cpython/Lib/test/test_regrtest.py

1723 lines
63 KiB
Python
Raw Normal View History

"""
Tests of regrtest.py.
Note: test_regrtest cannot be run twice in parallel.
"""
import contextlib
import glob
import io
import locale
import os.path
import platform
import re
import subprocess
import sys
import sysconfig
import tempfile
import textwrap
import time
import unittest
from test import libregrtest
from test import support
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
from test.support import os_helper, TestStats
from test.libregrtest import utils, setup
if not support.has_subprocess_support:
raise unittest.SkipTest("test module requires subprocess")
Py_DEBUG = hasattr(sys, 'gettotalrefcount')
ROOT_DIR = os.path.join(os.path.dirname(__file__), '..', '..')
ROOT_DIR = os.path.abspath(os.path.normpath(ROOT_DIR))
LOG_PREFIX = r'[0-9]+:[0-9]+:[0-9]+ (?:load avg: [0-9]+\.[0-9]{2} )?'
EXITCODE_BAD_TEST = 2
EXITCODE_ENV_CHANGED = 3
EXITCODE_NO_TESTS_RAN = 4
EXITCODE_INTERRUPTED = 130
TEST_INTERRUPTED = textwrap.dedent("""
from signal import SIGINT, raise_signal
try:
raise_signal(SIGINT)
except ImportError:
import os
os.kill(os.getpid(), SIGINT)
""")
class ParseArgsTestCase(unittest.TestCase):
"""
Test regrtest's argument parsing, function _parse_args().
"""
def checkError(self, args, msg):
with support.captured_stderr() as err, self.assertRaises(SystemExit):
libregrtest._parse_args(args)
self.assertIn(msg, err.getvalue())
def test_help(self):
for opt in '-h', '--help':
with self.subTest(opt=opt):
with support.captured_stdout() as out, \
self.assertRaises(SystemExit):
libregrtest._parse_args([opt])
self.assertIn('Run Python regression tests.', out.getvalue())
def test_timeout(self):
ns = libregrtest._parse_args(['--timeout', '4.2'])
self.assertEqual(ns.timeout, 4.2)
self.checkError(['--timeout'], 'expected one argument')
self.checkError(['--timeout', 'foo'], 'invalid float value')
def test_wait(self):
ns = libregrtest._parse_args(['--wait'])
self.assertTrue(ns.wait)
def test_worker_args(self):
ns = libregrtest._parse_args(['--worker-args', '[[], {}]'])
self.assertEqual(ns.worker_args, '[[], {}]')
self.checkError(['--worker-args'], 'expected one argument')
def test_start(self):
for opt in '-S', '--start':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt, 'foo'])
self.assertEqual(ns.start, 'foo')
self.checkError([opt], 'expected one argument')
def test_verbose(self):
ns = libregrtest._parse_args(['-v'])
self.assertEqual(ns.verbose, 1)
ns = libregrtest._parse_args(['-vvv'])
self.assertEqual(ns.verbose, 3)
ns = libregrtest._parse_args(['--verbose'])
self.assertEqual(ns.verbose, 1)
ns = libregrtest._parse_args(['--verbose'] * 3)
self.assertEqual(ns.verbose, 3)
ns = libregrtest._parse_args([])
self.assertEqual(ns.verbose, 0)
def test_verbose2(self):
for opt in '-w', '--verbose2':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt])
self.assertTrue(ns.verbose2)
def test_verbose3(self):
for opt in '-W', '--verbose3':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt])
self.assertTrue(ns.verbose3)
def test_quiet(self):
for opt in '-q', '--quiet':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt])
self.assertTrue(ns.quiet)
self.assertEqual(ns.verbose, 0)
def test_slowest(self):
for opt in '-o', '--slowest':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt])
self.assertTrue(ns.print_slow)
def test_header(self):
ns = libregrtest._parse_args(['--header'])
self.assertTrue(ns.header)
ns = libregrtest._parse_args(['--verbose'])
self.assertTrue(ns.header)
def test_randomize(self):
for opt in '-r', '--randomize':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt])
self.assertTrue(ns.randomize)
def test_randseed(self):
ns = libregrtest._parse_args(['--randseed', '12345'])
self.assertEqual(ns.random_seed, 12345)
self.assertTrue(ns.randomize)
self.checkError(['--randseed'], 'expected one argument')
self.checkError(['--randseed', 'foo'], 'invalid int value')
def test_fromfile(self):
for opt in '-f', '--fromfile':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt, 'foo'])
self.assertEqual(ns.fromfile, 'foo')
self.checkError([opt], 'expected one argument')
self.checkError([opt, 'foo', '-s'], "don't go together")
def test_exclude(self):
for opt in '-x', '--exclude':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt])
self.assertTrue(ns.exclude)
def test_single(self):
for opt in '-s', '--single':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt])
self.assertTrue(ns.single)
self.checkError([opt, '-f', 'foo'], "don't go together")
def test_ignore(self):
for opt in '-i', '--ignore':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt, 'pattern'])
self.assertEqual(ns.ignore_tests, ['pattern'])
self.checkError([opt], 'expected one argument')
self.addCleanup(os_helper.unlink, os_helper.TESTFN)
with open(os_helper.TESTFN, "w") as fp:
print('matchfile1', file=fp)
print('matchfile2', file=fp)
filename = os.path.abspath(os_helper.TESTFN)
ns = libregrtest._parse_args(['-m', 'match',
'--ignorefile', filename])
self.assertEqual(ns.ignore_tests,
['matchfile1', 'matchfile2'])
def test_match(self):
for opt in '-m', '--match':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt, 'pattern'])
self.assertEqual(ns.match_tests, ['pattern'])
self.checkError([opt], 'expected one argument')
ns = libregrtest._parse_args(['-m', 'pattern1',
'-m', 'pattern2'])
self.assertEqual(ns.match_tests, ['pattern1', 'pattern2'])
self.addCleanup(os_helper.unlink, os_helper.TESTFN)
with open(os_helper.TESTFN, "w") as fp:
print('matchfile1', file=fp)
print('matchfile2', file=fp)
filename = os.path.abspath(os_helper.TESTFN)
ns = libregrtest._parse_args(['-m', 'match',
'--matchfile', filename])
self.assertEqual(ns.match_tests,
['match', 'matchfile1', 'matchfile2'])
def test_failfast(self):
for opt in '-G', '--failfast':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt, '-v'])
self.assertTrue(ns.failfast)
ns = libregrtest._parse_args([opt, '-W'])
self.assertTrue(ns.failfast)
self.checkError([opt], '-G/--failfast needs either -v or -W')
def test_use(self):
for opt in '-u', '--use':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt, 'gui,network'])
self.assertEqual(ns.use_resources, ['gui', 'network'])
ns = libregrtest._parse_args([opt, 'gui,none,network'])
self.assertEqual(ns.use_resources, ['network'])
expected = list(libregrtest.ALL_RESOURCES)
expected.remove('gui')
ns = libregrtest._parse_args([opt, 'all,-gui'])
self.assertEqual(ns.use_resources, expected)
self.checkError([opt], 'expected one argument')
self.checkError([opt, 'foo'], 'invalid resource')
# all + a resource not part of "all"
ns = libregrtest._parse_args([opt, 'all,tzdata'])
self.assertEqual(ns.use_resources,
list(libregrtest.ALL_RESOURCES) + ['tzdata'])
# test another resource which is not part of "all"
ns = libregrtest._parse_args([opt, 'extralargefile'])
self.assertEqual(ns.use_resources, ['extralargefile'])
def test_memlimit(self):
for opt in '-M', '--memlimit':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt, '4G'])
self.assertEqual(ns.memlimit, '4G')
self.checkError([opt], 'expected one argument')
def test_testdir(self):
ns = libregrtest._parse_args(['--testdir', 'foo'])
self.assertEqual(ns.testdir, os.path.join(os_helper.SAVEDCWD, 'foo'))
self.checkError(['--testdir'], 'expected one argument')
def test_runleaks(self):
for opt in '-L', '--runleaks':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt])
self.assertTrue(ns.runleaks)
def test_huntrleaks(self):
for opt in '-R', '--huntrleaks':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt, ':'])
self.assertEqual(ns.huntrleaks, (5, 4, 'reflog.txt'))
ns = libregrtest._parse_args([opt, '6:'])
self.assertEqual(ns.huntrleaks, (6, 4, 'reflog.txt'))
ns = libregrtest._parse_args([opt, ':3'])
self.assertEqual(ns.huntrleaks, (5, 3, 'reflog.txt'))
ns = libregrtest._parse_args([opt, '6:3:leaks.log'])
self.assertEqual(ns.huntrleaks, (6, 3, 'leaks.log'))
self.checkError([opt], 'expected one argument')
self.checkError([opt, '6'],
'needs 2 or 3 colon-separated arguments')
self.checkError([opt, 'foo:'], 'invalid huntrleaks value')
self.checkError([opt, '6:foo'], 'invalid huntrleaks value')
def test_multiprocess(self):
for opt in '-j', '--multiprocess':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt, '2'])
self.assertEqual(ns.use_mp, 2)
self.checkError([opt], 'expected one argument')
self.checkError([opt, 'foo'], 'invalid int value')
self.checkError([opt, '2', '-T'], "don't go together")
self.checkError([opt, '0', '-T'], "don't go together")
def test_coverage(self):
for opt in '-T', '--coverage':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt])
self.assertTrue(ns.trace)
def test_coverdir(self):
for opt in '-D', '--coverdir':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt, 'foo'])
self.assertEqual(ns.coverdir,
os.path.join(os_helper.SAVEDCWD, 'foo'))
self.checkError([opt], 'expected one argument')
def test_nocoverdir(self):
for opt in '-N', '--nocoverdir':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt])
self.assertIsNone(ns.coverdir)
def test_threshold(self):
for opt in '-t', '--threshold':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt, '1000'])
self.assertEqual(ns.threshold, 1000)
self.checkError([opt], 'expected one argument')
self.checkError([opt, 'foo'], 'invalid int value')
def test_nowindows(self):
for opt in '-n', '--nowindows':
with self.subTest(opt=opt):
with contextlib.redirect_stderr(io.StringIO()) as stderr:
ns = libregrtest._parse_args([opt])
self.assertTrue(ns.nowindows)
err = stderr.getvalue()
self.assertIn('the --nowindows (-n) option is deprecated', err)
def test_forever(self):
for opt in '-F', '--forever':
with self.subTest(opt=opt):
ns = libregrtest._parse_args([opt])
self.assertTrue(ns.forever)
def test_unrecognized_argument(self):
self.checkError(['--xxx'], 'usage:')
def test_long_option__partial(self):
ns = libregrtest._parse_args(['--qui'])
self.assertTrue(ns.quiet)
self.assertEqual(ns.verbose, 0)
def test_two_options(self):
ns = libregrtest._parse_args(['--quiet', '--exclude'])
self.assertTrue(ns.quiet)
self.assertEqual(ns.verbose, 0)
self.assertTrue(ns.exclude)
def test_option_with_empty_string_value(self):
ns = libregrtest._parse_args(['--start', ''])
self.assertEqual(ns.start, '')
def test_arg(self):
ns = libregrtest._parse_args(['foo'])
self.assertEqual(ns.args, ['foo'])
def test_option_and_arg(self):
ns = libregrtest._parse_args(['--quiet', 'foo'])
self.assertTrue(ns.quiet)
self.assertEqual(ns.verbose, 0)
self.assertEqual(ns.args, ['foo'])
def test_arg_option_arg(self):
ns = libregrtest._parse_args(['test_unaryop', '-v', 'test_binop'])
self.assertEqual(ns.verbose, 1)
self.assertEqual(ns.args, ['test_unaryop', 'test_binop'])
def test_unknown_option(self):
self.checkError(['--unknown-option'],
'unrecognized arguments: --unknown-option')
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(os_helper.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
class Tests(unittest.TestCase):
def test_empty_test(self):
pass
""")
# 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(os_helper.unlink, path)
# Use 'x' mode to ensure that we do not override existing tests
try:
with open(path, 'x', encoding='utf-8') as fp:
fp.write(code)
except PermissionError as exc:
if not sysconfig.is_python_build():
self.skipTest("cannot write %s: %s" % (path, exc))
raise
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
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
def check_line(self, output, regex, full=False):
if full:
regex += '\n'
regex = re.compile(r'^' + regex, re.MULTILINE)
self.assertRegex(output, regex)
def parse_executed_tests(self, output):
regex = (r'^%s\[ *[0-9]+(?:/ *[0-9]+)*\] (%s)'
% (LOG_PREFIX, 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=(),
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
rerun={}, run_no_tests=(),
resource_denied=(),
randomize=False, interrupted=False,
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
fail_env_changed=False,
*, stats):
if isinstance(tests, str):
tests = [tests]
if isinstance(skipped, str):
skipped = [skipped]
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
if isinstance(resource_denied, str):
resource_denied = [resource_denied]
if isinstance(failed, str):
failed = [failed]
if isinstance(env_changed, str):
env_changed = [env_changed]
if isinstance(omitted, str):
omitted = [omitted]
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
if isinstance(run_no_tests, str):
run_no_tests = [run_no_tests]
if isinstance(stats, int):
stats = TestStats(stats)
executed = self.parse_executed_tests(output)
if randomize:
self.assertEqual(set(executed), set(tests), output)
else:
self.assertEqual(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.keys())
self.check_line(output, regex)
regex = LOG_PREFIX + r"Re-running failed tests in verbose mode"
self.check_line(output, regex)
for name, match in rerun.items():
regex = LOG_PREFIX + f"Re-running {name} in verbose mode \\(matching: {match}\\)"
self.check_line(output, regex)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
if run_no_tests:
regex = list_regex('%s test%s run no tests', run_no_tests)
self.check_line(output, regex)
good = (len(tests) - len(skipped) - len(failed)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
- len(omitted) - len(env_changed) - len(run_no_tests))
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.')
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
# Total tests
parts = [f'run={stats.tests_run:,}']
if stats.failures:
parts.append(f'failures={stats.failures:,}')
if stats.skipped:
parts.append(f'skipped={stats.skipped:,}')
line = fr'Total tests: {" ".join(parts)}'
self.check_line(output, line, full=True)
# Total test files
report = [f'success={good}']
if failed:
report.append(f'failed={len(failed)}')
if env_changed:
report.append(f'env_changed={len(env_changed)}')
if skipped:
report.append(f'skipped={len(skipped)}')
if resource_denied:
report.append(f'resource_denied={len(resource_denied)}')
if rerun:
report.append(f'rerun={len(rerun)}')
if run_no_tests:
report.append(f'run_no_tests={len(run_no_tests)}')
line = fr'Total test files: {" ".join(report)}'
self.check_line(output, line, full=True)
# Result
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)):
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
result.append("NO TESTS RAN")
elif not result:
result.append('SUCCESS')
result = ', '.join(result)
if rerun:
result = 'FAILURE then %s' % result
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
self.check_line(output, f'Result: {result}', full=True)
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.STDOUT
proc = subprocess.run(args,
universal_newlines=True,
input=input,
stdout=subprocess.PIPE,
**kw)
if proc.returncode != exitcode:
msg = ("Command %s failed with exit code %s\n"
"\n"
"stdout:\n"
"---\n"
"%s\n"
"---\n"
% (str(args), proc.returncode, proc.stdout))
if proc.stderr:
msg += ("\n"
"stderr:\n"
"---\n"
"%s"
"---\n"
% proc.stderr)
self.fail(msg)
return proc
def run_python(self, args, **kw):
args = [sys.executable, '-X', 'faulthandler', '-I', *args]
proc = self.run_command(args, **kw)
return proc.stdout
class CheckActualTests(BaseTestCase):
def test_finds_expected_number_of_tests(self):
"""
Check that regrtest appears to find the expected set of tests.
"""
args = ['-Wd', '-E', '-bb', '-m', 'test.regrtest', '--list-tests']
output = self.run_python(args)
rough_number_of_tests_found = len(output.splitlines())
actual_testsuite_glob = os.path.join(glob.escape(os.path.dirname(__file__)),
'test*.py')
rough_counted_test_py_files = len(glob.glob(actual_testsuite_glob))
# We're not trying to duplicate test finding logic in here,
# just give a rough estimate of how many there should be and
# be near that. This is a regression test to prevent mishaps
# such as https://bugs.python.org/issue37667 in the future.
# If you need to change the values in here during some
# mythical future test suite reorganization, don't go
# overboard with logic and keep that goal in mind.
self.assertGreater(rough_number_of_tests_found,
rough_counted_test_py_files*9//10,
msg='Unexpectedly low number of tests found in:\n'
f'{", ".join(output.splitlines())}')
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().setUp()
# Create NTEST tests doing nothing
self.tests = [self.create_test() for index in range(self.NTEST)]
self.python_args = ['-Wd', '-E', '-bb']
self.regrtest_args = ['-uall', '-rwW',
'--testdir=%s' % self.tmptestdir]
self.regrtest_args.extend(('--timeout', '3600', '-j4'))
if sys.platform == 'win32':
self.regrtest_args.append('-n')
def check_output(self, output):
self.parse_random_seed(output)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
self.check_executed_tests(output, self.tests,
randomize=True, stats=len(self.tests))
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)
@unittest.skipUnless(sysconfig.is_python_build(),
'run_tests.py script is not installed')
def test_tools_script_run_tests(self):
# Tools/scripts/run_tests.py
script = os.path.join(ROOT_DIR, 'Tools', 'scripts', 'run_tests.py')
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)
@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):
# Tools\buildbot\test.bat
script = os.path.join(ROOT_DIR, 'Tools', 'buildbot', 'test.bat')
test_args = ['--testdir=%s' % self.tmptestdir]
if platform.machine() == 'ARM64':
test_args.append('-arm64') # ARM 64-bit build
elif platform.machine() == 'ARM':
test_args.append('-arm32') # 32-bit ARM build
elif platform.architecture()[0] == '64bit':
test_args.append('-x64') # 64-bit build
if not Py_DEBUG:
test_args.append('+d') # Release build, use python.exe
self.run_batch(script, *test_args, *self.tests)
@unittest.skipUnless(sys.platform == 'win32', 'Windows only')
def test_pcbuild_rt(self):
# PCbuild\rt.bat
script = os.path.join(ROOT_DIR, r'PCbuild\rt.bat')
if not os.path.isfile(script):
self.skipTest(f'File "{script}" does not exist')
rt_args = ["-q"] # Quick, don't run tests twice
if platform.machine() == 'ARM64':
rt_args.append('-arm64') # ARM 64-bit build
elif platform.machine() == 'ARM':
rt_args.append('-arm32') # 32-bit ARM build
elif platform.architecture()[0] == '64bit':
rt_args.append('-x64') # 64-bit build
if Py_DEBUG:
rt_args.append('-d') # Debug build, use python_d.exe
self.run_batch(script, *rt_args, *self.regrtest_args, *self.tests)
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
class FailingTest(unittest.TestCase):
def test_failing(self):
self.fail("bug")
""")
test_ok = self.create_test('ok')
test_failing = self.create_test('failing', code=code)
tests = [test_ok, test_failing]
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests(*tests, exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, tests, failed=test_failing,
stats=TestStats(2, 1))
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
""" % 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)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
self.check_executed_tests(output, test_names, stats=2)
# -u audio: 1 resource enabled
output = self.run_tests('-uaudio', *test_names)
self.check_executed_tests(output, test_names,
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
skipped=tests['network'],
resource_denied=tests['network'],
stats=1)
# no option: 0 resources enabled
output = self.run_tests(*test_names)
self.check_executed_tests(output, test_names,
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
skipped=test_names,
resource_denied=test_names,
stats=0)
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
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests('-r', test, exitcode=EXITCODE_NO_TESTS_RAN)
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
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests('-r', '--randseed=%s' % randseed, test,
exitcode=EXITCODE_NO_TESTS_RAN)
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 = os_helper.TESTFN
self.addCleanup(os_helper.unlink, filename)
# test format '0:00:00 [2/7] test_opcodes -- test_grammar took 0 sec'
with open(filename, "w") as fp:
previous = None
for index, name in enumerate(tests, 1):
line = ("00:00:%02i [%s/%s] %s"
% (index, index, len(tests), name))
if previous:
line += " -- %s took 0 sec" % previous
print(line, file=fp)
previous = name
output = self.run_tests('--fromfile', filename)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
stats = len(tests)
self.check_executed_tests(output, tests, stats=stats)
# test format '[2/7] test_opcodes'
with open(filename, "w") as fp:
for index, name in enumerate(tests, 1):
print("[%s/%s] %s" % (index, len(tests), name), file=fp)
output = self.run_tests('--fromfile', filename)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
self.check_executed_tests(output, tests, stats=stats)
# test format 'test_opcodes'
with open(filename, "w") as fp:
for name in tests:
print(name, file=fp)
output = self.run_tests('--fromfile', filename)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
self.check_executed_tests(output, tests, stats=stats)
# test format 'Lib/test/test_opcodes.py'
with open(filename, "w") as fp:
for name in tests:
print('Lib/test/%s.py' % name, file=fp)
output = self.run_tests('--fromfile', filename)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
self.check_executed_tests(output, tests, stats=stats)
def test_interrupted(self):
code = TEST_INTERRUPTED
test = self.create_test('sigint', code=code)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests(test, exitcode=EXITCODE_INTERRUPTED)
self.check_executed_tests(output, test, omitted=test,
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
interrupted=True, stats=0)
def test_slowest(self):
# test --slowest
tests = [self.create_test() for index in range(3)]
output = self.run_tests("--slowest", *tests)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
self.check_executed_tests(output, tests, stats=len(tests))
regex = ('10 slowest tests:\n'
'(?:- %s: .*\n){%s}'
% (self.TESTNAME_REGEX, len(tests)))
self.check_line(output, regex)
def test_slowest_interrupted(self):
# Issue #25373: test --slowest with an interrupted test
code = TEST_INTERRUPTED
test = self.create_test("sigint", code=code)
for multiprocessing in (False, True):
with self.subTest(multiprocessing=multiprocessing):
if multiprocessing:
args = ("--slowest", "-j2", test)
else:
args = ("--slowest", test)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests(*args, exitcode=EXITCODE_INTERRUPTED)
self.check_executed_tests(output, test,
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
omitted=test, interrupted=True,
stats=0)
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)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
self.check_executed_tests(output, [test], stats=1)
regex = (r'lines +cov% +module +\(path\)\n'
r'(?: *[0-9]+ *[0-9]{1,2}% *[^ ]+ +\([^)]+\)+)+')
self.check_line(output, regex)
def test_wait(self):
# test --wait
test = self.create_test('wait')
output = self.run_tests("--wait", test, input='key')
self.check_line(output, 'Press any key to continue')
def test_forever(self):
# test --forever
code = textwrap.dedent("""
import builtins
import unittest
class ForeverTester(unittest.TestCase):
def test_run(self):
# Store the state in the builtins module, because the test
# module is reload at each run
if 'RUN' in builtins.__dict__:
builtins.__dict__['RUN'] += 1
if builtins.__dict__['RUN'] >= 3:
self.fail("fail at the 3rd runs")
else:
builtins.__dict__['RUN'] = 1
""")
test = self.create_test('forever', code=code)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests('--forever', test, exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, [test]*3, failed=test,
stats=TestStats(1, 1))
def check_leak(self, code, what):
test = self.create_test('huntrleaks', code=code)
filename = 'reflog.txt'
self.addCleanup(os_helper.unlink, filename)
output = self.run_tests('--huntrleaks', '3:3:', test,
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
exitcode=EXITCODE_BAD_TEST,
stderr=subprocess.STDOUT)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
self.check_executed_tests(output, [test], failed=test, stats=1)
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')
def test_huntrleaks(self):
# test --huntrleaks
code = textwrap.dedent("""
import unittest
GLOBAL_LIST = []
class RefLeakTest(unittest.TestCase):
def test_leak(self):
GLOBAL_LIST.append(object())
""")
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
class FDLeakTest(unittest.TestCase):
def test_leak(self):
fd = os.open(__file__, os.O_RDONLY)
# bug: never close the file descriptor
""")
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)
@support.cpython_only
def test_crashed(self):
# Any code which causes a crash
code = 'import faulthandler; faulthandler._sigsegv()'
crash_test = self.create_test(name="crash", code=code)
tests = [crash_test]
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests("-j2", *tests, exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, tests, failed=crash_test,
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
randomize=True, stats=0)
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_ignorefile(self):
code = textwrap.dedent("""
import unittest
class Tests(unittest.TestCase):
def test_method1(self):
pass
def test_method2(self):
pass
def test_method3(self):
pass
def test_method4(self):
pass
""")
all_methods = ['test_method1', 'test_method2',
'test_method3', 'test_method4']
testname = self.create_test(code=code)
# only run a subset
filename = os_helper.TESTFN
self.addCleanup(os_helper.unlink, filename)
subset = [
# only ignore the method name
'test_method1',
# ignore 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", "--ignorefile", filename, testname)
methods = self.parse_methods(output)
subset = ['test_method2', 'test_method4']
self.assertEqual(methods, subset)
def test_matchfile(self):
code = textwrap.dedent("""
import unittest
class Tests(unittest.TestCase):
def test_method1(self):
pass
def test_method2(self):
pass
def test_method3(self):
pass
def test_method4(self):
pass
""")
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 = os_helper.TESTFN
self.addCleanup(os_helper.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)
def test_env_changed(self):
code = textwrap.dedent("""
import unittest
class Tests(unittest.TestCase):
def test_env_changed(self):
open("env_changed", "w").close()
""")
testname = self.create_test(code=code)
# don't fail by default
output = self.run_tests(testname)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
self.check_executed_tests(output, [testname],
env_changed=testname, stats=1)
# fail with --fail-env-changed
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests("--fail-env-changed", testname,
exitcode=EXITCODE_ENV_CHANGED)
self.check_executed_tests(output, [testname], env_changed=testname,
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
fail_env_changed=True, stats=1)
def test_rerun_fail(self):
# FAILURE then FAILURE
code = textwrap.dedent("""
import unittest
class Tests(unittest.TestCase):
def test_succeed(self):
return
def test_fail_always(self):
# test that always fails
self.fail("bug")
""")
testname = self.create_test(code=code)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests("-w", testname, exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, [testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
failed=testname,
rerun={testname: "test_fail_always"},
stats=TestStats(1, 1))
def test_rerun_success(self):
# FAILURE then SUCCESS
code = textwrap.dedent("""
import builtins
import unittest
class Tests(unittest.TestCase):
def test_succeed(self):
return
def test_fail_once(self):
if not hasattr(builtins, '_test_failed'):
builtins._test_failed = True
self.fail("bug")
""")
testname = self.create_test(code=code)
output = self.run_tests("-w", testname, exitcode=0)
self.check_executed_tests(output, [testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
rerun={testname: "test_fail_once"},
stats=1)
def test_rerun_setup_class_hook_failure(self):
# FAILURE then FAILURE
code = textwrap.dedent("""
import unittest
class ExampleTests(unittest.TestCase):
@classmethod
def setUpClass(self):
raise RuntimeError('Fail')
def test_success(self):
return
""")
testname = self.create_test(code=code)
output = self.run_tests("-w", testname, exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, testname,
failed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
rerun={testname: "ExampleTests"},
stats=0)
def test_rerun_teardown_class_hook_failure(self):
# FAILURE then FAILURE
code = textwrap.dedent("""
import unittest
class ExampleTests(unittest.TestCase):
@classmethod
def tearDownClass(self):
raise RuntimeError('Fail')
def test_success(self):
return
""")
testname = self.create_test(code=code)
output = self.run_tests("-w", testname, exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, testname,
failed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
rerun={testname: "ExampleTests"},
stats=1)
def test_rerun_setup_module_hook_failure(self):
# FAILURE then FAILURE
code = textwrap.dedent("""
import unittest
def setUpModule():
raise RuntimeError('Fail')
class ExampleTests(unittest.TestCase):
def test_success(self):
return
""")
testname = self.create_test(code=code)
output = self.run_tests("-w", testname, exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, testname,
failed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
rerun={testname: testname},
stats=0)
def test_rerun_teardown_module_hook_failure(self):
# FAILURE then FAILURE
code = textwrap.dedent("""
import unittest
def tearDownModule():
raise RuntimeError('Fail')
class ExampleTests(unittest.TestCase):
def test_success(self):
return
""")
testname = self.create_test(code=code)
output = self.run_tests("-w", testname, exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, testname,
failed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
rerun={testname: testname},
stats=1)
def test_rerun_setup_hook_failure(self):
# FAILURE then FAILURE
code = textwrap.dedent("""
import unittest
class ExampleTests(unittest.TestCase):
def setUp(self):
raise RuntimeError('Fail')
def test_success(self):
return
""")
testname = self.create_test(code=code)
output = self.run_tests("-w", testname, exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, testname,
failed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
rerun={testname: "test_success"},
stats=1)
def test_rerun_teardown_hook_failure(self):
# FAILURE then FAILURE
code = textwrap.dedent("""
import unittest
class ExampleTests(unittest.TestCase):
def tearDown(self):
raise RuntimeError('Fail')
def test_success(self):
return
""")
testname = self.create_test(code=code)
output = self.run_tests("-w", testname, exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, testname,
failed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
rerun={testname: "test_success"},
stats=1)
def test_rerun_async_setup_hook_failure(self):
# FAILURE then FAILURE
code = textwrap.dedent("""
import unittest
class ExampleTests(unittest.IsolatedAsyncioTestCase):
async def asyncSetUp(self):
raise RuntimeError('Fail')
async def test_success(self):
return
""")
testname = self.create_test(code=code)
output = self.run_tests("-w", testname, exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, testname,
failed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
rerun={testname: "test_success"},
stats=1)
def test_rerun_async_teardown_hook_failure(self):
# FAILURE then FAILURE
code = textwrap.dedent("""
import unittest
class ExampleTests(unittest.IsolatedAsyncioTestCase):
async def asyncTearDown(self):
raise RuntimeError('Fail')
async def test_success(self):
return
""")
testname = self.create_test(code=code)
output = self.run_tests("-w", testname, exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, testname,
failed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
rerun={testname: "test_success"},
stats=1)
def test_no_tests_ran(self):
code = textwrap.dedent("""
import unittest
class Tests(unittest.TestCase):
def test_bug(self):
pass
""")
testname = self.create_test(code=code)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests(testname, "-m", "nosuchtest",
exitcode=EXITCODE_NO_TESTS_RAN)
self.check_executed_tests(output, [testname],
run_no_tests=testname,
stats=0)
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)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests(testname)
self.check_executed_tests(output, [testname],
stats=TestStats(1, skipped=1))
def test_no_tests_ran_multiple_tests_nonexistent(self):
code = textwrap.dedent("""
import unittest
class Tests(unittest.TestCase):
def test_bug(self):
pass
""")
testname = self.create_test(code=code)
testname2 = self.create_test(code=code)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests(testname, testname2, "-m", "nosuchtest",
exitcode=EXITCODE_NO_TESTS_RAN)
self.check_executed_tests(output, [testname, testname2],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
run_no_tests=[testname, testname2],
stats=0)
def test_no_test_ran_some_test_exist_some_not(self):
code = textwrap.dedent("""
import unittest
class Tests(unittest.TestCase):
def test_bug(self):
pass
""")
testname = self.create_test(code=code)
other_code = textwrap.dedent("""
import unittest
class Tests(unittest.TestCase):
def test_other_bug(self):
pass
""")
testname2 = self.create_test(code=other_code)
output = self.run_tests(testname, testname2, "-m", "nosuchtest",
"-m", "test_other_bug", exitcode=0)
self.check_executed_tests(output, [testname, testname2],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
run_no_tests=[testname],
stats=1)
@support.cpython_only
def test_uncollectable(self):
code = textwrap.dedent(r"""
import _testcapi
import gc
import unittest
@_testcapi.with_tp_del
class Garbage:
def __tp_del__(self):
pass
class Tests(unittest.TestCase):
def test_garbage(self):
# create an uncollectable object
obj = Garbage()
obj.ref_cycle = obj
obj = None
""")
testname = self.create_test(code=code)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests("--fail-env-changed", testname,
exitcode=EXITCODE_ENV_CHANGED)
self.check_executed_tests(output, [testname],
env_changed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
fail_env_changed=True,
stats=1)
def test_multiprocessing_timeout(self):
code = textwrap.dedent(r"""
import time
import unittest
try:
import faulthandler
except ImportError:
faulthandler = None
class Tests(unittest.TestCase):
# test hangs and so should be stopped by the timeout
def test_sleep(self):
# we want to test regrtest multiprocessing timeout,
# not faulthandler timeout
if faulthandler is not None:
faulthandler.cancel_dump_traceback_later()
time.sleep(60 * 5)
""")
testname = self.create_test(code=code)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests("-j2", "--timeout=1.0", testname,
exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, [testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
failed=testname, stats=0)
self.assertRegex(output,
re.compile('%s timed out' % testname, re.MULTILINE))
def test_unraisable_exc(self):
# --fail-env-changed must catch unraisable exception.
# The exception must be displayed even if sys.stderr is redirected.
code = textwrap.dedent(r"""
import unittest
import weakref
from test.support import captured_stderr
class MyObject:
pass
def weakref_callback(obj):
raise Exception("weakref callback bug")
class Tests(unittest.TestCase):
def test_unraisable_exc(self):
obj = MyObject()
ref = weakref.ref(obj, weakref_callback)
with captured_stderr() as stderr:
# call weakref_callback() which logs
# an unraisable exception
obj = None
self.assertEqual(stderr.getvalue(), '')
""")
testname = self.create_test(code=code)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests("--fail-env-changed", "-v", testname,
exitcode=EXITCODE_ENV_CHANGED)
self.check_executed_tests(output, [testname],
env_changed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
fail_env_changed=True,
stats=1)
self.assertIn("Warning -- Unraisable exception", output)
self.assertIn("Exception: weakref callback bug", output)
def test_threading_excepthook(self):
# --fail-env-changed must catch uncaught thread exception.
# The exception must be displayed even if sys.stderr is redirected.
code = textwrap.dedent(r"""
import threading
import unittest
from test.support import captured_stderr
class MyObject:
pass
def func_bug():
raise Exception("bug in thread")
class Tests(unittest.TestCase):
def test_threading_excepthook(self):
with captured_stderr() as stderr:
thread = threading.Thread(target=func_bug)
thread.start()
thread.join()
self.assertEqual(stderr.getvalue(), '')
""")
testname = self.create_test(code=code)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests("--fail-env-changed", "-v", testname,
exitcode=EXITCODE_ENV_CHANGED)
self.check_executed_tests(output, [testname],
env_changed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
fail_env_changed=True,
stats=1)
self.assertIn("Warning -- Uncaught thread exception", output)
self.assertIn("Exception: bug in thread", output)
def test_print_warning(self):
# bpo-45410: The order of messages must be preserved when -W and
# support.print_warning() are used.
code = textwrap.dedent(r"""
import sys
import unittest
from test import support
class MyObject:
pass
def func_bug():
raise Exception("bug in thread")
class Tests(unittest.TestCase):
def test_print_warning(self):
print("msg1: stdout")
support.print_warning("msg2: print_warning")
# Fail with ENV CHANGED to see print_warning() log
support.environment_altered = True
""")
testname = self.create_test(code=code)
# Expect an output like:
#
# test_threading_excepthook (test.test_x.Tests) ... msg1: stdout
# Warning -- msg2: print_warning
# ok
regex = (r"test_print_warning.*msg1: stdout\n"
r"Warning -- msg2: print_warning\n"
r"ok\n")
for option in ("-v", "-W"):
with self.subTest(option=option):
cmd = ["--fail-env-changed", option, testname]
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
output = self.run_tests(*cmd, exitcode=EXITCODE_ENV_CHANGED)
self.check_executed_tests(output, [testname],
env_changed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
fail_env_changed=True,
stats=1)
self.assertRegex(output, regex)
def test_unicode_guard_env(self):
guard = os.environ.get(setup.UNICODE_GUARD_ENV)
self.assertIsNotNone(guard, f"{setup.UNICODE_GUARD_ENV} not set")
if guard.isascii():
# Skip to signify that the env var value was changed by the user;
# possibly to something ASCII to work around Unicode issues.
self.skipTest("Modified guard")
def test_cleanup(self):
dirname = os.path.join(self.tmptestdir, "test_python_123")
os.mkdir(dirname)
filename = os.path.join(self.tmptestdir, "test_python_456")
open(filename, "wb").close()
names = [dirname, filename]
cmdargs = ['-m', 'test',
'--tempdir=%s' % self.tmptestdir,
'--cleanup']
self.run_python(cmdargs)
for name in names:
self.assertFalse(os.path.exists(name), name)
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
@unittest.skipIf(support.is_wasi,
'checking temp files is not implemented on WASI')
def test_leak_tmp_file(self):
code = textwrap.dedent(r"""
import os.path
import tempfile
import unittest
class FileTests(unittest.TestCase):
def test_leak_tmp_file(self):
filename = os.path.join(tempfile.gettempdir(), 'mytmpfile')
with open(filename, "wb") as fp:
fp.write(b'content')
""")
testnames = [self.create_test(code=code) for _ in range(3)]
output = self.run_tests("--fail-env-changed", "-v", "-j2", *testnames,
exitcode=EXITCODE_ENV_CHANGED)
self.check_executed_tests(output, testnames,
env_changed=testnames,
fail_env_changed=True,
randomize=True,
stats=len(testnames))
for testname in testnames:
self.assertIn(f"Warning -- {testname} leaked temporary "
f"files (1): mytmpfile",
output)
def test_mp_decode_error(self):
# gh-101634: If a worker stdout cannot be decoded, report a failed test
# and a non-zero exit code.
if sys.platform == 'win32':
encoding = locale.getencoding()
else:
encoding = sys.stdout.encoding
if encoding is None:
encoding = sys.__stdout__.encoding
if encoding is None:
self.skipTest(f"cannot get regrtest worker encoding")
nonascii = b"byte:\xa0\xa9\xff\n"
try:
nonascii.decode(encoding)
except UnicodeDecodeError:
pass
else:
self.skipTest(f"{encoding} can decode non-ASCII bytes {nonascii!a}")
code = textwrap.dedent(fr"""
import sys
# bytes which cannot be decoded from UTF-8
nonascii = {nonascii!a}
sys.stdout.buffer.write(nonascii)
sys.stdout.buffer.flush()
""")
testname = self.create_test(code=code)
output = self.run_tests("--fail-env-changed", "-v", "-j1", testname,
exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, [testname],
failed=[testname],
[3.11] gh-108822: Backport libregrtest changes from the main branch (#108820) * Revert "[3.11] gh-101634: regrtest reports decoding error as failed test (#106169) (#106175)" This reverts commit d5418e97fc524420011a370ba3c2c3cf6a89a74f. * Revert "[3.11] bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (GH-30895) (GH-103342)" This reverts commit ecb09a849689764193e0115d27e220f82b5f6d9f. * Revert "gh-95027: Fix regrtest stdout encoding on Windows (GH-98492)" This reverts commit b2aa28eec56d07b9c6777b02b7247cf21839de9f. * Revert "[3.11] gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) (GH-94408)" This reverts commit 0122ab235b5acb52dd99fd05d8802a00f438b828. * Revert "Run Tools/scripts/reindent.py (GH-94225)" This reverts commit f0f3a424afb00a15ce8c0140dd218f5b33929be6. * Revert "gh-94052: Don't re-run failed tests with --python option (GH-94054)" This reverts commit 1347607db12012f6458ffcba48d8ad797083812e. * Revert "[3.11] gh-84461: Fix Emscripten umask and permission issues (GH-94002) (GH-94006)" This reverts commit 10731849184a3101ed18683b0128d689f1671c3f. * gh-93353: regrtest checks for leaked temporary files (#93776) When running tests with -jN, create a temporary directory per process and mark a test as "environment changed" if a test leaks a temporary file or directory. (cherry picked from commit e566ce5496f1bad81c431aaee65e36d5e44771c5) * gh-93353: Fix regrtest for -jN with N >= 2 (GH-93813) (cherry picked from commit 36934a16e86f34d69ba2d41630fb5b4d06d59cff) * gh-93353: regrtest supports checking tmp files with -j2 (#93909) regrtest now also implements checking for leaked temporary files and directories when using -jN for N >= 2. Use tempfile.mkdtemp() to create the temporary directory. Skip this check on WASI. (cherry picked from commit 4f85cec9e2077681b3dacc3108e646d509b720bf) * gh-84461: Fix Emscripten umask and permission issues (GH-94002) - Emscripten's default umask is too strict, see https://github.com/emscripten-core/emscripten/issues/17269 - getuid/getgid and geteuid/getegid are stubs that always return 0 (root). Disable effective uid/gid syscalls and fix tests that use chmod() current user. - Cannot drop X bit from directory. (cherry picked from commit 2702e408fd0e0dd7aec396b4cf8c7ce9caae81d8) * gh-94052: Don't re-run failed tests with --python option (#94054) (cherry picked from commit 0ff7b996f5d836e63cdaf652c7aa734285261096) * Run Tools/scripts/reindent.py (#94225) Reindent files which were not properly formatted (PEP 8: 4 spaces). Remove also some trailing spaces. (cherry picked from commit e87ada48a9e5d9d03f9759138869216df0d7383a) * gh-94026: Buffer regrtest worker stdout in temporary file (GH-94253) Co-authored-by: Victor Stinner <vstinner@python.org> (cherry picked from commit 199ba233248ab279f445e0809c2077976f0711bc) * gh-96465: Clear fractions hash lru_cache under refleak testing (GH-96689) Automerge-Triggered-By: GH:zware (cherry picked from commit 9c8f3794337457b1d905a9fa0f38c2978fe32abd) * gh-95027: Fix regrtest stdout encoding on Windows (#98492) On Windows, when the Python test suite is run with the -jN option, the ANSI code page is now used as the encoding for the stdout temporary file, rather than using UTF-8 which can lead to decoding errors. (cherry picked from commit ec1f6f5f139868dc2c1116a7c7c878c38c668d53) * gh-98903: Test suite fails with exit code 4 if no tests ran (#98904) The Python test suite now fails wit exit code 4 if no tests ran. It should help detecting typos in test names and test methods. * Add "EXITCODE_" constants to Lib/test/libregrtest/main.py. * Fix a typo: "NO TEST RUN" becomes "NO TESTS RAN" (cherry picked from commit c76db37c0d23174cbffd6fa978d39693890ef020) * gh-100086: Add build info to test.libregrtest (#100093) The Python test runner (libregrtest) now logs Python build information like "debug" vs "release" build, or LTO and PGO optimizations. (cherry picked from commit 3c892022472eb975360fb3f0caa6f6fcc6fbf220) * bpo-46523: fix tests rerun when `setUp[Class|Module]` fails (#30895) Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com> Co-authored-by: Ɓukasz Langa <lukasz@langa.pl> (cherry picked from commit 995386071f96e4cfebfa027a71ca9134e4651d2a) * gh-82054: allow test runner to split test_asyncio to execute in parallel by sharding. (#103927) This runs test_asyncio sub-tests in parallel using sharding from Cinder. This suite is typically the longest-pole in runs because it is a test package with a lot of further sub-tests otherwise run serially. By breaking out the sub-tests as independent modules we can run a lot more in parallel. After porting we can see the direct impact on a multicore system. Without this change: Running make test is 5 min 26 seconds With this change: Running make test takes 3 min 39 seconds That'll vary based on system and parallelism. On a `-j 4` run similar to what CI and buildbot systems often do, it reduced the overall test suite completion latency by 10%. The drawbacks are that this implementation is hacky and due to the sorting of the tests it obscures when the asyncio tests occur and involves changing CPython test infrastructure but, the wall time saved it is worth it, especially in low-core count CI runs as it pulls a long tail. The win for productivity and reserved CI resource usage is significant. Future tests that deserve to be refactored into split up suites to benefit from are test_concurrent_futures and the way the _test_multiprocessing suite gets run for all start methods. As exposed by passing the -o flag to python -m test to get a list of the 10 longest running tests. --------- Co-authored-by: Carl Meyer <carl@oddbird.net> Co-authored-by: Gregory P. Smith <greg@krypto.org> [Google, LLC] (cherry picked from commit 9e011e7c77dad7d0bbb944c44891531606caeb21) * Display the sanitizer config in the regrtest header. (#105301) Display the sanitizers present in libregrtest. Having this in the CI output for tests with the relevant environment variable displayed will help make it easier to do what we need to create an equivalent local test run. (cherry picked from commit 852348ab65783601e0844b6647ea033668b45c11) * gh-101634: regrtest reports decoding error as failed test (#106169) When running the Python test suite with -jN option, if a worker stdout cannot be decoded from the locale encoding report a failed testn so the exitcode is non-zero. (cherry picked from commit 2ac3eec103cf450aaaebeb932e51155d2e7fb37b) * gh-108223: test.pythoninfo and libregrtest log Py_NOGIL (#108238) Enable with --disable-gil --without-pydebug: $ make pythoninfo|grep NOGIL sysconfig[Py_NOGIL]: 1 $ ./python -m test ... == Python build: nogil debug ... (cherry picked from commit 5afe0c17ca14df430736e549542a4b85e7e7c7ac) * gh-90791: test.pythoninfo logs ASAN_OPTIONS env var (#108289) * Cleanup libregrtest code logging ASAN_OPTIONS. * Fix a typo on "ASAN_OPTIONS" vs "MSAN_OPTIONS". (cherry picked from commit 3a1ac87f8f89d3206b46a0df4908afae629d669d) * gh-108388: regrtest splits test_asyncio package (#108393) Currently, test_asyncio package is only splitted into sub-tests when using command "./python -m test". With this change, it's also splitted when passing it on the command line: "./python -m test test_asyncio". Remove the concept of "STDTESTS". Python is now mature enough to not have to bother with that anymore. Removing STDTESTS simplify the code. (cherry picked from commit 174e9da0836844a2138cc8915dd305cb2cd7a583) * regrtest computes statistics (#108793) test_netrc, test_pep646_syntax and test_xml_etree now return results in the test_main() function. Changes: * Rewrite TestResult as a dataclass with a new State class. * Add test.support.TestStats class and Regrtest.stats_dict attribute. * libregrtest.runtest functions now modify a TestResult instance in-place. * libregrtest summary lists the number of run tests and skipped tests, and denied resources. * Add TestResult.has_meaningful_duration() method. * Compute TestResult duration in the upper function. * Use time.perf_counter() instead of time.monotonic(). * Regrtest: rename 'resource_denieds' attribute to 'resource_denied'. * Rename CHILD_ERROR to MULTIPROCESSING_ERROR. * Use match/case syntadx to have different code depending on the test state. Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com> (cherry picked from commit d4e534cbb35678c82b3a1276826af55d7bfc23b6) * gh-108822: Add Changelog entry for regrtest statistics (#108821) --------- Co-authored-by: Christian Heimes <christian@python.org> Co-authored-by: Zachary Ware <zach@python.org> Co-authored-by: Nikita Sobolev <mail@sobolevn.me> Co-authored-by: Joshua Herman <zitterbewegung@gmail.com> Co-authored-by: Gregory P. Smith <greg@krypto.org>
2023-09-03 19:21:53 +02:00
randomize=True,
stats=0)
def test_doctest(self):
code = textwrap.dedent(fr'''
import doctest
import sys
from test import support
def my_function():
"""
Pass:
>>> 1 + 1
2
Failure:
>>> 2 + 3
23
>>> 1 + 1
11
Skipped test (ignored):
>>> id(1.0) # doctest: +SKIP
7948648
"""
def test_main():
testmod = sys.modules[__name__]
return support.run_doctest(testmod)
''')
testname = self.create_test(code=code)
output = self.run_tests("--fail-env-changed", "-v", "-j1", testname,
exitcode=EXITCODE_BAD_TEST)
self.check_executed_tests(output, [testname],
failed=[testname],
randomize=True,
stats=TestStats(3, 2, 0))
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),
'1.5 sec')
self.assertEqual(utils.format_duration(1),
'1.0 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')
if __name__ == '__main__':
unittest.main()