bpo-37531, bpo-38207: On timeout, regrtest no longer attempts to call
`popen.communicate() again: it can hang until all child processes
using stdout and stderr pipes completes. Kill the worker process and
ignores its output.
Reenable test_regrtest.test_multiprocessing_timeout().
bpo-37531: Change also the faulthandler timeout of the main process
from 1 minute to 5 minutes, for Python slowest buildbots.
(cherry picked from commit 0ec618af98)
Co-authored-by: Victor Stinner <vstinner@python.org>
* bpo-36670, regrtest: Fix WindowsLoadTracker() for partial line (GH-16550)
WindowsLoadTracker.read_output() now uses a short buffer for
incomplete line.
(cherry picked from commit 3e04cd268e)
* bpo-36670: Enhance regrtest WindowsLoadTracker (GH-16553)
The last line is now passed to the parser even if it does not end
with a newline, but only if it's a valid value.
(cherry picked from commit c65119d5bf)
* bpo-36670: Enhance regrtest (GH-16556)
* Add log() method: add timestamp and load average prefixes
to main messages.
* WindowsLoadTracker:
* LOAD_FACTOR_1 is now computed using SAMPLING_INTERVAL
* Initialize the load to the arithmetic mean of the first 5 values
of the Processor Queue Length value (so over 5 seconds), rather
than 0.0.
* Handle BrokenPipeError and when typeperf exit.
* format_duration(1.5) now returns '1.5 sec', rather than
'1 sec 500 ms'
(cherry picked from commit 098e25672f)
(cherry picked from commit de3195c937)
Co-authored-by: Victor Stinner <vstinner@python.org>
* Windows: Fix counter name in WindowsLoadTracker. Counter names are
localized: use the registry to get the counter name. Original
change written by Lorenz Mende.
* Regrtest.main() now ensures that the Windows load tracker is also
killed if an exception is raised
* TestWorkerProcess now ensures that worker processes are no longer
running before exiting: kill also worker processes when an
exception is raised.
* Enhance regrtest messages and warnings: include test name,
duration, add a worker identifier, etc.
* Rename MultiprocessRunner to TestWorkerProcess
* Use print_warning() to display warnings.
Co-Authored-By: Lorenz Mende <Lorenz.mende@gmail.com>
(cherry picked from commit 982bfa4da0)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
When using multiprocesss (-jN), the main process now uses a timeout
of 60 seconds instead of the double of the --timeout value. The
buildbot server stops a job which does not produce any output in 1200
seconds.
(cherry picked from commit 46b0b81220)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
bpo-37531: Enhance regrtest multiprocess timeout (GH-15345)
* Write a message when killing a worker process
* Put a timeout on the second popen.communicate() call
(after killing the process)
* Put a timeout on popen.wait() call
* Catch popen.kill() and popen.wait() exceptions
(cherry picked from commit de2d9eed8b)
When using multiprocessing (-jN option), worker processes now create
their temporary directory inside the temporary directory of the
main process. So the main process is able to remove temporary
directories of worker processes even if they crash or when they are
killed by regrtest on KeyboardInterrupt (CTRL+c).
Rework also how multiprocessing arguments are parsed in main.py.
(cherry picked from commit 3c93153f7d)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
MultiprocessThread.kill() now closes stdout and stderr to prevent
popen.communicate() to hang.
(cherry picked from commit c923c3449f)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
"python3 -m test -jN ..." now continues the execution of next tests
when a worker process crash (CHILD_ERROR state). Previously, the test
suite stopped immediately. Use --failfast to stop at the first error.
Moreover, --forever now also implies --failfast.
(cherry picked from commit b0917df329)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
* Clean up code which checked presence of os.{stat,lstat,chmod} (GH-11643)
(cherry picked from commit 8377cd4fcd)
* bpo-36725: regrtest: add TestResult type (GH-12960)
* Add TestResult and MultiprocessResult types to ensure that results
always have the same fields.
* runtest() now handles KeyboardInterrupt
* accumulate_result() and format_test_result() now takes a TestResult
* cleanup_test_droppings() is now called by runtest() and mark the
test as ENV_CHANGED if the test leaks support.TESTFN file.
* runtest() now includes code "around" the test in the test timing
* Add print_warning() in test.libregrtest.utils to standardize how
libregrtest logs warnings to ease parsing the test output.
* support.unload() is now called with abstest rather than test_name
* Rename 'test' variable/parameter to 'test_name'
* dash_R(): remove unused the_module parameter
* Remove unused imports
(cherry picked from commit 4d29983185)
* bpo-36725: Refactor regrtest multiprocessing code (GH-12961)
Rewrite run_tests_multiprocess() function as a new MultiprocessRunner
class with multiple methods to better report errors and stop
immediately when needed.
Changes:
* Worker processes are now killed immediately if tests are
interrupted or if a test does crash (CHILD_ERROR): worker
processes are killed.
* Rewrite how errors in a worker thread are reported to
the main thread. No longer ignore BaseException or parsing errors
silently.
* Remove 'finished' variable: use worker.is_alive() instead
* Always compute omitted tests. Add Regrtest.get_executed() method.
(cherry picked from commit 3cde440f20)
* bpo-36719: regrtest always detect uncollectable objects (GH-12951)
regrtest now always detects uncollectable objects. Previously, the
check was only enabled by --findleaks. The check now also works with
-jN/--multiprocess N.
--findleaks becomes a deprecated alias to --fail-env-changed.
(cherry picked from commit 75120d2205)
* bpo-34060: Report system load when running test suite for Windows (GH-8357)
While Windows exposes the system processor queue length, the raw value
used for load calculations on Unix systems, it does not provide an API
to access the averaged value. Hence to calculate the load we must track
and average it ourselves. We can't use multiprocessing or a thread to
read it in the background while the tests run since using those would
conflict with test_multiprocessing and test_xxsubprocess.
Thus, we use Window's asynchronous IO API to run the tracker in the
background with it sampling at the correct rate. When we wish to access
the load we check to see if there's new data on the stream, if there is,
we update our load values.
(cherry picked from commit e16467af0b)
* bpo-36719: Fix regrtest re-run (GH-12964)
Properly handle a test which fail but then pass.
Add test_rerun_success() unit test.
(cherry picked from commit 837acc1957)
* bpo-36719: regrtest closes explicitly WindowsLoadTracker (GH-12965)
Regrtest.finalize() now closes explicitly the WindowsLoadTracker
instance.
(cherry picked from commit 00db7c73af)
* bpo-34605, libregrtest: Rename --slaveargs to --worker-args (GH-9099)
Rename also run_tests_slave() to run_tests_worker().
(cherry picked from commit 012f5b968a)
* bpo-34279, regrtest: Issue a warning if no tests have been executed (GH-10150)
(cherry picked from commit 9724348b43)
* test_regrtest: remove unused threading import
* bpo-33718: regrtest: use format_duration() to display failed tests (GH-7686)
* Enhance also format_duration(): work on integers and rounds towards
+infinity (math.ceil).
* Write unit tests on format_duration()
(cherry picked from commit 4ffe9c2b25)
* bpo-33873: regrtest: Add warning on -R 1:3 (GH-7736)
regrtest: Add warning when using less than 3 warmup runs like -R 1:3.
(cherry picked from commit cac4fef886)
* bpo-33873: Fix bug in `runtest.py` and add checks for invalid `-R` parameters (GH-7735)
Fix bug in `Lib/test/libregrtest/runtest.py` that makes running tests an extra time than the specified number of runs.
Add check for invalid --huntrleaks/-R parameters.
(cherry picked from commit 58ed7307ea)
Issue #29362: Catch a crash of a worker process as a normal failure and
continue to run next tests. It allows to get the usual test summary: single
line result (OK/FAIL), total duration, etc.
* Fix "-m test --forever": replace _test_forever() with self._test_forever()
* Add unit test for --forever
* Add unit test for a failing test
* Fix also some pyflakes warnings in libregrtest
* Remove runtest_ns(): pass directly ns to runtest().
* Create also Regrtest.rerun_failed_tests() method.
* Inline again Regrtest.run_test(): it's no more justified to have a method
Slaves (child processes running tests for regrtest -jN) now inherit
--memlimit/-M, --threshold/-t and --nowindows/-n options.
* -M, -t and -n are now supported with -jN
* Factorize code to run tests.
* run_test_in_subprocess() now pass the whole "ns" namespace to the child
process.
Running the Python test suite with -jN now:
- Display the duration of tests which took longer than 30 seconds
- Display the tests currently running since at least 30 seconds
- Display the tests we are waiting for when the test suite is interrupted
Clenaup also run_test_in_subprocess() code.
Python doesn't display the refcount anymore by default. It only displays it
when -X showrefcount command line option is used, which is not the case here.
regrtest can be run with -X showrefcount, the option is not inherited by child
processes.
Move the code to run tests in multiple processes using threading and subprocess
to a new submodule.
Move also slave_runner() (renamed to run_tests_slave()) and
run_test_in_subprocess() (renamed to run_tests_in_subprocess()) there.