from __future__ import print_function from functools import partial import random from typing import Any, Callable, Dict, Iterable, List, Optional, Set, Tuple, \ Text, Type from unittest import loader, runner # type: ignore # Mypy cannot pick these up. from unittest.result import TestResult from django.conf import settings from django.db import connections, ProgrammingError from django.test import TestCase from django.test import runner as django_runner from django.test.runner import DiscoverRunner from django.test.signals import template_rendered from zerver.lib import test_classes, test_helpers from zerver.lib.cache import bounce_key_prefix_for_testing from zerver.lib.test_classes import flush_caches_for_testing from zerver.lib.sqlalchemy_utils import get_sqlalchemy_connection from zerver.lib.test_helpers import ( get_all_templates, write_instrumentation_reports, append_instrumentation_data ) import os import subprocess import sys import time import traceback import unittest if False: # Only needed by mypy. from multiprocessing.sharedctypes import Synchronized _worker_id = 0 # Used to identify the worker process. def slow(slowness_reason): # type: (str) -> Callable[[Callable], Callable] ''' This is a decorate that annotates a test as being "known to be slow." The decorator will set expected_run_time and slowness_reason as atributes of the function. Other code can use this annotation as needed, e.g. to exclude these tests in "fast" mode. ''' def decorator(f): # type: (Any) -> Any f.slowness_reason = slowness_reason return f return decorator def is_known_slow_test(test_method): # type: (Any) -> bool return hasattr(test_method, 'slowness_reason') def full_test_name(test): # type: (TestCase) -> str test_module = test.__module__ test_class = test.__class__.__name__ test_method = test._testMethodName return '%s.%s.%s' % (test_module, test_class, test_method) def get_test_method(test): # type: (TestCase) -> Callable[[], None] return getattr(test, test._testMethodName) # Each tuple is delay, test_name, slowness_reason TEST_TIMINGS = [] # type: List[Tuple[float, str, str]] def report_slow_tests(): # type: () -> None timings = sorted(TEST_TIMINGS, reverse=True) print('SLOWNESS REPORT') print(' delay test') print(' ---- ----') for delay, test_name, slowness_reason in timings[:15]: if not slowness_reason: slowness_reason = 'UNKNOWN WHY SLOW, please investigate' print(' %0.3f %s\n %s\n' % (delay, test_name, slowness_reason)) print('...') for delay, test_name, slowness_reason in timings[100:]: if slowness_reason: print(' %.3f %s is not that slow' % (delay, test_name)) print(' consider removing @slow decorator') print(' This may no longer be true: %s' % (slowness_reason,)) def enforce_timely_test_completion(test_method, test_name, delay, result): # type: (Any, str, float, TestResult) -> None if hasattr(test_method, 'slowness_reason'): max_delay = 1.1 # seconds else: max_delay = 0.4 # seconds if delay > max_delay: msg = '** Test is TOO slow: %s (%.3f s)\n' % (test_name, delay) result.addInfo(test_method, msg) def fast_tests_only(): # type: () -> bool return "FAST_TESTS_ONLY" in os.environ def run_test(test, result): # type: (TestCase, TestResult) -> bool failed = False test_method = get_test_method(test) if fast_tests_only() and is_known_slow_test(test_method): return failed test_name = full_test_name(test) bounce_key_prefix_for_testing(test_name) flush_caches_for_testing() if not hasattr(test, "_pre_setup"): # test_name is likely of the form unittest.loader.ModuleImportFailure.zerver.tests.test_upload import_failure_prefix = 'unittest.loader.ModuleImportFailure.' if test_name.startswith(import_failure_prefix): actual_test_name = test_name[len(import_failure_prefix):] error_msg = ("\nActual test to be run is %s, but import failed.\n" "Importing test module directly to generate clearer " "traceback:\n") % (actual_test_name,) result.addInfo(test, error_msg) try: command = [sys.executable, "-c", "import %s" % (actual_test_name,)] msg = "Import test command: `%s`" % (' '.join(command),) result.addInfo(test, msg) subprocess.check_call(command) except subprocess.CalledProcessError: msg = ("If that traceback is confusing, try doing the " "import inside `./manage.py shell`") result.addInfo(test, msg) result.addError(test, sys.exc_info()) return True msg = ("Import unexpectedly succeeded! Something is wrong. Try " "running `import %s` inside `./manage.py shell`.\n" "If that works, you may have introduced an import " "cycle.") % (actual_test_name,) import_error = (Exception, Exception(msg), None) # type: Tuple[Any, Any, Any] result.addError(test, import_error) return True else: msg = "Test doesn't have _pre_setup; something is wrong." error_pre_setup = (Exception, Exception(msg), None) # type: Tuple[Any, Any, Any] result.addError(test, error_pre_setup) return True test._pre_setup() start_time = time.time() test(result) # unittest will handle skipping, error, failure and success. delay = time.time() - start_time enforce_timely_test_completion(test_method, test_name, delay, result) slowness_reason = getattr(test_method, 'slowness_reason', '') TEST_TIMINGS.append((delay, test_name, slowness_reason)) test._post_teardown() return failed class TextTestResult(runner.TextTestResult): """ This class has unpythonic function names because base class follows this style. """ def __init__(self, *args, **kwargs): # type: (*Any, **Any) -> None super(TextTestResult, self).__init__(*args, **kwargs) self.failed_tests = [] # type: List[str] def addInfo(self, test, msg): # type: (TestCase, Text) -> None self.stream.write(msg) self.stream.flush() def addInstrumentation(self, test, data): # type: (TestCase, Dict[str, Any]) -> None append_instrumentation_data(data) def startTest(self, test): # type: (TestCase) -> None TestResult.startTest(self, test) self.stream.writeln("Running {}".format(full_test_name(test))) self.stream.flush() def addSuccess(self, *args, **kwargs): # type: (*Any, **Any) -> None TestResult.addSuccess(self, *args, **kwargs) def addError(self, *args, **kwargs): # type: (*Any, **Any) -> None TestResult.addError(self, *args, **kwargs) def addFailure(self, *args, **kwargs): # type: (*Any, **Any) -> None TestResult.addFailure(self, *args, **kwargs) test_name = full_test_name(args[0]) self.failed_tests.append(test_name) def addSkip(self, test, reason): # type: (TestCase, Text) -> None TestResult.addSkip(self, test, reason) self.stream.writeln("** Skipping {}: {}".format(full_test_name(test), reason)) self.stream.flush() class RemoteTestResult(django_runner.RemoteTestResult): """ The class follows the unpythonic style of function names of the base class. """ def addInfo(self, test, msg): # type: (TestCase, Text) -> None self.events.append(('addInfo', self.test_index, msg)) def addInstrumentation(self, test, data): # type: (TestCase, Dict[str, Any]) -> None # Some elements of data['info'] cannot be serialized. if 'info' in data: del data['info'] self.events.append(('addInstrumentation', self.test_index, data)) def process_instrumented_calls(func): # type: (Callable) -> None for call in test_helpers.INSTRUMENTED_CALLS: func(call) def run_subsuite(args): # type: (Tuple[int, Tuple[Type[Iterable[TestCase]], List[str]], bool]) -> Tuple[int, Any] subsuite_index, subsuite, failfast = args runner = RemoteTestRunner(failfast=failfast) result = runner.run(deserialize_suite(subsuite)) # Now we send instrumentation related events. This data will be # appended to the data structure in the main thread. For Mypy, # type of Partial is different from Callable. All the methods of # TestResult are passed TestCase as the first argument but # addInstrumentation does not need it. process_instrumented_calls(partial(result.addInstrumentation, None)) # type: ignore return subsuite_index, result.events # Monkey-patch database creation to fix unnecessary sleep(1) from django.db.backends.postgresql.creation import DatabaseCreation def _replacement_destroy_test_db(self, test_database_name, verbosity): # type: (Any, str, Any) -> None """Replacement for Django's _destroy_test_db that removes the unnecessary sleep(1).""" with self.connection._nodb_connection.cursor() as cursor: cursor.execute("DROP DATABASE %s" % self.connection.ops.quote_name(test_database_name)) DatabaseCreation._destroy_test_db = _replacement_destroy_test_db def destroy_test_databases(database_id=None): # type: (Optional[int]) -> None """ When database_id is None, the name of the databases is picked up by the database settings. """ for alias in connections: connection = connections[alias] try: connection.creation.destroy_test_db(number=database_id) except ProgrammingError: # DB doesn't exist. No need to do anything. pass def create_test_databases(database_id): # type: (int) -> None for alias in connections: connection = connections[alias] connection.creation.clone_test_db( number=database_id, keepdb=True, ) settings_dict = connection.creation.get_test_db_clone_settings(database_id) # connection.settings_dict must be updated in place for changes to be # reflected in django.db.connections. If the following line assigned # connection.settings_dict = settings_dict, new threads would connect # to the default database instead of the appropriate clone. connection.settings_dict.update(settings_dict) connection.close() def init_worker(counter): # type: (Synchronized) -> None """ This function runs only under parallel mode. It initializes the individual processes which are also called workers. """ global _worker_id test_classes.API_KEYS = {} # Clear the cache from zerver.lib.cache import get_cache_backend cache = get_cache_backend(None) cache.clear() # Close all connections connections.close_all() with counter.get_lock(): counter.value += 1 _worker_id = counter.value destroy_test_databases(_worker_id) create_test_databases(_worker_id) class TestSuite(unittest.TestSuite): def run(self, result, debug=False): # type: (TestResult, Optional[bool]) -> TestResult """ This function mostly contains the code from unittest.TestSuite.run. The need to override this function occurred because we use run_test to run the testcase. """ topLevel = False if getattr(result, '_testRunEntered', False) is False: result._testRunEntered = topLevel = True for test in self: # type: ignore # Mypy cannot recognize this # but this is correct. Taken from unittest. if result.shouldStop: break if isinstance(test, TestSuite): test.run(result, debug=debug) else: self._tearDownPreviousClass(test, result) # type: ignore self._handleModuleFixture(test, result) # type: ignore self._handleClassSetUp(test, result) # type: ignore result._previousTestClass = test.__class__ if (getattr(test.__class__, '_classSetupFailed', False) or getattr(result, '_moduleSetUpFailed', False)): continue failed = run_test(test, result) if failed or result.shouldStop: result.shouldStop = True break if topLevel: self._tearDownPreviousClass(None, result) # type: ignore self._handleModuleTearDown(result) # type: ignore result._testRunEntered = False return result class TestLoader(loader.TestLoader): suiteClass = TestSuite class ParallelTestSuite(django_runner.ParallelTestSuite): run_subsuite = run_subsuite init_worker = init_worker def __init__(self, suite, processes, failfast): # type: (TestSuite, int, bool) -> None super(ParallelTestSuite, self).__init__(suite, processes, failfast) self.subsuites = SubSuiteList(self.subsuites) # type: SubSuiteList class Runner(DiscoverRunner): test_suite = TestSuite test_loader = TestLoader() parallel_test_suite = ParallelTestSuite def __init__(self, *args, **kwargs): # type: (*Any, **Any) -> None DiscoverRunner.__init__(self, *args, **kwargs) # `templates_rendered` holds templates which were rendered # in proper logical tests. self.templates_rendered = set() # type: Set[str] # `shallow_tested_templates` holds templates which were rendered # in `zerver.tests.test_templates`. self.shallow_tested_templates = set() # type: Set[str] template_rendered.connect(self.on_template_rendered) self.database_id = random.randint(1, 10000) def get_resultclass(self): # type: () -> Type[TestResult] return TextTestResult def on_template_rendered(self, sender, context, **kwargs): # type: (Any, Dict[str, Any], **Any) -> None if hasattr(sender, 'template'): template_name = sender.template.name if template_name not in self.templates_rendered: if context.get('shallow_tested') and template_name not in self.templates_rendered: self.shallow_tested_templates.add(template_name) else: self.templates_rendered.add(template_name) self.shallow_tested_templates.discard(template_name) def get_shallow_tested_templates(self): # type: () -> Set[str] return self.shallow_tested_templates def setup_test_environment(self, *args, **kwargs): # type: (*Any, **Any) -> Any settings.DATABASES['default']['NAME'] = settings.BACKEND_DATABASE_TEMPLATE destroy_test_databases(self.database_id) create_test_databases(self.database_id) return super(Runner, self).setup_test_environment(*args, **kwargs) def teardown_test_environment(self, *args, **kwargs): # type: (*Any, **Any) -> Any # No need to pass the database id now. It will be picked up # automatically through settings. destroy_test_databases() return super(Runner, self).teardown_test_environment(*args, **kwargs) def run_tests(self, test_labels, extra_tests=None, full_suite=False, **kwargs): # type: (List[str], Optional[List[TestCase]], bool, **Any) -> Tuple[bool, List[str]] self.setup_test_environment() try: suite = self.build_suite(test_labels, extra_tests) except AttributeError: traceback.print_exc() print() print(" This is often caused by a test module/class/function that doesn't exist or ") print(" import properly. You can usually debug in a `manage.py shell` via e.g. ") print(" import zerver.tests.test_messages") print(" from zerver.tests.test_messages import StreamMessagesTest") print(" StreamMessagesTest.test_message_to_stream") print() sys.exit(1) # We have to do the next line to avoid flaky scenarios where we # run a single test and getting an SA connection causes data from # a Django connection to be rolled back mid-test. get_sqlalchemy_connection() result = self.run_suite(suite) self.teardown_test_environment() failed = self.suite_result(suite, result) if not failed: write_instrumentation_reports(full_suite=full_suite) return failed, result.failed_tests def get_test_names(suite): # type: (TestSuite) -> List[str] return [full_test_name(t) for t in get_tests_from_suite(suite)] def get_tests_from_suite(suite): # type: (TestSuite) -> TestCase for test in suite: # type: ignore if isinstance(test, TestSuite): for child in get_tests_from_suite(test): yield child else: yield test def serialize_suite(suite): # type: (TestSuite) -> Tuple[Type[TestSuite], List[str]] return type(suite), get_test_names(suite) def deserialize_suite(args): # type: (Tuple[Type[Iterable[TestCase]], List[str]]) -> Iterable[TestCase] suite_class, test_names = args suite = suite_class() # type: ignore # Gives abstract type error. tests = TestLoader().loadTestsFromNames(test_names) for test in get_tests_from_suite(tests): suite.addTest(test) return suite class RemoteTestRunner(django_runner.RemoteTestRunner): resultclass = RemoteTestResult class SubSuiteList(list): """ This class allows us to avoid changing the main logic of ParallelTestSuite and still make it serializable. """ def __init__(self, suites): # type: (List[TestSuite]) -> None serialized_suites = [serialize_suite(s) for s in suites] super(SubSuiteList, self).__init__(serialized_suites) def __getitem__(self, index): # type: (Any) -> Any suite = super(SubSuiteList, self).__getitem__(index) return deserialize_suite(suite)