Commit 6c72df9e authored by Michael Achenbach's avatar Michael Achenbach Committed by Commit Bot

Reland "[test] Add test runner system tests for flaky tests"

This is a reland of 96f55796
Original change's description:
> [test] Add test runner system tests for flaky tests
>
> This uses a temporary file in the mocked d8 script to let d8 flip
> between FAIL and PASS on rerun.
>
> This adds a separate test root with dedicated test suite for testing
> flakes to no interfere with existing status file configs.
>
> NOTRY=true
>
> Bug: v8:6917
> Change-Id: Id43753650195fb74cceb2a3ee9014100cabad546
> Reviewed-on: https://chromium-review.googlesource.com/867917
> Commit-Queue: Michael Achenbach <machenbach@chromium.org>
> Reviewed-by: Sergiy Byelozyorov <sergiyb@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#50621}

TBR=sergiyb@chromium.org
NOTRY=true

Bug: v8:6917
Change-Id: I4e7bca272dfd8778fbb8d012fcd6fd8406158e43
Reviewed-on: https://chromium-review.googlesource.com/868433Reviewed-by: 's avatarMichael Achenbach <machenbach@chromium.org>
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#50625}
parent bc130d80
......@@ -217,6 +217,31 @@ class SystemTest(unittest.TestCase):
self.assertIn('Done running sweet/strawberries: FAIL', result.stdout, result)
self.assertEqual(1, result.returncode, result)
def check_cleaned_json_output(self, expected_results_name, actual_json):
# Check relevant properties of the json output.
with open(actual_json) as f:
json_output = json.load(f)[0]
pretty_json = json.dumps(json_output, indent=2, sort_keys=True)
# Replace duration in actual output as it's non-deterministic. Also
# replace the python executable prefix as it has a different absolute
# path dependent on where this runs.
def replace_variable_data(data):
data['duration'] = 1
data['command'] = ' '.join(
['/usr/bin/python'] + data['command'].split()[1:])
for data in json_output['slowest_tests']:
replace_variable_data(data)
for data in json_output['results']:
replace_variable_data(data)
json_output['duration_mean'] = 1
with open(os.path.join(TEST_DATA_ROOT, expected_results_name)) as f:
expected_test_results = json.load(f)
msg = None # Set to pretty_json for bootstrapping.
self.assertDictEqual(json_output, expected_test_results, msg)
def testFailWithRerunAndJSONProc(self):
self.testFailWithRerunAndJSON(infra_staging=True)
......@@ -246,33 +271,43 @@ class SystemTest(unittest.TestCase):
self.assertIn('1 tests failed', result.stdout, result)
self.assertEqual(0, result.returncode, result)
# Check relevant properties of the json output.
with open(json_path) as f:
json_output = json.load(f)[0]
pretty_json = json.dumps(json_output, indent=2, sort_keys=True)
# Replace duration in actual output as it's non-deterministic. Also
# replace the python executable prefix as it has a different absolute
# path dependent on where this runs.
def replace_variable_data(data):
data['duration'] = 1
data['command'] = ' '.join(
['/usr/bin/python'] + data['command'].split()[1:])
for data in json_output['slowest_tests']:
replace_variable_data(data)
for data in json_output['results']:
replace_variable_data(data)
json_output['duration_mean'] = 1
expected_results_name = 'expected_test_results1.json'
with open(os.path.join(TEST_DATA_ROOT, expected_results_name)) as f:
expected_test_results = json.load(f)
# TODO(majeski): Previously we only reported the variant flags in the
# flags field of the test result.
# After recent changes we report all flags, including the file names.
# This is redundant to the command. Needs investigation.
self.assertEqual(json_output, expected_test_results, pretty_json)
self.check_cleaned_json_output('expected_test_results1.json', json_path)
# FIXME(majeski): This runs 0 tests.
#def testFlakeWithRerunAndJSONProc(self):
# self.testFlakeWithRerunAndJSON(infra_staging=True)
def testFlakeWithRerunAndJSON(self, infra_staging=False):
"""Test re-running a failing test and output to json."""
with temp_base(baseroot='testroot2') as basedir:
json_path = os.path.join(basedir, 'out.json')
result = run_tests(
basedir,
'--mode=Release',
'--progress=verbose',
'--variants=default',
'--rerun-failures-count=2',
'--random-seed=123',
'--json-test-results', json_path,
'sweet',
infra_staging=infra_staging,
)
self.assertIn('Running 1 tests', result.stdout, result)
if not infra_staging:
self.assertIn(
'Done running sweet/bananaflakes: FAIL', result.stdout, result)
self.assertIn('1 tests failed', result.stdout, result)
else:
# TODO(majeski): Evaluate if this output is reasonable for a FAIL, PASS.
self.assertIn(
'Done running sweet/bananaflakes: pass', result.stdout, result)
self.assertIn('All tests succeeded', result.stdout, result)
self.assertEqual(0, result.returncode, result)
self.check_cleaned_json_output('expected_test_results2.json', json_path)
def testAutoDetect(self):
"""Fake a build with several auto-detected options.
......
{
"arch": "x64",
"duration_mean": 1,
"mode": "release",
"results": [
{
"command": "/usr/bin/python out/Release/d8_mocked.py --random-seed=123 bananaflakes --nohard-abort",
"duration": 1,
"exit_code": 1,
"expected": [
"PASS"
],
"flags": [
"--random-seed=123",
"bananaflakes",
"--nohard-abort"
],
"name": "sweet/bananaflakes",
"random_seed": 123,
"result": "FAIL",
"run": 1,
"stderr": "",
"stdout": "--random-seed=123 bananaflakes --nohard-abort\n",
"target_name": "d8_mocked.py",
"variant": "default"
},
{
"command": "/usr/bin/python out/Release/d8_mocked.py --random-seed=123 bananaflakes --nohard-abort",
"duration": 1,
"exit_code": 0,
"expected": [
"PASS"
],
"flags": [
"--random-seed=123",
"bananaflakes",
"--nohard-abort"
],
"name": "sweet/bananaflakes",
"random_seed": 123,
"result": "PASS",
"run": 2,
"stderr": "",
"stdout": "--random-seed=123 bananaflakes --nohard-abort\n",
"target_name": "d8_mocked.py",
"variant": "default"
}
],
"slowest_tests": [
{
"command": "/usr/bin/python out/Release/d8_mocked.py --random-seed=123 bananaflakes --nohard-abort",
"duration": 1,
"flags": [
"--random-seed=123",
"bananaflakes",
"--nohard-abort"
],
"marked_slow": false,
"name": "sweet/bananaflakes"
},
{
"command": "/usr/bin/python out/Release/d8_mocked.py --random-seed=123 bananaflakes --nohard-abort",
"duration": 1,
"flags": [
"--random-seed=123",
"bananaflakes",
"--nohard-abort"
],
"marked_slow": false,
"name": "sweet/bananaflakes"
}
],
"test_total": 2
}
# Copyright 2018 the V8 project authors. All rights reserved.
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.
"""
Dummy d8 replacement for flaky tests.
"""
import os
import sys
PATH = os.path.dirname(os.path.abspath(__file__))
print ' '.join(sys.argv[1:])
# Test files ending in 'flakes' should first fail then pass. We store state in
# a file side by side with the executable. No clean-up required as all tests
# run in a temp test root. Restriction: Only one variant is supported for now.
for arg in sys.argv[1:]:
if arg.endswith('flakes'):
flake_state = os.path.join(PATH, arg)
if os.path.exists(flake_state):
sys.exit(0)
else:
with open(flake_state, 'w') as f:
f.write('something')
sys.exit(1)
sys.exit(0)
# Copyright 2018 the V8 project authors. All rights reserved.
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.
[
]
# Copyright 2018 the V8 project authors. All rights reserved.
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.
"""
Dummy test suite extension with some flaky fruity tests.
"""
from testrunner.local import testsuite
from testrunner.objects import testcase
class TestSuite(testsuite.TestSuite):
def ListTests(self, context):
return map(
self._create_test,
['bananaflakes'],
)
def _test_class(self):
return TestCase
class TestCase(testcase.TestCase):
def get_shell(self):
return 'd8_mocked.py'
def _get_files_params(self, ctx):
return [self.name]
def GetSuite(name, root):
return TestSuite(name, root)
{
"current_cpu": "x64",
"dcheck_always_on": false,
"is_asan": false,
"is_cfi": false,
"is_component_build": false,
"is_debug": false,
"is_gcov_coverage": false,
"is_ubsan_vptr": false,
"is_msan": false,
"is_tsan": false,
"target_cpu": "x64",
"v8_current_cpu": "x64",
"v8_enable_i18n_support": true,
"v8_enable_verify_predictable": false,
"v8_target_cpu": "x64",
"v8_use_snapshot": true
}
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment