1# Copyright (C) 2010, 2012 Google Inc. All rights reserved. 2# 3# Redistribution and use in source and binary forms, with or without 4# modification, are permitted provided that the following conditions are 5# met: 6# 7# * Redistributions of source code must retain the above copyright 8# notice, this list of conditions and the following disclaimer. 9# * Redistributions in binary form must reproduce the above 10# copyright notice, this list of conditions and the following disclaimer 11# in the documentation and/or other materials provided with the 12# distribution. 13# * Neither the name of Google Inc. nor the names of its 14# contributors may be used to endorse or promote products derived from 15# this software without specific prior written permission. 16# 17# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 18# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 19# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR 20# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT 21# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 22# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT 23# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 24# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 25# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 26# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE 27# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 28 29"""Package that handles non-debug, non-file output for run-webkit-tests.""" 30 31import math 32import optparse 33 34from webkitpy.tool import grammar 35from webkitpy.layout_tests.models import test_expectations 36from webkitpy.layout_tests.models.test_expectations import TestExpectations, TestExpectationParser 37from webkitpy.layout_tests.views.metered_stream import MeteredStream 38 39 40NUM_SLOW_TESTS_TO_LOG = 10 41 42 43def print_options(): 44 return [ 45 optparse.make_option('-q', '--quiet', action='store_true', default=False, 46 help='run quietly (errors, warnings, and progress only)'), 47 optparse.make_option('--timing', action='store_true', default=False, 48 help='display test times (summary plus per-test w/ --verbose)'), 49 optparse.make_option('-v', '--verbose', action='store_true', default=False, 50 help='print a summarized result for every test (one line per test)'), 51 optparse.make_option('--details', action='store_true', default=False, 52 help='print detailed results for every test'), 53 optparse.make_option('--debug-rwt-logging', action='store_true', default=False, 54 help='print timestamps and debug information for run-webkit-tests itself'), 55 ] 56 57 58class Printer(object): 59 """Class handling all non-debug-logging printing done by run-webkit-tests.""" 60 61 def __init__(self, port, options, regular_output, logger=None): 62 self.num_completed = 0 63 self.num_tests = 0 64 self._port = port 65 self._options = options 66 self._meter = MeteredStream(regular_output, options.debug_rwt_logging, logger=logger, 67 number_of_columns=self._port.host.platform.terminal_width()) 68 self._running_tests = [] 69 self._completed_tests = [] 70 71 def cleanup(self): 72 self._meter.cleanup() 73 74 def __del__(self): 75 self.cleanup() 76 77 def print_config(self, results_directory): 78 self._print_default("Using port '%s'" % self._port.name()) 79 self._print_default("Test configuration: %s" % self._port.test_configuration()) 80 self._print_default("View the test results at file://%s/results.html" % results_directory) 81 82 # FIXME: should these options be in printing_options? 83 if self._options.new_baseline: 84 self._print_default("Placing new baselines in %s" % self._port.baseline_path()) 85 86 fs = self._port.host.filesystem 87 fallback_path = [fs.split(x)[1] for x in self._port.baseline_search_path()] 88 self._print_default("Baseline search path: %s -> generic" % " -> ".join(fallback_path)) 89 90 self._print_default("Using %s build" % self._options.configuration) 91 if self._options.pixel_tests: 92 self._print_default("Pixel tests enabled") 93 else: 94 self._print_default("Pixel tests disabled") 95 96 self._print_default("Regular timeout: %s, slow test timeout: %s" % 97 (self._options.time_out_ms, self._options.slow_time_out_ms)) 98 99 self._print_default('Command line: ' + ' '.join(self._port.driver_cmd_line())) 100 self._print_default('') 101 102 def print_found(self, num_all_test_files, num_to_run, repeat_each, iterations): 103 found_str = 'Found %s; running %d' % (grammar.pluralize('test', num_all_test_files), num_to_run) 104 if repeat_each * iterations > 1: 105 found_str += ' (%d times each: --repeat-each=%d --iterations=%d)' % (repeat_each * iterations, repeat_each, iterations) 106 found_str += ', skipping %d' % (num_all_test_files - num_to_run) 107 self._print_default(found_str + '.') 108 109 def print_expected(self, run_results, tests_with_result_type_callback): 110 self._print_expected_results_of_type(run_results, test_expectations.PASS, "passes", tests_with_result_type_callback) 111 self._print_expected_results_of_type(run_results, test_expectations.FAIL, "failures", tests_with_result_type_callback) 112 self._print_expected_results_of_type(run_results, test_expectations.FLAKY, "flaky", tests_with_result_type_callback) 113 self._print_debug('') 114 115 def print_workers_and_shards(self, num_workers, num_shards, num_locked_shards): 116 driver_name = self._port.driver_name() 117 if num_workers == 1: 118 self._print_default("Running 1 %s." % driver_name) 119 self._print_debug("(%s)." % grammar.pluralize('shard', num_shards)) 120 else: 121 self._print_default("Running %d %ss in parallel." % (num_workers, driver_name)) 122 self._print_debug("(%d shards; %d locked)." % (num_shards, num_locked_shards)) 123 self._print_default('') 124 125 def _print_expected_results_of_type(self, run_results, result_type, result_type_str, tests_with_result_type_callback): 126 tests = tests_with_result_type_callback(result_type) 127 now = run_results.tests_by_timeline[test_expectations.NOW] 128 wontfix = run_results.tests_by_timeline[test_expectations.WONTFIX] 129 130 # We use a fancy format string in order to print the data out in a 131 # nicely-aligned table. 132 fmtstr = ("Expect: %%5d %%-8s (%%%dd now, %%%dd wontfix)" 133 % (self._num_digits(now), self._num_digits(wontfix))) 134 self._print_debug(fmtstr % (len(tests), result_type_str, len(tests & now), len(tests & wontfix))) 135 136 def _num_digits(self, num): 137 ndigits = 1 138 if len(num): 139 ndigits = int(math.log10(len(num))) + 1 140 return ndigits 141 142 def print_results(self, run_time, run_results, summarized_results): 143 self._print_timing_statistics(run_time, run_results) 144 self._print_one_line_summary(run_time, run_results) 145 146 def _print_timing_statistics(self, total_time, run_results): 147 self._print_debug("Test timing:") 148 self._print_debug(" %6.2f total testing time" % total_time) 149 self._print_debug("") 150 151 self._print_worker_statistics(run_results, int(self._options.child_processes)) 152 self._print_aggregate_test_statistics(run_results) 153 self._print_individual_test_times(run_results) 154 self._print_directory_timings(run_results) 155 156 def _print_worker_statistics(self, run_results, num_workers): 157 self._print_debug("Thread timing:") 158 stats = {} 159 cuml_time = 0 160 for result in run_results.results_by_name.values(): 161 stats.setdefault(result.worker_name, {'num_tests': 0, 'total_time': 0}) 162 stats[result.worker_name]['num_tests'] += 1 163 stats[result.worker_name]['total_time'] += result.total_run_time 164 cuml_time += result.total_run_time 165 166 for worker_name in stats: 167 self._print_debug(" %10s: %5d tests, %6.2f secs" % (worker_name, stats[worker_name]['num_tests'], stats[worker_name]['total_time'])) 168 self._print_debug(" %6.2f cumulative, %6.2f optimal" % (cuml_time, cuml_time / num_workers)) 169 self._print_debug("") 170 171 def _print_aggregate_test_statistics(self, run_results): 172 times_for_dump_render_tree = [result.test_run_time for result in run_results.results_by_name.values()] 173 self._print_statistics_for_test_timings("PER TEST TIME IN TESTSHELL (seconds):", times_for_dump_render_tree) 174 175 def _print_individual_test_times(self, run_results): 176 # Reverse-sort by the time spent in the driver. 177 178 individual_test_timings = sorted(run_results.results_by_name.values(), key=lambda result: result.test_run_time, reverse=True) 179 num_printed = 0 180 slow_tests = [] 181 timeout_or_crash_tests = [] 182 unexpected_slow_tests = [] 183 for test_tuple in individual_test_timings: 184 test_name = test_tuple.test_name 185 is_timeout_crash_or_slow = False 186 if test_name in run_results.slow_tests: 187 is_timeout_crash_or_slow = True 188 slow_tests.append(test_tuple) 189 190 if test_name in run_results.failures_by_name: 191 result = run_results.results_by_name[test_name].type 192 if (result == test_expectations.TIMEOUT or 193 result == test_expectations.CRASH): 194 is_timeout_crash_or_slow = True 195 timeout_or_crash_tests.append(test_tuple) 196 197 if (not is_timeout_crash_or_slow and num_printed < NUM_SLOW_TESTS_TO_LOG): 198 num_printed = num_printed + 1 199 unexpected_slow_tests.append(test_tuple) 200 201 self._print_debug("") 202 if unexpected_slow_tests: 203 self._print_test_list_timing("%s slowest tests that are not marked as SLOW and did not timeout/crash:" % 204 NUM_SLOW_TESTS_TO_LOG, unexpected_slow_tests) 205 self._print_debug("") 206 207 if slow_tests: 208 self._print_test_list_timing("Tests marked as SLOW:", slow_tests) 209 self._print_debug("") 210 211 if timeout_or_crash_tests: 212 self._print_test_list_timing("Tests that timed out or crashed:", timeout_or_crash_tests) 213 self._print_debug("") 214 215 def _print_test_list_timing(self, title, test_list): 216 self._print_debug(title) 217 for test_tuple in test_list: 218 test_run_time = round(test_tuple.test_run_time, 1) 219 self._print_debug(" %s took %s seconds" % (test_tuple.test_name, test_run_time)) 220 221 def _print_directory_timings(self, run_results): 222 stats = {} 223 for result in run_results.results_by_name.values(): 224 stats.setdefault(result.shard_name, {'num_tests': 0, 'total_time': 0}) 225 stats[result.shard_name]['num_tests'] += 1 226 stats[result.shard_name]['total_time'] += result.total_run_time 227 228 min_seconds_to_print = 15 229 230 timings = [] 231 for directory in stats: 232 rounded_time = round(stats[directory]['total_time'], 1) 233 if rounded_time > min_seconds_to_print: 234 timings.append((directory, rounded_time, stats[directory]['num_tests'])) 235 236 if not timings: 237 return 238 239 timings.sort() 240 241 self._print_debug("Time to process slowest subdirectories:") 242 for timing in timings: 243 self._print_debug(" %s took %s seconds to run %s tests." % timing) 244 self._print_debug("") 245 246 def _print_statistics_for_test_timings(self, title, timings): 247 self._print_debug(title) 248 timings.sort() 249 250 num_tests = len(timings) 251 if not num_tests: 252 return 253 percentile90 = timings[int(.9 * num_tests)] 254 percentile99 = timings[int(.99 * num_tests)] 255 256 if num_tests % 2 == 1: 257 median = timings[((num_tests - 1) / 2) - 1] 258 else: 259 lower = timings[num_tests / 2 - 1] 260 upper = timings[num_tests / 2] 261 median = (float(lower + upper)) / 2 262 263 mean = sum(timings) / num_tests 264 265 for timing in timings: 266 sum_of_deviations = math.pow(timing - mean, 2) 267 268 std_deviation = math.sqrt(sum_of_deviations / num_tests) 269 self._print_debug(" Median: %6.3f" % median) 270 self._print_debug(" Mean: %6.3f" % mean) 271 self._print_debug(" 90th percentile: %6.3f" % percentile90) 272 self._print_debug(" 99th percentile: %6.3f" % percentile99) 273 self._print_debug(" Standard dev: %6.3f" % std_deviation) 274 self._print_debug("") 275 276 def _print_one_line_summary(self, total_time, run_results): 277 if self._options.timing: 278 parallel_time = sum(result.total_run_time for result in run_results.results_by_name.values()) 279 280 # There is serial overhead in layout_test_runner.run() that we can't easily account for when 281 # really running in parallel, but taking the min() ensures that in the worst case 282 # (if parallel time is less than run_time) we do account for it. 283 serial_time = total_time - min(run_results.run_time, parallel_time) 284 285 speedup = (parallel_time + serial_time) / total_time 286 timing_summary = ' in %.2fs (%.2fs in rwt, %.2gx)' % (total_time, serial_time, speedup) 287 else: 288 timing_summary = '' 289 290 total = run_results.total - run_results.expected_skips 291 expected = run_results.expected - run_results.expected_skips 292 unexpected = run_results.unexpected 293 incomplete = total - expected - unexpected 294 incomplete_str = '' 295 if incomplete: 296 self._print_default("") 297 incomplete_str = " (%d didn't run)" % incomplete 298 299 if self._options.verbose or self._options.debug_rwt_logging or unexpected: 300 self.writeln("") 301 302 expected_summary_str = '' 303 if run_results.expected_failures > 0: 304 expected_summary_str = " (%d passed, %d didn't)" % (expected - run_results.expected_failures, run_results.expected_failures) 305 306 summary = '' 307 if unexpected == 0: 308 if expected == total: 309 if expected > 1: 310 summary = "All %d tests ran as expected%s%s." % (expected, expected_summary_str, timing_summary) 311 else: 312 summary = "The test ran as expected%s%s." % (expected_summary_str, timing_summary) 313 else: 314 summary = "%s ran as expected%s%s%s." % (grammar.pluralize('test', expected), expected_summary_str, incomplete_str, timing_summary) 315 else: 316 summary = "%s ran as expected%s, %d didn't%s%s:" % (grammar.pluralize('test', expected), expected_summary_str, unexpected, incomplete_str, timing_summary) 317 318 self._print_quiet(summary) 319 self._print_quiet("") 320 321 def _test_status_line(self, test_name, suffix): 322 format_string = '[%d/%d] %s%s' 323 status_line = format_string % (self.num_completed, self.num_tests, test_name, suffix) 324 if len(status_line) > self._meter.number_of_columns(): 325 overflow_columns = len(status_line) - self._meter.number_of_columns() 326 ellipsis = '...' 327 if len(test_name) < overflow_columns + len(ellipsis) + 2: 328 # We don't have enough space even if we elide, just show the test filename. 329 fs = self._port.host.filesystem 330 test_name = fs.split(test_name)[1] 331 else: 332 new_length = len(test_name) - overflow_columns - len(ellipsis) 333 prefix = int(new_length / 2) 334 test_name = test_name[:prefix] + ellipsis + test_name[-(new_length - prefix):] 335 return format_string % (self.num_completed, self.num_tests, test_name, suffix) 336 337 def print_started_test(self, test_name): 338 self._running_tests.append(test_name) 339 if len(self._running_tests) > 1: 340 suffix = ' (+%d)' % (len(self._running_tests) - 1) 341 else: 342 suffix = '' 343 if self._options.verbose: 344 write = self._meter.write_update 345 else: 346 write = self._meter.write_throttled_update 347 write(self._test_status_line(test_name, suffix)) 348 349 def print_finished_test(self, result, expected, exp_str, got_str): 350 self.num_completed += 1 351 test_name = result.test_name 352 353 result_message = self._result_message(result.type, result.failures, expected, 354 self._options.timing, result.test_run_time) 355 356 if self._options.details: 357 self._print_test_trace(result, exp_str, got_str) 358 elif self._options.verbose or not expected: 359 self.writeln(self._test_status_line(test_name, result_message)) 360 elif self.num_completed == self.num_tests: 361 self._meter.write_update('') 362 else: 363 if test_name == self._running_tests[0]: 364 self._completed_tests.insert(0, [test_name, result_message]) 365 else: 366 self._completed_tests.append([test_name, result_message]) 367 368 for test_name, result_message in self._completed_tests: 369 self._meter.write_throttled_update(self._test_status_line(test_name, result_message)) 370 self._completed_tests = [] 371 self._running_tests.remove(test_name) 372 373 def _result_message(self, result_type, failures, expected, timing, test_run_time): 374 exp_string = ' unexpectedly' if not expected else '' 375 timing_string = ' %.4fs' % test_run_time if timing else '' 376 if result_type == test_expectations.PASS: 377 return ' passed%s%s' % (exp_string, timing_string) 378 else: 379 return ' failed%s (%s)%s' % (exp_string, ', '.join(failure.message() for failure in failures), timing_string) 380 381 def _print_test_trace(self, result, exp_str, got_str): 382 test_name = result.test_name 383 self._print_default(self._test_status_line(test_name, '')) 384 385 base = self._port.lookup_virtual_test_base(test_name) 386 if base: 387 args = ' '.join(self._port.lookup_virtual_test_args(test_name)) 388 self._print_default(' base: %s' % base) 389 self._print_default(' args: %s' % args) 390 391 references = self._port.reference_files(test_name) 392 if references: 393 for _, filename in references: 394 self._print_default(' ref: %s' % self._port.relative_test_filename(filename)) 395 else: 396 for extension in ('.txt', '.png', '.wav'): 397 self._print_baseline(test_name, extension) 398 399 self._print_default(' exp: %s' % exp_str) 400 self._print_default(' got: %s' % got_str) 401 self._print_default(' took: %-.3f' % result.test_run_time) 402 self._print_default('') 403 404 def _print_baseline(self, test_name, extension): 405 baseline = self._port.expected_filename(test_name, extension) 406 if self._port._filesystem.exists(baseline): 407 relpath = self._port.relative_test_filename(baseline) 408 else: 409 relpath = '<none>' 410 self._print_default(' %s: %s' % (extension[1:], relpath)) 411 412 def _print_quiet(self, msg): 413 self.writeln(msg) 414 415 def _print_default(self, msg): 416 if not self._options.quiet: 417 self.writeln(msg) 418 419 def _print_debug(self, msg): 420 if self._options.debug_rwt_logging: 421 self.writeln(msg) 422 423 def write_throttled_update(self, msg): 424 self._meter.write_throttled_update(msg) 425 426 def write_update(self, msg): 427 self._meter.write_update(msg) 428 429 def writeln(self, msg): 430 self._meter.writeln(msg) 431 432 def flush(self): 433 self._meter.flush() 434