ladybird/Userland/Libraries/LibTest/TestSuite.cpp
Martin Janiczek a60e3b17b1 LibTest: Expand test result bool to a TestResult
This will be used in the randomized tests a lot more than it is in the
unit tests / benchmarks; randomized tests will run the test function
multiple times, check the result and optionally start shrinking the
failing input. Generators will also be able to fail, resulting in some
of the new TestResult variants.
2023-10-26 17:26:52 -06:00

249 lines
8.1 KiB
C++

/*
* Copyright (c) 2018-2020, Andreas Kling <kling@serenityos.org>
* Copyright (c) 2021, Andrew Kaster <akaster@serenityos.org>
*
* SPDX-License-Identifier: BSD-2-Clause
*/
#include <LibTest/Macros.h> // intentionally first -- we redefine VERIFY and friends in here
#include <AK/Function.h>
#include <LibCore/ArgsParser.h>
#include <LibTest/TestResult.h>
#include <LibTest/TestSuite.h>
#include <math.h>
#include <stdlib.h>
#include <sys/time.h>
namespace Test {
TestSuite* TestSuite::s_global = nullptr;
class TestElapsedTimer {
public:
TestElapsedTimer() { restart(); }
void restart() { gettimeofday(&m_started, nullptr); }
u64 elapsed_milliseconds()
{
struct timeval now = {};
gettimeofday(&now, nullptr);
struct timeval delta = {};
timersub(&now, &m_started, &delta);
return delta.tv_sec * 1000 + delta.tv_usec / 1000;
}
private:
struct timeval m_started = {};
};
// Declared in Macros.h
TestResult current_test_result()
{
return TestSuite::the().current_test_result();
}
// Declared in Macros.h
void set_current_test_result(TestResult result)
{
TestSuite::the().set_current_test_result(result);
}
// Declared in TestCase.h
void add_test_case_to_suite(NonnullRefPtr<TestCase> const& test_case)
{
TestSuite::the().add_case(test_case);
}
// Declared in TestCase.h
void set_suite_setup_function(Function<void()> setup)
{
TestSuite::the().set_suite_setup(move(setup));
}
static DeprecatedString test_result_to_string(TestResult result)
{
switch (result) {
case TestResult::NotRun:
return "Not run";
case TestResult::Passed:
return "Completed";
case TestResult::Failed:
return "Failed";
default:
return "Unknown TestResult";
}
}
int TestSuite::main(DeprecatedString const& suite_name, Span<StringView> arguments)
{
m_suite_name = suite_name;
Core::ArgsParser args_parser;
bool do_tests_only = getenv("TESTS_ONLY") != nullptr;
bool do_benchmarks_only = false;
bool do_list_cases = false;
StringView search_string = "*"sv;
args_parser.add_option(do_tests_only, "Only run tests.", "tests", 0);
args_parser.add_option(do_benchmarks_only, "Only run benchmarks.", "bench", 0);
args_parser.add_option(m_benchmark_repetitions, "Number of times to repeat each benchmark (default 1)", "benchmark_repetitions", 0, "N");
args_parser.add_option(do_list_cases, "List available test cases.", "list", 0);
args_parser.add_positional_argument(search_string, "Only run matching cases.", "pattern", Core::ArgsParser::Required::No);
args_parser.parse(arguments);
if (m_setup)
m_setup();
auto const& matching_tests = find_cases(search_string, !do_benchmarks_only, !do_tests_only);
if (do_list_cases) {
outln("Available cases for {}:", suite_name);
for (auto const& test : matching_tests) {
outln(" {}", test->name());
}
return 0;
}
outln("Running {} cases out of {}.", matching_tests.size(), m_cases.size());
return run(matching_tests);
}
Vector<NonnullRefPtr<TestCase>> TestSuite::find_cases(DeprecatedString const& search, bool find_tests, bool find_benchmarks)
{
Vector<NonnullRefPtr<TestCase>> matches;
for (auto& t : m_cases) {
if (!search.is_empty() && !t->name().matches(search, CaseSensitivity::CaseInsensitive)) {
continue;
}
if (!find_tests && !t->is_benchmark()) {
continue;
}
if (!find_benchmarks && t->is_benchmark()) {
continue;
}
matches.append(t);
}
return matches;
}
int TestSuite::run(Vector<NonnullRefPtr<TestCase>> const& tests)
{
size_t test_count = 0;
size_t test_passed_count = 0;
size_t test_failed_count = 0;
size_t benchmark_count = 0;
size_t benchmark_passed_count = 0;
size_t benchmark_failed_count = 0;
TestElapsedTimer global_timer;
for (auto const& t : tests) {
auto const test_type = t->is_benchmark() ? "benchmark" : "test";
auto const repetitions = t->is_benchmark() ? m_benchmark_repetitions : 1;
warnln("Running {} '{}'.", test_type, t->name());
m_current_test_result = TestResult::NotRun;
u64 total_time = 0;
u64 sum_of_squared_times = 0;
u64 min_time = NumericLimits<u64>::max();
u64 max_time = 0;
for (u64 i = 0; i < repetitions; ++i) {
TestElapsedTimer timer;
t->func()();
auto const iteration_time = timer.elapsed_milliseconds();
total_time += iteration_time;
sum_of_squared_times += iteration_time * iteration_time;
min_time = min(min_time, iteration_time);
max_time = max(max_time, iteration_time);
// Non-randomized tests don't touch the test result when passing.
if (m_current_test_result == TestResult::NotRun)
m_current_test_result = TestResult::Passed;
}
if (repetitions != 1) {
double average = total_time / double(repetitions);
double average_squared = average * average;
double standard_deviation = sqrt((sum_of_squared_times + repetitions * average_squared - 2 * total_time * average) / (repetitions - 1));
dbgln("{} {} '{}' on average in {:.1f}±{:.1f}ms (min={}ms, max={}ms, total={}ms)",
test_result_to_string(m_current_test_result), test_type, t->name(),
average, standard_deviation, min_time, max_time, total_time);
} else {
dbgln("{} {} '{}' in {}ms", test_result_to_string(m_current_test_result), test_type, t->name(), total_time);
}
if (t->is_benchmark()) {
m_benchtime += total_time;
benchmark_count++;
switch (m_current_test_result) {
case TestResult::Passed:
benchmark_passed_count++;
break;
case TestResult::Failed:
benchmark_failed_count++;
break;
default:
break;
}
} else {
m_testtime += total_time;
test_count++;
switch (m_current_test_result) {
case TestResult::Passed:
test_passed_count++;
break;
case TestResult::Failed:
test_failed_count++;
break;
default:
break;
}
}
}
dbgln("Finished {} tests and {} benchmarks in {}ms ({}ms tests, {}ms benchmarks, {}ms other).",
test_count,
benchmark_count,
global_timer.elapsed_milliseconds(),
m_testtime,
m_benchtime,
global_timer.elapsed_milliseconds() - (m_testtime + m_benchtime));
if (test_count != 0) {
if (test_passed_count == test_count) {
dbgln("All {} tests passed.", test_count);
} else if (test_passed_count + test_failed_count == test_count) {
dbgln("Out of {} tests, {} passed and {} failed.", test_count, test_passed_count, test_failed_count);
} else {
dbgln("Out of {} tests, {} passed, {} failed and {} didn't finish for other reasons.", test_count, test_passed_count, test_failed_count, test_count - test_passed_count - test_failed_count);
}
}
if (benchmark_count != 0) {
if (benchmark_passed_count == benchmark_count) {
dbgln("All {} benchmarks passed.", benchmark_count);
} else if (benchmark_passed_count + benchmark_failed_count == benchmark_count) {
dbgln("Out of {} benchmarks, {} passed and {} failed.", benchmark_count, benchmark_passed_count, benchmark_failed_count);
} else {
dbgln("Out of {} benchmarks, {} passed, {} failed and {} didn't finish for other reasons.", benchmark_count, benchmark_passed_count, benchmark_failed_count, benchmark_count - benchmark_passed_count - benchmark_failed_count);
}
}
// We have multiple TestResults, all except for Passed being "bad".
// Let's get a count of them:
return (int)(test_count - test_passed_count + benchmark_count - benchmark_passed_count);
}
} // namespace Test