summaryrefslogtreecommitdiff
path: root/Lib/test/test_regrtest.py
blob: 872ba646110738ba198f9cb1506ed3e53bcf2151 (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
"""
Tests of regrtest.py.

Note: test_regrtest cannot be run twice in parallel.
"""
from __future__ import print_function

import collections
import errno
import os.path
import platform
import re
import subprocess
import sys
import sysconfig
import tempfile
import textwrap
import unittest
from test import support
# Use utils alias to use the same code for TestUtils in master and 2.7 branches
import regrtest as utils


Py_DEBUG = hasattr(sys, 'getobjects')
ROOT_DIR = os.path.join(os.path.dirname(__file__), '..', '..')
ROOT_DIR = os.path.abspath(os.path.normpath(ROOT_DIR))

TEST_INTERRUPTED = textwrap.dedent("""
    from signal import SIGINT
    try:
        from _testcapi import raise_signal
        raise_signal(SIGINT)
    except ImportError:
        import os
        os.kill(os.getpid(), SIGINT)
    """)


SubprocessRun = collections.namedtuple('SubprocessRun',
                                       'returncode stdout stderr')


class BaseTestCase(unittest.TestCase):
    TEST_UNIQUE_ID = 1
    TESTNAME_PREFIX = 'test_regrtest_'
    TESTNAME_REGEX = r'test_[a-zA-Z0-9_]+'

    def setUp(self):
        self.testdir = os.path.realpath(os.path.dirname(__file__))

        self.tmptestdir = tempfile.mkdtemp()
        self.addCleanup(support.rmtree, self.tmptestdir)

    def create_test(self, name=None, code=None):
        if not name:
            name = 'noop%s' % BaseTestCase.TEST_UNIQUE_ID
            BaseTestCase.TEST_UNIQUE_ID += 1

        if code is None:
            code = textwrap.dedent("""
                    import unittest
                    from test import support

                    class Tests(unittest.TestCase):
                        def test_empty_test(self):
                            pass

                    def test_main():
                        support.run_unittest(Tests)
                """)

        # test_regrtest cannot be run twice in parallel because
        # of setUp() and create_test()
        name = self.TESTNAME_PREFIX + name
        path = os.path.join(self.tmptestdir, name + '.py')

        self.addCleanup(support.unlink, path)
        # Use O_EXCL to ensure that we do not override existing tests
        try:
            fd = os.open(path, os.O_WRONLY | os.O_CREAT | os.O_EXCL)
        except OSError as exc:
            if (exc.errno in (errno.EACCES, errno.EPERM)
               and not sysconfig.is_python_build()):
                self.skipTest("cannot write %s: %s" % (path, exc))
            else:
                raise
        else:
            with os.fdopen(fd, 'w') as fp:
                fp.write(code)
        return name

    def regex_search(self, regex, output):
        match = re.search(regex, output, re.MULTILINE)
        if not match:
            self.fail("%r not found in %r" % (regex, output))
        return match

    def check_line(self, output, regex):
        regex = re.compile(r'^' + regex, re.MULTILINE)
        self.assertRegexpMatches(output, regex)

    def parse_executed_tests(self, output):
        regex = (r'^[0-9]+:[0-9]+:[0-9]+ (?:load avg: [0-9]+\.[0-9]{2} )?\[ *[0-9]+(?:/ *[0-9]+)*\] (%s)'
                 % self.TESTNAME_REGEX)
        parser = re.finditer(regex, output, re.MULTILINE)
        return list(match.group(1) for match in parser)

    def check_executed_tests(self, output, tests, skipped=(), failed=(),
                             env_changed=(), omitted=(),
                             rerun=(), no_test_ran=(),
                             randomize=False, interrupted=False,
                             fail_env_changed=False):
        if isinstance(tests, str):
            tests = [tests]
        if isinstance(skipped, str):
            skipped = [skipped]
        if isinstance(failed, str):
            failed = [failed]
        if isinstance(env_changed, str):
            env_changed = [env_changed]
        if isinstance(omitted, str):
            omitted = [omitted]
        if isinstance(rerun, str):
            rerun = [rerun]
        if isinstance(no_test_ran, str):
            no_test_ran = [no_test_ran]

        executed = self.parse_executed_tests(output)
        if randomize:
            self.assertEqual(set(executed), set(tests), output)
        else:
            self.assertEqual(executed, tests, (executed, tests, output))

        def plural(count):
            return 's' if count != 1 else ''

        def list_regex(line_format, tests):
            count = len(tests)
            names = ' '.join(sorted(tests))
            regex = line_format % (count, plural(count))
            regex = r'%s:\n    %s$' % (regex, names)
            return regex

        if skipped:
            regex = list_regex('%s test%s skipped', skipped)
            self.check_line(output, regex)

        if failed:
            regex = list_regex('%s test%s failed', failed)
            self.check_line(output, regex)

        if env_changed:
            regex = list_regex('%s test%s altered the execution environment',
                               env_changed)
            self.check_line(output, regex)

        if omitted:
            regex = list_regex('%s test%s omitted', omitted)
            self.check_line(output, regex)

        if rerun:
            regex = list_regex('%s re-run test%s', rerun)
            self.check_line(output, regex)
            self.check_line(output, "Re-running failed tests in verbose mode")
            for name in rerun:
                regex = "Re-running test %r in verbose mode" % name
                self.check_line(output, regex)

        good = (len(tests) - len(skipped) - len(failed)
                - len(omitted) - len(env_changed) - len(no_test_ran))
        if good:
            regex = r'%s test%s OK\.$' % (good, plural(good))
            if not skipped and not failed and good > 1:
                regex = 'All %s' % regex
            self.check_line(output, regex)

        if interrupted:
            self.check_line(output, 'Test suite interrupted by signal SIGINT.')

        result = []
        if failed:
            result.append('FAILURE')
        elif fail_env_changed and env_changed:
            result.append('ENV CHANGED')
        if interrupted:
            result.append('INTERRUPTED')
        if not any((good, result, failed, interrupted, skipped,
                    env_changed, fail_env_changed)):
            result.append("NO TEST RUN")
        elif not result:
            result.append('SUCCESS')
        result = ', '.join(result)
        if rerun:
            self.check_line(output, 'Tests result: %s' % result)
            result = 'FAILURE then %s' % result

        self.check_line(output, 'Tests result: %s' % result)

    def parse_random_seed(self, output):
        match = self.regex_search(r'Using random seed ([0-9]+)', output)
        randseed = int(match.group(1))
        self.assertTrue(0 <= randseed <= 10000000, randseed)
        return randseed

    def run_command(self, args, input=None, exitcode=0, **kw):
        if not input:
            input = ''
        if 'stderr' not in kw:
            kw['stderr'] = subprocess.PIPE
        proc = subprocess.Popen(args,
                                universal_newlines=True,
                                stdout=subprocess.PIPE,
                                **kw)
        stdout, stderr = proc.communicate(input=input)
        if proc.returncode != exitcode:
            msg = ("Command %s failed with exit code %s\n"
                   "\n"
                   "stdout:\n"
                   "---\n"
                   "%s\n"
                   "---\n"
                   % (str(args), proc.returncode, stdout))
            if proc.stderr:
                msg += ("\n"
                        "stderr:\n"
                        "---\n"
                        "%s"
                        "---\n"
                        % stderr)
            self.fail(msg)
        return SubprocessRun(proc.returncode, stdout, stderr)

    def run_python(self, args, **kw):
        args = [sys.executable] + list(args)
        proc = self.run_command(args, **kw)
        return proc.stdout


class ProgramsTestCase(BaseTestCase):
    """
    Test various ways to run the Python test suite. Use options close
    to options used on the buildbot.
    """

    NTEST = 4

    def setUp(self):
        super(ProgramsTestCase, self).setUp()

        # Create NTEST tests doing nothing
        self.tests = [self.create_test() for index in range(self.NTEST)]

        self.python_args = ['-Wd', '-3', '-E', '-bb', '-tt']
        self.regrtest_args = ['-uall', '-rwW',
                              '--testdir=%s' % self.tmptestdir]

    def check_output(self, output):
        self.parse_random_seed(output)
        self.check_executed_tests(output, self.tests, randomize=True)

    def run_tests(self, args):
        output = self.run_python(args)
        self.check_output(output)

    def test_script_regrtest(self):
        # Lib/test/regrtest.py
        script = os.path.join(self.testdir, 'regrtest.py')

        args = self.python_args + [script] + self.regrtest_args + self.tests
        self.run_tests(args)

    def test_module_test(self):
        # -m test
        args = self.python_args + ['-m', 'test'] + self.regrtest_args + self.tests
        self.run_tests(args)

    def test_module_regrtest(self):
        # -m test.regrtest
        args = self.python_args + ['-m', 'test.regrtest'] + self.regrtest_args + self.tests
        self.run_tests(args)

    def test_module_autotest(self):
        # -m test.autotest
        args = self.python_args + ['-m', 'test.autotest'] + self.regrtest_args + self.tests
        self.run_tests(args)

    def test_module_from_test_autotest(self):
        # from test import autotest
        code = 'from test import autotest'
        args = self.python_args + ['-c', code] + self.regrtest_args + self.tests
        self.run_tests(args)

    def test_script_autotest(self):
        # Lib/test/autotest.py
        script = os.path.join(self.testdir, 'autotest.py')
        args = self.python_args + [script] + self.regrtest_args + self.tests
        self.run_tests(args)

    def run_batch(self, *args):
        proc = self.run_command(args)
        self.check_output(proc.stdout)

    def need_pcbuild(self):
        exe = os.path.normpath(os.path.abspath(sys.executable))
        parts = exe.split(os.path.sep)
        if len(parts) < 3:
            # it's not a python build, python is likely to be installed
            return

        build_dir = parts[-3]
        if build_dir.lower() != 'pcbuild':
            self.skipTest("Tools/buildbot/test.bat requires PCbuild build, "
                          "found %s" % build_dir)

    @unittest.skipUnless(sysconfig.is_python_build(),
                         'test.bat script is not installed')
    @unittest.skipUnless(sys.platform == 'win32', 'Windows only')
    def test_tools_buildbot_test(self):
        self.need_pcbuild()

        # Tools\buildbot\test.bat
        script = os.path.join(ROOT_DIR, 'Tools', 'buildbot', 'test.bat')
        test_args = ['--testdir=%s' % self.tmptestdir]
        if platform.architecture()[0] == '64bit':
            test_args.append('-x64')   # 64-bit build
        if not Py_DEBUG:
            test_args.append('+d')     # Release build, use python.exe

        args = [script] + test_args + self.tests
        self.run_batch(*args)

    @unittest.skipUnless(sys.platform == 'win32', 'Windows only')
    def test_pcbuild_rt(self):
        self.need_pcbuild()

        # PCbuild\rt.bat
        script = os.path.join(ROOT_DIR, r'PCbuild\rt.bat')
        rt_args = ["-q"]             # Quick, don't run tests twice
        if platform.architecture()[0] == '64bit':
            rt_args.append('-x64')   # 64-bit build
        if Py_DEBUG:
            rt_args.append('-d')     # Debug build, use python_d.exe
        args = [script] + rt_args + self.regrtest_args + self.tests
        self.run_batch(*args)


class ArgsTestCase(BaseTestCase):
    """
    Test arguments of the Python test suite.
    """

    def run_tests(self, *testargs, **kw):
        cmdargs = ('-m', 'test', '--testdir=%s' % self.tmptestdir) + testargs
        return self.run_python(cmdargs, **kw)

    def test_failing_test(self):
        # test a failing test
        code = textwrap.dedent("""
            import unittest
            from test import support

            class FailingTest(unittest.TestCase):
                def test_failing(self):
                    self.fail("bug")

            def test_main():
                support.run_unittest(FailingTest)
        """)
        test_ok = self.create_test('ok')
        test_failing = self.create_test('failing', code=code)
        tests = [test_ok, test_failing]

        output = self.run_tests(*tests, exitcode=2)
        self.check_executed_tests(output, tests, failed=test_failing)

    def test_resources(self):
        # test -u command line option
        tests = {}
        for resource in ('audio', 'network'):
            code = textwrap.dedent("""
                        from test import support; support.requires(%r)
                        import unittest
                        class PassingTest(unittest.TestCase):
                            def test_pass(self):
                                pass

                        def test_main():
                            support.run_unittest(PassingTest)
                    """ % resource)

            tests[resource] = self.create_test(resource, code)
        test_names = sorted(tests.values())

        # -u all: 2 resources enabled
        output = self.run_tests('-u', 'all', *test_names)
        self.check_executed_tests(output, test_names)

        # -u audio: 1 resource enabled
        output = self.run_tests('-uaudio', *test_names)
        self.check_executed_tests(output, test_names,
                                  skipped=tests['network'])

        # no option: 0 resources enabled
        output = self.run_tests(*test_names)
        self.check_executed_tests(output, test_names,
                                  skipped=test_names)

    def test_random(self):
        # test -r and --randseed command line option
        code = textwrap.dedent("""
            import random
            print("TESTRANDOM: %s" % random.randint(1, 1000))
        """)
        test = self.create_test('random', code)

        # first run to get the output with the random seed
        output = self.run_tests('-r', '-v', test)
        randseed = self.parse_random_seed(output)
        match = self.regex_search(r'TESTRANDOM: ([0-9]+)', output)
        test_random = int(match.group(1))

        # try to reproduce with the random seed
        output = self.run_tests('-r', '-v', '--randseed=%s' % randseed, test)
        randseed2 = self.parse_random_seed(output)
        self.assertEqual(randseed2, randseed)

        match = self.regex_search(r'TESTRANDOM: ([0-9]+)', output)
        test_random2 = int(match.group(1))
        self.assertEqual(test_random2, test_random)

    def test_fromfile(self):
        # test --fromfile
        tests = [self.create_test() for index in range(5)]

        # Write the list of files using a format similar to regrtest output:
        # [1/2] test_1
        # [2/2] test_2
        filename = support.TESTFN
        self.addCleanup(support.unlink, filename)

        # test format 'test_opcodes'
        with open(filename, "w") as fp:
            for name in tests:
                print(name, file=fp)

        output = self.run_tests('--fromfile', filename)
        self.check_executed_tests(output, tests)

    def test_interrupted(self):
        code = TEST_INTERRUPTED
        test = self.create_test('sigint', code=code)
        output = self.run_tests(test, exitcode=130)
        self.check_executed_tests(output, test, omitted=test,
                                  interrupted=True)

    def test_slowest(self):
        # test --slow
        tests = [self.create_test() for index in range(3)]
        output = self.run_tests("--slowest", *tests)
        self.check_executed_tests(output, tests)
        regex = ('10 slowest tests:\n'
                 '(?:- %s: .*\n){%s}'
                 % (self.TESTNAME_REGEX, len(tests)))
        self.check_line(output, regex)

    def test_slow_interrupted(self):
        # Issue #25373: test --slowest with an interrupted test
        code = TEST_INTERRUPTED
        test = self.create_test("sigint", code=code)

        try:
            import threading
            tests = (False, True)
        except ImportError:
            tests = (False,)
        for multiprocessing in tests:
            if multiprocessing:
                args = ("--slowest", "-j2", test)
            else:
                args = ("--slowest", test)
            output = self.run_tests(*args, exitcode=130)
            self.check_executed_tests(output, test,
                                      omitted=test, interrupted=True)

            regex = ('10 slowest tests:\n')
            self.check_line(output, regex)

    def test_coverage(self):
        # test --coverage
        test = self.create_test('coverage')
        output = self.run_tests("--coverage", test)
        self.check_executed_tests(output, [test])
        regex = (r'lines +cov% +module +\(path\)\n'
                 r'(?: *[0-9]+ *[0-9]{1,2}% *[^ ]+ +\([^)]+\)+)+')
        self.check_line(output, regex)

    def test_forever(self):
        # test --forever
        code = textwrap.dedent("""
            import __builtin__
            import unittest
            from test import support

            class ForeverTester(unittest.TestCase):
                def test_run(self):
                    # Store the state in the __builtin__ module, because the test
                    # module is reload at each run
                    if 'RUN' in __builtin__.__dict__:
                        __builtin__.__dict__['RUN'] += 1
                        if __builtin__.__dict__['RUN'] >= 3:
                            self.fail("fail at the 3rd runs")
                    else:
                        __builtin__.__dict__['RUN'] = 1

            def test_main():
                support.run_unittest(ForeverTester)
        """)
        test = self.create_test('forever', code=code)
        output = self.run_tests('--forever', test, exitcode=2)
        self.check_executed_tests(output, [test]*3, failed=test)

    def check_leak(self, code, what):
        test = self.create_test('huntrleaks', code=code)

        filename = 'reflog.txt'
        self.addCleanup(support.unlink, filename)
        output = self.run_tests('--huntrleaks', '3:3:', test,
                                exitcode=2,
                                stderr=subprocess.STDOUT)
        self.check_executed_tests(output, [test], failed=test)

        line = 'beginning 6 repetitions\n123456\n......\n'
        self.check_line(output, re.escape(line))

        line2 = '%s leaked [1, 1, 1] %s, sum=3\n' % (test, what)
        self.assertIn(line2, output)

        with open(filename) as fp:
            reflog = fp.read()
            self.assertIn(line2, reflog)

    @unittest.skipUnless(Py_DEBUG, 'need a debug build')
    @support.requires_type_collecting
    def test_huntrleaks(self):
        # test --huntrleaks
        code = textwrap.dedent("""
            import unittest
            from test import support

            GLOBAL_LIST = []

            class RefLeakTest(unittest.TestCase):
                def test_leak(self):
                    GLOBAL_LIST.append(object())

            def test_main():
                support.run_unittest(RefLeakTest)
        """)
        self.check_leak(code, 'references')

    @unittest.skipUnless(Py_DEBUG, 'need a debug build')
    def test_huntrleaks_fd_leak(self):
        # test --huntrleaks for file descriptor leak
        code = textwrap.dedent("""
            import os
            import unittest
            from test import support

            class FDLeakTest(unittest.TestCase):
                def test_leak(self):
                    fd = os.open(__file__, os.O_RDONLY)
                    # bug: never close the file descriptor

            def test_main():
                support.run_unittest(FDLeakTest)
        """)
        self.check_leak(code, 'file descriptors')

    def test_list_tests(self):
        # test --list-tests
        tests = [self.create_test() for i in range(5)]
        output = self.run_tests('--list-tests', *tests)
        self.assertEqual(output.rstrip().splitlines(),
                         tests)

    def test_list_cases(self):
        # test --list-cases
        code = textwrap.dedent("""
            import unittest

            class Tests(unittest.TestCase):
                def test_method1(self):
                    pass
                def test_method2(self):
                    pass
        """)
        testname = self.create_test(code=code)

        # Test --list-cases
        all_methods = ['%s.Tests.test_method1' % testname,
                       '%s.Tests.test_method2' % testname]
        output = self.run_tests('--list-cases', testname)
        self.assertEqual(output.splitlines(), all_methods)

        # Test --list-cases with --match
        all_methods = ['%s.Tests.test_method1' % testname]
        output = self.run_tests('--list-cases',
                                '-m', 'test_method1',
                                testname)
        self.assertEqual(output.splitlines(), all_methods)

    @unittest.skipIf(sys.platform.startswith('aix'),
                     "support._crash_python() doesn't work on AIX")
    def test_crashed(self):
        # Any code which causes a crash
        code = 'import test.support; test.support._crash_python()'
        crash_test = self.create_test(name="crash", code=code)
        ok_test = self.create_test(name="ok")

        tests = [crash_test, ok_test]
        output = self.run_tests("-j2", *tests, exitcode=2)
        self.check_executed_tests(output, tests, failed=crash_test,
                                  randomize=True)

    def parse_methods(self, output):
        regex = re.compile("^(test[^ ]+).*ok$", flags=re.MULTILINE)
        return [match.group(1) for match in regex.finditer(output)]

    def test_matchfile(self):
        # Any code which causes a crash
        code = textwrap.dedent("""
            import unittest
            from test import support

            class Tests(unittest.TestCase):
                def test_method1(self):
                    pass
                def test_method2(self):
                    pass
                def test_method3(self):
                    pass
                def test_method4(self):
                    pass

            def test_main():
                support.run_unittest(Tests)
        """)
        all_methods = ['test_method1', 'test_method2',
                       'test_method3', 'test_method4']
        testname = self.create_test(code=code)

        # by default, all methods should be run
        output = self.run_tests("-v", testname)
        methods = self.parse_methods(output)
        self.assertEqual(methods, all_methods)

        # only run a subset
        filename = support.TESTFN
        self.addCleanup(support.unlink, filename)

        subset = [
            # only match the method name
            'test_method1',
            # match the full identifier
            '%s.Tests.test_method3' % testname]
        with open(filename, "w") as fp:
            for name in subset:
                print(name, file=fp)

        output = self.run_tests("-v", "--matchfile", filename, testname)
        methods = self.parse_methods(output)
        subset = ['test_method1', 'test_method3']
        self.assertEqual(methods, subset)

    # bpo-34021: The test fails randomly for an unknown reason
    # on "x86 Windows XP VS9.0 2.7" buildbot worker.
    @unittest.skipIf(sys.platform == "win32", "test fails randomly on Windows")
    def test_env_changed(self):
        code = textwrap.dedent("""
            import unittest
            from test import support

            class Tests(unittest.TestCase):
                def test_env_changed(self):
                    open("env_changed", "w").close()

            def test_main():
                support.run_unittest(Tests)
        """)
        testname = self.create_test(code=code)

        # don't fail by default
        output = self.run_tests(testname)
        self.check_executed_tests(output, [testname], env_changed=testname)

        # fail with --fail-env-changed
        output = self.run_tests("--fail-env-changed", testname, exitcode=3)
        self.check_executed_tests(output, [testname], env_changed=testname,
                                  fail_env_changed=True)

    def test_rerun_fail(self):
        code = textwrap.dedent("""
            import unittest
            from test import support

            class Tests(unittest.TestCase):
                def test_bug(self):
                    # test always fail
                    self.fail("bug")

            def test_main():
                support.run_unittest(Tests)
        """)
        testname = self.create_test(code=code)

        output = self.run_tests("-w", testname, exitcode=2)
        self.check_executed_tests(output, [testname],
                                  failed=testname, rerun=testname)

    def test_no_tests_ran(self):
        code = textwrap.dedent("""
            import unittest
            from test import support

            class Tests(unittest.TestCase):
                def test_bug(self):
                    pass

            def test_main():
                support.run_unittest(Tests)
        """)
        testname = self.create_test(code=code)

        output = self.run_tests("-m", "nosuchtest", testname, exitcode=0)
        self.check_executed_tests(output, [testname], no_test_ran=testname)

    def test_no_tests_ran_skip(self):
        code = textwrap.dedent("""
            import unittest

            class Tests(unittest.TestCase):
                def test_skipped(self):
                    self.skipTest("because")
        """)
        testname = self.create_test(code=code)

        output = self.run_tests(testname, exitcode=0)
        self.check_executed_tests(output, [testname])

    def test_no_tests_ran_multiple_tests_nonexistent(self):
        code = textwrap.dedent("""
            import unittest
            from test import support

            class Tests(unittest.TestCase):
                def test_bug(self):
                    pass

            def test_main():
                support.run_unittest(Tests)
        """)
        testname = self.create_test(code=code)
        testname2 = self.create_test(code=code)

        output = self.run_tests("-m", "nosuchtest",
                                testname, testname2,
                                exitcode=0)
        self.check_executed_tests(output, [testname, testname2],
                                  no_test_ran=[testname, testname2])

    def test_no_test_ran_some_test_exist_some_not(self):
        code = textwrap.dedent("""
            import unittest
            from test import support

            class Tests(unittest.TestCase):
                def test_bug(self):
                    pass

            def test_main():
                support.run_unittest(Tests)
        """)
        testname = self.create_test(code=code)
        other_code = textwrap.dedent("""
            import unittest
            from test import support

            class Tests(unittest.TestCase):
                def test_other_bug(self):
                    pass

            def test_main():
                support.run_unittest(Tests)
        """)
        testname2 = self.create_test(code=other_code)

        output = self.run_tests("-m", "nosuchtest", "-m", "test_other_bug",
                                testname, testname2,
                                exitcode=0)
        self.check_executed_tests(output, [testname, testname2],
                                  no_test_ran=[testname])


class TestUtils(unittest.TestCase):
    def test_format_duration(self):
        self.assertEqual(utils.format_duration(0),
                         '0 ms')
        self.assertEqual(utils.format_duration(1e-9),
                         '1 ms')
        self.assertEqual(utils.format_duration(10e-3),
                         '10 ms')
        self.assertEqual(utils.format_duration(1.5),
                         '1 sec 500 ms')
        self.assertEqual(utils.format_duration(1),
                         '1 sec')
        self.assertEqual(utils.format_duration(2 * 60),
                         '2 min')
        self.assertEqual(utils.format_duration(2 * 60 + 1),
                         '2 min 1 sec')
        self.assertEqual(utils.format_duration(3 * 3600),
                         '3 hour')
        self.assertEqual(utils.format_duration(3 * 3600  + 2 * 60 + 1),
                         '3 hour 2 min')
        self.assertEqual(utils.format_duration(3 * 3600 + 1),
                         '3 hour 1 sec')


def test_main():
    support.run_unittest(ProgramsTestCase, ArgsTestCase, TestUtils)


if __name__ == "__main__":
    test_main()