commit 6f696e98c155434d6ac95806052eb634d36e1e8e Author: Damian Johnson atagar@torproject.org Date: Sat Jan 4 15:14:31 2014 -0800

Defaulting to provide less verbose test output

Our test output is pretty sweet, but as we've added more tests it's gotten to be pretty overwhelming. This is especially troublesome for the RUN_ALL target which our jenkins test use.

Dropping output on individual tests unless there's a failure. Instead, we now just print the test modules and their runtime. Users can get the more verbose output by including '--verbose' in their test run. --- run_tests.py | 44 +++++++++++++++++++++++++++++++++++--------- test/settings.cfg | 1 + 2 files changed, 36 insertions(+), 9 deletions(-)

diff --git a/run_tests.py b/run_tests.py index 9559311..664df35 100755 --- a/run_tests.py +++ b/run_tests.py @@ -25,7 +25,7 @@ import test.output import test.runner import test.util

-from test.output import STATUS, SUCCESS, ERROR, println +from test.output import STATUS, SUCCESS, ERROR, NO_NL, println from test.util import STEM_BASE, Target, Task

# Our default arguments. The _get_args() function provides a named tuple of @@ -50,11 +50,12 @@ ARGS = { 'tor_path': 'tor', 'run_targets': [Target.RUN_OPEN], 'attribute_targets': [], + 'verbose': False, 'print_help': False, }

-OPT = "auit:l:h" -OPT_EXPANDED = ["all", "unit", "integ", "python3", "clean", "targets=", "test=", "log=", "tor=", "help"] +OPT = "auit:l:vh" +OPT_EXPANDED = ["all", "unit", "integ", "python3", "clean", "targets=", "test=", "log=", "tor=", "verbose", "help"]

CONFIG = stem.util.conf.config_dict("test", { "target.torrc": {}, @@ -212,7 +213,7 @@ def main(): error_tracker.set_category("UNIT TEST")

for test_class in test.util.get_unit_tests(args.test_prefix): - run_result = _run_test(test_class, output_filters, logging_buffer) + run_result = _run_test(args, test_class, output_filters, logging_buffer) skipped_tests += len(getattr(run_result, 'skipped', []))

println() @@ -248,7 +249,7 @@ def main(): owner = integ_runner.get_tor_controller(True) # controller to own our main Tor process

for test_class in test.util.get_integ_tests(args.test_prefix): - run_result = _run_test(test_class, output_filters, logging_buffer) + run_result = _run_test(args, test_class, output_filters, logging_buffer) skipped_tests += len(getattr(run_result, 'skipped', []))

if owner: @@ -384,6 +385,8 @@ def _get_args(argv): args['logging_runlevel'] = arg elif opt in ("--tor"): args['tor_path'] = arg + elif opt in ("-v", "--verbose"): + args['verbose'] = True elif opt in ("-h", "--help"): args['print_help'] = True

@@ -416,15 +419,38 @@ def _print_static_issues(static_check_issues): println()

-def _run_test(test_class, output_filters, logging_buffer): - test.output.print_divider(test_class.__module__) +def _run_test(args, test_class, output_filters, logging_buffer): + start_time = time.time() + + if args.verbose: + test.output.print_divider(test_class.__module__) + else: + label = test_class.__module__ + + if label.startswith('test.unit.'): + label = label[10:] + elif label.startswith('test.integ.'): + label = label[11:] + + label = " %s..." % label + label = "%-54s" % label + + println(label, STATUS, NO_NL) + suite = unittest.TestLoader().loadTestsFromTestCase(test_class)

test_results = StringIO.StringIO() run_result = unittest.TextTestRunner(test_results, verbosity=2).run(suite)

- sys.stdout.write(test.output.apply_filters(test_results.getvalue(), *output_filters)) - println() + if args.verbose: + sys.stdout.write(test.output.apply_filters(test_results.getvalue(), *output_filters)) + println() + elif not run_result.failures: + println(" success (%0.2fs)" % (time.time() - start_time), SUCCESS) + else: + println(" failed (%0.2fs)" % (time.time() - start_time), ERROR) + sys.stdout.write(test.output.apply_filters(test_results.getvalue(), *output_filters)) + test.output.print_logging(logging_buffer)

return run_result diff --git a/test/settings.cfg b/test/settings.cfg index 6fd436a..e78ddb4 100644 --- a/test/settings.cfg +++ b/test/settings.cfg @@ -41,6 +41,7 @@ msg.help | -l, --log RUNLEVEL includes logging output with test results, runlevels: | TRACE, DEBUG, INFO, NOTICE, WARN, ERROR | --tor PATH custom tor binary to run testing against +| -v, --verbose provides additional test output | -h, --help presents this help | | Examples: