summaryrefslogtreecommitdiff
path: root/runtests.py
blob: 541a47ea090917afd60049a2e90f46f58112dd3e (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
#!/usr/bin/env python
"""Run trollius unittests.

Usage:
  python3 runtests.py [flags] [pattern] ...

Patterns are matched against the fully qualified name of the test,
including package, module, class and method,
e.g. 'tests.test_events.PolicyTests.testPolicy'.

For full help, try --help.

runtests.py --coverage is equivalent of:

  $(COVERAGE) run --branch runtests.py -v
  $(COVERAGE) html $(list of files)
  $(COVERAGE) report -m $(list of files)

"""

# Originally written by Beech Horn (for NDB).

from __future__ import print_function
import optparse
import gc
import logging
import os
import random
import re
import sys
import textwrap
PY2 = (sys.version_info < (3,))
PY33 = (sys.version_info >= (3, 3))
if PY33:
    import importlib.machinery
else:
    import imp
try:
    import coverage
except ImportError:
    coverage = None
if PY2:
    sys.exc_clear()

try:
    import unittest2 as unittest
    from unittest2.signals import installHandler
except ImportError:
    import unittest
    from unittest.signals import installHandler

ARGS = optparse.OptionParser(description="Run all unittests.", usage="%prog [options] [pattern] [pattern2 ...]")
ARGS.add_option(
    '-v', '--verbose', type=int, dest='verbose',
    default=0, help='verbose')
ARGS.add_option(
    '-x', action="store_true", dest='exclude', help='exclude tests')
ARGS.add_option(
    '-f', '--failfast', action="store_true", default=False,
    dest='failfast', help='Stop on first fail or error')
ARGS.add_option(
    '--no-ssl', action="store_true", default=False,
    help='Disable the SSL module')
ARGS.add_option(
    '--no-concurrent', action="store_true", default=False,
    help='Disable the concurrent module')
ARGS.add_option(
    '-c', '--catch', action="store_true", default=False,
    dest='catchbreak', help='Catch control-C and display results')
ARGS.add_option(
    '--forever', action="store_true", dest='forever', default=False,
    help='run tests forever to catch sporadic errors')
ARGS.add_option(
    '--findleaks', action='store_true', dest='findleaks',
    help='detect tests that leak memory')
ARGS.add_option(
    '-r', '--randomize', action='store_true',
    help='randomize test execution order.')
ARGS.add_option(
    '--seed', type=int,
    help='random seed to reproduce a previous random run')
ARGS.add_option(
    '-q', action="store_true", dest='quiet', help='quiet')
ARGS.add_option(
    '--tests', action="store", dest='testsdir', default='tests',
    help='tests directory')
ARGS.add_option(
    '--coverage', action="store_true", dest='coverage',
    help='enable html coverage report')


if PY33:
    def load_module(modname, sourcefile):
        loader = importlib.machinery.SourceFileLoader(modname, sourcefile)
        return loader.load_module()
else:
    def load_module(modname, sourcefile):
        return imp.load_source(modname, sourcefile)


def load_modules(basedir, suffix='.py'):
    import trollius.test_utils

    def list_dir(prefix, dir):
        files = []

        modpath = os.path.join(dir, '__init__.py')
        if os.path.isfile(modpath):
            mod = os.path.split(dir)[-1]
            files.append(('{0}{1}'.format(prefix, mod), modpath))

            prefix = '{0}{1}.'.format(prefix, mod)

        for name in os.listdir(dir):
            path = os.path.join(dir, name)

            if os.path.isdir(path):
                files.extend(list_dir('{0}{1}.'.format(prefix, name), path))
            else:
                if (name != '__init__.py' and
                    name.endswith(suffix) and
                    not name.startswith(('.', '_'))):
                    files.append(('{0}{1}'.format(prefix, name[:-3]), path))

        return files

    mods = []
    for modname, sourcefile in list_dir('', basedir):
        if modname == 'runtests':
            continue
        if modname == 'test_asyncio' and not PY33:
            print("Skipping '{0}': need at least Python 3.3".format(modname),
                  file=sys.stderr)
            continue
        try:
            mod = load_module(modname, sourcefile)
            mods.append((mod, sourcefile))
        except SyntaxError:
            raise
        except trollius.test_utils.SkipTest as err:
            print("Skipping '{0}': {1}".format(modname, err), file=sys.stderr)

    return mods


def randomize_tests(tests, seed):
    if seed is None:
        seed = random.randrange(10000000)
    random.seed(seed)
    print("Randomize test execution order (seed: %s)" % seed)
    random.shuffle(tests._tests)


class TestsFinder:

    def __init__(self, testsdir, includes=(), excludes=()):
        self._testsdir = testsdir
        self._includes = includes
        self._excludes = excludes
        self.find_available_tests()

    def find_available_tests(self):
        """
        Find available test classes without instantiating them.
        """
        self._test_factories = []
        mods = [mod for mod, _ in load_modules(self._testsdir)]
        for mod in mods:
            for name in set(dir(mod)):
                if name.endswith('Tests'):
                    self._test_factories.append(getattr(mod, name))

    def load_tests(self):
        """
        Load test cases from the available test classes and apply
        optional include / exclude filters.
        """
        loader = unittest.TestLoader()
        suite = unittest.TestSuite()
        for test_factory in self._test_factories:
            tests = loader.loadTestsFromTestCase(test_factory)
            if self._includes:
                tests = [test
                         for test in tests
                         if any(re.search(pat, test.id())
                                for pat in self._includes)]
            if self._excludes:
                tests = [test
                         for test in tests
                         if not any(re.search(pat, test.id())
                                    for pat in self._excludes)]
            suite.addTests(tests)
        return suite


class TestResult(unittest.TextTestResult):

    def __init__(self, stream, descriptions, verbosity):
        super().__init__(stream, descriptions, verbosity)
        self.leaks = []

    def startTest(self, test):
        super().startTest(test)
        gc.collect()

    def addSuccess(self, test):
        super().addSuccess(test)
        gc.collect()
        if gc.garbage:
            if self.showAll:
                self.stream.writeln(
                    "    Warning: test created {} uncollectable "
                    "object(s).".format(len(gc.garbage)))
            # move the uncollectable objects somewhere so we don't see
            # them again
            self.leaks.append((self.getDescription(test), gc.garbage[:]))
            del gc.garbage[:]


class TestRunner(unittest.TextTestRunner):
    resultclass = TestResult

    def run(self, test):
        result = super().run(test)
        if result.leaks:
            self.stream.writeln("{0} tests leaks:".format(len(result.leaks)))
            for name, leaks in result.leaks:
                self.stream.writeln(' '*4 + name + ':')
                for leak in leaks:
                    self.stream.writeln(' '*8 + repr(leak))
        return result


def _runtests(args, tests):
    v = 0 if args.quiet else args.verbose + 1
    runner_factory = TestRunner if args.findleaks else unittest.TextTestRunner
    if args.randomize:
        randomize_tests(tests, args.seed)
    runner = runner_factory(verbosity=v, failfast=args.failfast)
    sys.stdout.flush()
    sys.stderr.flush()
    return runner.run(tests)


def runtests():
    args, pattern = ARGS.parse_args()

    if args.no_ssl:
        sys.modules['ssl'] = None

    if args.no_concurrent:
        sys.modules['concurrent'] = None

    if args.coverage and coverage is None:
        URL = "bitbucket.org/pypa/setuptools/raw/bootstrap/ez_setup.py"
        print(textwrap.dedent("""
            coverage package is not installed.

            To install coverage3 for Python 3, you need:
              - Setuptools (https://pypi.python.org/pypi/setuptools)

              What worked for me:
              - download {0}
                 * curl -O https://{0}
              - python3 ez_setup.py
              - python3 -m easy_install coverage
        """.format(URL)).strip())
        sys.exit(1)

    testsdir = os.path.abspath(args.testsdir)
    if not os.path.isdir(testsdir):
        print("Tests directory is not found: {0}\n".format(testsdir))
        ARGS.print_help()
        return

    excludes = includes = []
    if args.exclude:
        excludes = pattern
    else:
        includes = pattern

    v = 0 if args.quiet else args.verbose + 1
    failfast = args.failfast

    if args.coverage:
        cov = coverage.coverage(branch=True,
                                source=['asyncio'],
                                )
        cov.start()

    if v == 0:
        level = logging.CRITICAL
    elif v == 1:
        level = logging.ERROR
    elif v == 2:
        level = logging.WARNING
    elif v == 3:
        level = logging.INFO
    elif v >= 4:
        level = logging.DEBUG
    logging.basicConfig(level=level)

    finder = TestsFinder(args.testsdir, includes, excludes)
    if args.catchbreak:
        installHandler()
    import trollius.coroutines
    if trollius.coroutines._DEBUG:
        print("Run tests in debug mode")
    else:
        print("Run tests in release mode")
    try:
        tests = finder.load_tests()
        if args.forever:
            while True:
                result = _runtests(args, tests)
                if not result.wasSuccessful():
                    sys.exit(1)
        else:
            result = _runtests(args, tests)
            sys.exit(not result.wasSuccessful())
    finally:
        if args.coverage:
            cov.stop()
            cov.save()
            cov.html_report(directory='htmlcov')
            print("\nCoverage report:")
            cov.report(show_missing=False)
            here = os.path.dirname(os.path.abspath(__file__))
            print("\nFor html report:")
            print("open file://{0}/htmlcov/index.html".format(here))


if __name__ == '__main__':
    runtests()