cpython/Lib/test/libregrtest/main.py

867 lines
30 KiB
Python
Raw Normal View History

import faulthandler
2016-09-08 21:46:56 -07:00
import locale
import os
import platform
import random
import re
import sys
import sysconfig
import tempfile
import time
import unittest
from test.libregrtest.cmdline import _parse_args
from test.libregrtest.runtest import (
[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: Łukasz 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
findtests, split_test_packages, runtest, get_abs_module,
PROGRESS_MIN_TIME, State)
from test.libregrtest.setup import setup_tests
from test.libregrtest.pgo import setup_pgo_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: Łukasz 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.libregrtest.utils import (removepy, count, format_duration,
printlist, get_build_info)
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: Łukasz 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 TestStats
from test.support import os_helper
from test.support import threading_helper
# bpo-38203: Maximum delay in seconds to exit Python (call Py_Finalize()).
# Used to protect against threading._shutdown() hang.
# Must be smaller than buildbot "1200 seconds without output" limit.
EXIT_TIMEOUT = 120.0
# gh-90681: When rerunning tests, we might need to rerun the whole
# class or module suite if some its life-cycle hooks fail.
# Test level hooks are not affected.
_TEST_LIFECYCLE_HOOKS = frozenset((
'setUpClass', 'tearDownClass',
'setUpModule', 'tearDownModule',
))
EXITCODE_BAD_TEST = 2
EXITCODE_INTERRUPTED = 130
EXITCODE_ENV_CHANGED = 3
EXITCODE_NO_TESTS_RAN = 4
class Regrtest:
"""Execute a test suite.
This also parses command-line options and modifies its behavior
accordingly.
tests -- a list of strings containing test names (optional)
testdir -- the directory in which to look for tests (optional)
Users other than the Python test suite will certainly want to
specify testdir; if it's omitted, the directory containing the
Python test suite is searched for.
If the tests argument is omitted, the tests listed on the
command-line will be used. If that's empty, too, then all *.py
files beginning with test_ will be used.
The other default arguments (verbose, quiet, exclude,
single, randomize, use_resources, trace, coverdir,
print_slow, and random_seed) allow programmers calling main()
directly to set the values that would normally be set by flags
on the command line.
"""
def __init__(self):
# Namespace of command line options
self.ns = None
# tests
self.tests = []
self.selected = []
# test results
self.good = []
self.bad = []
self.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: Łukasz 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.resource_denied = []
self.environment_changed = []
self.run_no_tests = []
self.need_rerun = []
self.rerun = []
self.first_result = None
self.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: Łukasz 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.stats_dict: dict[str, TestStats] = {}
# used by --slow
self.test_times = []
# used by --coverage, trace.Trace instance
self.tracer = None
# used to display the progress bar "[ 3/100]"
[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: Łukasz 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.start_time = time.perf_counter()
self.test_count = ''
self.test_count_width = 1
# used by --single
self.next_single_test = None
self.next_single_filename = None
# used by --junit-xml
self.testsuite_xml = None
# misc
self.win_load_tracker = None
self.tmp_dir = None
self.worker_test_name = None
def get_executed(self):
return (set(self.good) | set(self.bad) | set(self.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: Łukasz 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
| set(self.resource_denied) | set(self.environment_changed)
| set(self.run_no_tests))
def accumulate_result(self, result, rerun=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: Łukasz 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
test_name = result.test_name
if result.has_meaningful_duration() and not rerun:
self.test_times.append((result.duration, test_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: Łukasz 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
match result.state:
case State.PASSED:
self.good.append(test_name)
case State.ENV_CHANGED:
self.environment_changed.append(test_name)
case State.SKIPPED:
self.skipped.append(test_name)
case State.RESOURCE_DENIED:
self.skipped.append(test_name)
self.resource_denied.append(test_name)
case State.INTERRUPTED:
self.interrupted = True
case State.DID_NOT_RUN:
self.run_no_tests.append(test_name)
case _:
if result.is_failed(self.ns.fail_env_changed):
if not rerun:
self.bad.append(test_name)
self.need_rerun.append(result)
else:
raise ValueError(f"invalid test state: {state!r}")
if result.stats is not None:
self.stats_dict[result.test_name] = result.stats
if rerun and not(result.is_failed(False) or result.state == State.INTERRUPTED):
self.bad.remove(test_name)
xml_data = result.xml_data
if xml_data:
import xml.etree.ElementTree as ET
for e in xml_data:
try:
self.testsuite_xml.append(ET.fromstring(e))
except ET.ParseError:
print(xml_data, file=sys.__stderr__)
raise
def log(self, line=''):
empty = not line
# add the system load prefix: "load avg: 1.80 "
load_avg = self.getloadavg()
if load_avg is not None:
line = f"load avg: {load_avg:.2f} {line}"
# add the timestamp prefix: "0:01:05 "
[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: Łukasz 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
test_time = time.perf_counter() - self.start_time
mins, secs = divmod(int(test_time), 60)
hours, mins = divmod(mins, 60)
test_time = "%d:%02d:%02d" % (hours, mins, secs)
line = f"{test_time} {line}"
if empty:
line = line[:-1]
print(line, flush=True)
def display_progress(self, test_index, text):
if self.ns.quiet:
return
# "[ 51/405/1] test_tcl passed"
line = f"{test_index:{self.test_count_width}}{self.test_count}"
fails = len(self.bad) + len(self.environment_changed)
if fails and not self.ns.pgo:
line = f"{line}/{fails}"
self.log(f"[{line}] {text}")
def parse_args(self, kwargs):
ns = _parse_args(sys.argv[1:], **kwargs)
if ns.xmlpath:
support.junit_xml_list = self.testsuite_xml = []
worker_args = ns.worker_args
if worker_args is not None:
from test.libregrtest.runtest_mp import parse_worker_args
ns, test_name = parse_worker_args(ns.worker_args)
ns.worker_args = worker_args
self.worker_test_name = test_name
# Strip .py extensions.
removepy(ns.args)
if ns.huntrleaks:
warmup, repetitions, _ = ns.huntrleaks
if warmup < 1 or repetitions < 1:
msg = ("Invalid values for the --huntrleaks/-R parameters. The "
"number of warmups and repetitions must be at least 1 "
"each (1:1).")
print(msg, file=sys.stderr, flush=True)
sys.exit(2)
if ns.tempdir:
ns.tempdir = os.path.expanduser(ns.tempdir)
self.ns = ns
def find_tests(self, tests):
self.tests = tests
if self.ns.single:
self.next_single_filename = os.path.join(self.tmp_dir, 'pynexttest')
try:
with open(self.next_single_filename, 'r') as fp:
next_test = fp.read().strip()
self.tests = [next_test]
except OSError:
pass
if self.ns.fromfile:
self.tests = []
# regex to match 'test_builtin' in line:
# '0:00:00 [ 4/400] test_builtin -- test_dict took 1 sec'
regex = re.compile(r'\btest_[a-zA-Z0-9_]+\b')
with open(os.path.join(os_helper.SAVEDCWD, self.ns.fromfile)) as fp:
for line in fp:
line = line.split('#', 1)[0]
line = line.strip()
match = regex.search(line)
if match is not None:
self.tests.append(match.group())
removepy(self.tests)
if self.ns.pgo:
# add default PGO tests if no tests are specified
setup_pgo_tests(self.ns)
[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: Łukasz 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
exclude = set()
if self.ns.exclude:
for arg in self.ns.args:
[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: Łukasz 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
exclude.add(arg)
self.ns.args = []
[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: Łukasz 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
alltests = findtests(testdir=self.ns.testdir, exclude=exclude)
if not self.ns.fromfile:
[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: Łukasz 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.selected = self.tests or self.ns.args
if self.selected:
self.selected = split_test_packages(self.selected)
else:
self.selected = alltests
else:
self.selected = self.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: Łukasz 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 self.ns.single:
self.selected = self.selected[:1]
try:
pos = alltests.index(self.selected[0])
self.next_single_test = alltests[pos + 1]
except IndexError:
pass
# Remove all the selected tests that precede start if it's set.
if self.ns.start:
try:
del self.selected[:self.selected.index(self.ns.start)]
except ValueError:
print("Couldn't find starting test (%s), using all tests"
% self.ns.start, file=sys.stderr)
if self.ns.randomize:
if self.ns.random_seed is None:
self.ns.random_seed = random.randrange(10000000)
random.seed(self.ns.random_seed)
random.shuffle(self.selected)
def list_tests(self):
for name in self.selected:
print(name)
def _list_cases(self, suite):
for test in suite:
if isinstance(test, unittest.loader._FailedTest):
continue
if isinstance(test, unittest.TestSuite):
self._list_cases(test)
elif isinstance(test, unittest.TestCase):
if support.match_test(test):
print(test.id())
def list_cases(self):
support.verbose = False
support.set_match_tests(self.ns.match_tests, self.ns.ignore_tests)
for test_name in self.selected:
abstest = get_abs_module(self.ns, test_name)
try:
suite = unittest.defaultTestLoader.loadTestsFromName(abstest)
self._list_cases(suite)
except unittest.SkipTest:
self.skipped.append(test_name)
if self.skipped:
print(file=sys.stderr)
print(count(len(self.skipped), "test"), "skipped:", file=sys.stderr)
printlist(self.skipped, file=sys.stderr)
def rerun_failed_tests(self):
self.log()
if self.ns.python:
# Temp patch for https://github.com/python/cpython/issues/94052
self.log(
"Re-running failed tests is not supported with --python "
"host runner option."
)
return
self.ns.verbose = True
self.ns.failfast = False
self.ns.verbose3 = False
self.first_result = self.get_tests_result()
self.log("Re-running failed tests in verbose mode")
rerun_list = list(self.need_rerun)
self.need_rerun.clear()
for result in rerun_list:
[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: Łukasz 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
test_name = result.test_name
self.rerun.append(test_name)
errors = result.errors or []
failures = result.failures or []
error_names = [
self.normalize_test_name(test_full_name, is_error=True)
for (test_full_name, *_) in errors]
failure_names = [
self.normalize_test_name(test_full_name)
for (test_full_name, *_) in failures]
self.ns.verbose = True
orig_match_tests = self.ns.match_tests
if errors or failures:
if self.ns.match_tests is None:
self.ns.match_tests = []
self.ns.match_tests.extend(error_names)
self.ns.match_tests.extend(failure_names)
matching = "matching: " + ", ".join(self.ns.match_tests)
self.log(f"Re-running {test_name} in verbose mode ({matching})")
else:
self.log(f"Re-running {test_name} in verbose mode")
result = runtest(self.ns, test_name)
self.ns.match_tests = orig_match_tests
self.accumulate_result(result, rerun=True)
[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: Łukasz 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 result.state == State.INTERRUPTED:
break
if self.bad:
print(count(len(self.bad), 'test'), "failed again:")
printlist(self.bad)
self.display_result()
def normalize_test_name(self, test_full_name, *, is_error=False):
short_name = test_full_name.split(" ")[0]
if is_error and short_name in _TEST_LIFECYCLE_HOOKS:
# This means that we have a failure in a life-cycle hook,
# we need to rerun the whole module or class suite.
# Basically the error looks like this:
# ERROR: setUpClass (test.test_reg_ex.RegTest)
# or
# ERROR: setUpModule (test.test_reg_ex)
# So, we need to parse the class / module name.
lpar = test_full_name.index('(')
rpar = test_full_name.index(')')
return test_full_name[lpar + 1: rpar].split('.')[-1]
return short_name
def display_result(self):
# If running the test suite for PGO then no one cares about results.
if self.ns.pgo:
return
print()
print("== Tests result: %s ==" % self.get_tests_result())
if self.interrupted:
print("Test suite interrupted by signal SIGINT.")
omitted = set(self.selected) - self.get_executed()
if omitted:
print()
print(count(len(omitted), "test"), "omitted:")
printlist(omitted)
if self.good and not self.ns.quiet:
print()
if (not self.bad
and not self.skipped
and not self.interrupted
and len(self.good) > 1):
print("All", end=' ')
print(count(len(self.good), "test"), "OK.")
if self.ns.print_slow:
self.test_times.sort(reverse=True)
print()
print("10 slowest tests:")
for test_time, test in self.test_times[:10]:
print("- %s: %s" % (test, format_duration(test_time)))
if self.bad:
print()
print(count(len(self.bad), "test"), "failed:")
printlist(self.bad)
if self.environment_changed:
print()
print("{} altered the execution environment:".format(
count(len(self.environment_changed), "test")))
printlist(self.environment_changed)
if self.skipped and not self.ns.quiet:
print()
print(count(len(self.skipped), "test"), "skipped:")
printlist(self.skipped)
if self.rerun:
print()
print("%s:" % count(len(self.rerun), "re-run test"))
printlist(self.rerun)
if self.run_no_tests:
print()
print(count(len(self.run_no_tests), "test"), "run no tests:")
printlist(self.run_no_tests)
def run_tests_sequential(self):
if self.ns.trace:
import trace
self.tracer = trace.Trace(trace=False, count=True)
save_modules = sys.modules.keys()
msg = "Run tests sequentially"
if self.ns.timeout:
msg += " (timeout: %s)" % format_duration(self.ns.timeout)
self.log(msg)
previous_test = None
for test_index, test_name in enumerate(self.tests, 1):
[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: Łukasz 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
start_time = time.perf_counter()
text = test_name
if previous_test:
text = '%s -- %s' % (text, previous_test)
self.display_progress(test_index, text)
if self.tracer:
# If we're tracing code coverage, then we don't exit with status
# if on a false return value from main.
cmd = ('result = runtest(self.ns, test_name); '
'self.accumulate_result(result)')
ns = dict(locals())
self.tracer.runctx(cmd, globals=globals(), locals=ns)
result = ns['result']
else:
result = runtest(self.ns, test_name)
self.accumulate_result(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: Łukasz 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 result.state == State.INTERRUPTED:
break
previous_test = str(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: Łukasz 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
test_time = time.perf_counter() - start_time
if test_time >= PROGRESS_MIN_TIME:
previous_test = "%s in %s" % (previous_test, format_duration(test_time))
[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: Łukasz 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
elif result.state == State.PASSED:
# be quiet: say nothing if the test passed shortly
previous_test = None
# Unload the newly imported modules (best effort finalization)
for module in sys.modules.keys():
if module not in save_modules and module.startswith("test."):
support.unload(module)
[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: Łukasz 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 self.ns.failfast and result.is_failed(self.ns.fail_env_changed):
break
if previous_test:
print(previous_test)
def _test_forever(self, tests):
while True:
for test_name in tests:
yield test_name
if self.bad:
return
if self.ns.fail_env_changed and self.environment_changed:
return
def display_header(self):
# Print basic platform information
print("==", platform.python_implementation(), *sys.version.split())
print("==", platform.platform(aliased=True),
"%s-endian" % sys.byteorder)
[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: Łukasz 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
print("== Python build:", ' '.join(get_build_info()))
print("== cwd:", os.getcwd())
cpu_count = os.cpu_count()
if cpu_count:
print("== CPU count:", cpu_count)
print("== encodings: locale=%s, FS=%s"
% (locale.getencoding(), sys.getfilesystemencoding()))
[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: Łukasz 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.display_sanitizers()
def display_sanitizers(self):
# This makes it easier to remember what to set in your local
# environment when trying to reproduce a sanitizer failure.
asan = support.check_sanitizer(address=True)
msan = support.check_sanitizer(memory=True)
ubsan = support.check_sanitizer(ub=True)
sanitizers = []
if asan:
sanitizers.append("address")
if msan:
sanitizers.append("memory")
if ubsan:
sanitizers.append("undefined behavior")
if not sanitizers:
return
print(f"== sanitizers: {', '.join(sanitizers)}")
for sanitizer, env_var in (
(asan, "ASAN_OPTIONS"),
(msan, "MSAN_OPTIONS"),
(ubsan, "UBSAN_OPTIONS"),
):
options= os.environ.get(env_var)
if sanitizer and options is not None:
print(f"== {env_var}={options!r}")
def no_tests_run(self):
return not any((self.good, self.bad, self.skipped, self.interrupted,
self.environment_changed))
def get_tests_result(self):
result = []
if self.bad:
result.append("FAILURE")
elif self.ns.fail_env_changed and self.environment_changed:
result.append("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: Łukasz 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
elif self.no_tests_run():
result.append("NO TESTS RAN")
if self.interrupted:
result.append("INTERRUPTED")
if not result:
result.append("SUCCESS")
result = ', '.join(result)
if self.first_result:
result = '%s then %s' % (self.first_result, result)
return result
def run_tests(self):
# For a partial run, we do not need to clutter the output.
if (self.ns.header
or not(self.ns.pgo or self.ns.quiet or self.ns.single
or self.tests or self.ns.args)):
self.display_header()
if self.ns.huntrleaks:
warmup, repetitions, _ = self.ns.huntrleaks
if warmup < 3:
msg = ("WARNING: Running tests with --huntrleaks/-R and less than "
"3 warmup repetitions can give false positives!")
print(msg, file=sys.stdout, flush=True)
if self.ns.randomize:
print("Using random seed", self.ns.random_seed)
if self.ns.forever:
self.tests = self._test_forever(list(self.selected))
self.test_count = ''
self.test_count_width = 3
else:
self.tests = iter(self.selected)
self.test_count = '/{}'.format(len(self.selected))
self.test_count_width = len(self.test_count) - 1
if self.ns.use_mp:
from test.libregrtest.runtest_mp import run_tests_multiprocess
# If we're on windows and this is the parent runner (not a worker),
# track the load average.
if sys.platform == 'win32' and self.worker_test_name is None:
from test.libregrtest.win_utils import WindowsLoadTracker
try:
self.win_load_tracker = WindowsLoadTracker()
except PermissionError as error:
# Standard accounts may not have access to the performance
# counters.
print(f'Failed to create WindowsLoadTracker: {error}')
try:
run_tests_multiprocess(self)
finally:
if self.win_load_tracker is not None:
self.win_load_tracker.close()
self.win_load_tracker = None
else:
self.run_tests_sequential()
def finalize(self):
if self.next_single_filename:
if self.next_single_test:
with open(self.next_single_filename, 'w') as fp:
fp.write(self.next_single_test + '\n')
else:
os.unlink(self.next_single_filename)
if self.tracer:
r = self.tracer.results()
r.write_results(show_missing=True, summary=True,
coverdir=self.ns.coverdir)
print()
[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: Łukasz 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.display_summary()
if self.ns.runleaks:
os.system("leaks %d" % os.getpid())
[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: Łukasz 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 display_summary(self):
duration = time.perf_counter() - self.start_time
# Total duration
print("Total duration: %s" % format_duration(duration))
# Total tests
total = TestStats()
for stats in self.stats_dict.values():
total.accumulate(stats)
stats = [f'run={total.tests_run:,}']
if total.failures:
stats.append(f'failures={total.failures:,}')
if total.skipped:
stats.append(f'skipped={total.skipped:,}')
print(f"Total tests: {' '.join(stats)}")
# Total test files
report = [f'success={len(self.good)}']
if self.bad:
report.append(f'failed={len(self.bad)}')
if self.environment_changed:
report.append(f'env_changed={len(self.environment_changed)}')
if self.skipped:
report.append(f'skipped={len(self.skipped)}')
if self.resource_denied:
report.append(f'resource_denied={len(self.resource_denied)}')
if self.rerun:
report.append(f'rerun={len(self.rerun)}')
if self.run_no_tests:
report.append(f'run_no_tests={len(self.run_no_tests)}')
print(f"Total test files: {' '.join(report)}")
# Result
result = self.get_tests_result()
print(f"Result: {result}")
def save_xml_result(self):
if not self.ns.xmlpath and not self.testsuite_xml:
return
import xml.etree.ElementTree as ET
root = ET.Element("testsuites")
# Manually count the totals for the overall summary
totals = {'tests': 0, 'errors': 0, 'failures': 0}
for suite in self.testsuite_xml:
root.append(suite)
for k in totals:
try:
totals[k] += int(suite.get(k, 0))
except ValueError:
pass
for k, v in totals.items():
root.set(k, str(v))
xmlpath = os.path.join(os_helper.SAVEDCWD, self.ns.xmlpath)
with open(xmlpath, 'wb') as f:
for s in ET.tostringlist(root):
f.write(s)
def fix_umask(self):
if support.is_emscripten:
# Emscripten has default umask 0o777, which breaks some tests.
# see https://github.com/emscripten-core/emscripten/issues/17269
old_mask = os.umask(0)
if old_mask == 0o777:
os.umask(0o027)
else:
os.umask(old_mask)
def set_temp_dir(self):
if self.ns.tempdir:
self.tmp_dir = self.ns.tempdir
if not self.tmp_dir:
# When tests are run from the Python build directory, it is best practice
# to keep the test files in a subfolder. This eases the cleanup of leftover
# files using the "make distclean" command.
if sysconfig.is_python_build():
self.tmp_dir = sysconfig.get_config_var('abs_builddir')
if self.tmp_dir is None:
# bpo-30284: On Windows, only srcdir is available. Using
# abs_builddir mostly matters on UNIX when building Python
# out of the source tree, especially when the source tree
# is read only.
self.tmp_dir = sysconfig.get_config_var('srcdir')
self.tmp_dir = os.path.join(self.tmp_dir, 'build')
else:
self.tmp_dir = tempfile.gettempdir()
self.tmp_dir = os.path.abspath(self.tmp_dir)
def create_temp_dir(self):
os.makedirs(self.tmp_dir, exist_ok=True)
# Define a writable temp dir that will be used as cwd while running
# the tests. The name of the dir includes the pid to allow parallel
# testing (see the -j option).
# Emscripten and WASI have stubbed getpid(), Emscripten has only
# milisecond clock resolution. Use randint() instead.
if sys.platform in {"emscripten", "wasi"}:
nounce = random.randint(0, 1_000_000)
else:
nounce = os.getpid()
if self.worker_test_name is not None:
test_cwd = 'test_python_worker_{}'.format(nounce)
else:
test_cwd = 'test_python_{}'.format(nounce)
test_cwd += os_helper.FS_NONASCII
test_cwd = os.path.join(self.tmp_dir, test_cwd)
return test_cwd
def cleanup(self):
import glob
path = os.path.join(glob.escape(self.tmp_dir), 'test_python_*')
print("Cleanup %s directory" % self.tmp_dir)
for name in glob.glob(path):
if os.path.isdir(name):
print("Remove directory: %s" % name)
os_helper.rmtree(name)
else:
print("Remove file: %s" % name)
os_helper.unlink(name)
def main(self, tests=None, **kwargs):
self.parse_args(kwargs)
self.set_temp_dir()
self.fix_umask()
if self.ns.cleanup:
self.cleanup()
sys.exit(0)
test_cwd = self.create_temp_dir()
try:
# Run the tests in a context manager that temporarily changes the CWD
# to a temporary and writable directory. If it's not possible to
# create or change the CWD, the original CWD will be used.
# The original CWD is available from os_helper.SAVEDCWD.
with os_helper.temp_cwd(test_cwd, quiet=True):
# When using multiprocessing, worker processes will use test_cwd
# as their parent temporary directory. So when the main process
# exit, it removes also subdirectories of worker processes.
self.ns.tempdir = test_cwd
self._main(tests, kwargs)
except SystemExit as exc:
# bpo-38203: Python can hang at exit in Py_Finalize(), especially
# on threading._shutdown() call: put a timeout
if threading_helper.can_start_thread:
faulthandler.dump_traceback_later(EXIT_TIMEOUT, exit=True)
sys.exit(exc.code)
def getloadavg(self):
if self.win_load_tracker is not None:
return self.win_load_tracker.getloadavg()
if hasattr(os, 'getloadavg'):
return os.getloadavg()[0]
return None
def _main(self, tests, kwargs):
if self.worker_test_name is not None:
from test.libregrtest.runtest_mp import run_tests_worker
run_tests_worker(self.ns, self.worker_test_name)
if self.ns.wait:
input("Press any key to continue...")
support.PGO = self.ns.pgo
support.PGO_EXTENDED = self.ns.pgo_extended
setup_tests(self.ns)
self.find_tests(tests)
if self.ns.list_tests:
self.list_tests()
sys.exit(0)
if self.ns.list_cases:
self.list_cases()
sys.exit(0)
self.run_tests()
self.display_result()
if self.ns.verbose2 and self.bad:
self.rerun_failed_tests()
self.finalize()
self.save_xml_result()
if self.bad:
[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: Łukasz 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
sys.exit(EXITCODE_BAD_TEST)
if self.interrupted:
[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: Łukasz 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
sys.exit(EXITCODE_INTERRUPTED)
if self.ns.fail_env_changed and self.environment_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: Łukasz 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
sys.exit(EXITCODE_ENV_CHANGED)
if self.no_tests_run():
sys.exit(EXITCODE_NO_TESTS_RAN)
sys.exit(0)
def main(tests=None, **kwargs):
"""Run the Python suite."""
Regrtest().main(tests=tests, **kwargs)