l.index(1), 0)
        self.assertEquals(l.index(5, size - 5), size - 1)
        self.assertEquals(l.index(5, size - 5, size), size - 1)
        self.assertRaises(ValueError, l.index, 1, size - 4, size)
        self.assertRaises(ValueError, l.index, 6L)

    # This tests suffers from overallocation, just like test_append.
    @bigmemtest(minsize=_2G + 10, memuse=9)
    def test_insert(self, size):
        l = [1.0] * size
        l.insert(size - 1, "A")
        size += 1
        self.assertEquals(len(l), size)
        self.assertEquals(l[-3:], [1.0, "A", 1.0])

        l.insert(size + 1, "B")
        size += 1
        self.assertEquals(len(l), size)
        self.assertEquals(l[-3:], ["A", 1.0, "B"])

        l.insert(1, "C")
        size += 1
        self.assertEquals(len(l), size)
        self.assertEquals(l[:3], [1.0, "C", 1.0])
        self.assertEquals(l[size - 3:], ["A", 1.0, "B"])

    @bigmemtest(minsize=_2G // 5 + 4, memuse=8 * 5)
    def test_pop(self, size):
        l = [u"a", u"b", u"c", u"d", u"e"] * size
        size *= 5
        self.assertEquals(len(l), size)

        item = l.pop()
        size -= 1
        self.assertEquals(len(l), size)
        self.assertEquals(item, u"e")
        self.assertEquals(l[-2:], [u"c", u"d"])

        item = l.pop(0)
        size -= 1
        self.assertEquals(len(l), size)
        self.assertEquals(item, u"a")
        self.assertEquals(l[:2], [u"b", u"c"])

        item = l.pop(size - 2)
        size -= 1
        self.assertEquals(len(l), size)
        self.assertEquals(item, u"c")
        self.assertEquals(l[-2:], [u"b", u"d"])

    @bigmemtest(minsize=_2G + 10, memuse=8)
    def test_remove(self, size):
        l = [10] * size
        self.assertEquals(len(l), size)

        l.remove(10)
        size -= 1
        self.assertEquals(len(l), size)

        # Because of the earlier l.remove(), this append doesn't trigger
        # a resize.
        l.append(5)
        size += 1
        self.assertEquals(len(l), size)
        self.assertEquals(l[-2:], [10, 5])
        l.remove(5)
        size -= 1
        self.assertEquals(len(l), size)
        self.assertEquals(l[-2:], [10, 10])

    @bigmemtest(minsize=_2G // 5 + 2, memuse=8 * 5)
    def test_reverse(self, size):
        l = [1, 2, 3, 4, 5] * size
        l.reverse()
        self.assertEquals(len(l), size * 5)
        self.assertEquals(l[-5:], [5, 4, 3, 2, 1])
        self.assertEquals(l[:5], [5, 4, 3, 2, 1])

    @bigmemtest(minsize=_2G // 5 + 2, memuse=8 * 5)
    def test_sort(self, size):
        l = [1, 2, 3, 4, 5] * size
        l.sort()
        self.assertEquals(len(l), size * 5)
        self.assertEquals(l.count(1), size)
        self.assertEquals(l[:10], [1] * 10)
        self.assertEquals(l[-10:], [5] * 10)

def test_main():
    test_support.run_unittest(StrTest, TupleTest, ListTest)

if __name__ == '__main__':
    if len(sys.argv) > 1:
        test_support.set_memlimit(sys.argv[1])
    test_main()