[test] Check output on the worker process.

I added additional exception logging in the execution.py since
errors in processing results were really difficult to debug.

There is a problem on Windows with class serialization when
it comes from dynamically loaded module. To fix it I moved all
output processors to the tools/testrunner/outproc/ and import
them in test/*/testcfg.py.

Bug: v8:6917
Change-Id: Ida604641d659b006e91faf1d56a37769ec47f5f3
Reviewed-on: https://chromium-review.googlesource.com/842784
Commit-Queue: Michał Majewski <majeski@google.com>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#50361}

test/inspector/testcfg.py

@@ -6,8 +6,8 @@ import os
 
 from testrunner.local import testsuite
 from testrunner.local import utils
-from testrunner.objects import outproc
 from testrunner.objects import testcase
+from testrunner.outproc import base as outproc
 
 PROTOCOL_TEST_JS = "protocol-test.js"
 EXPECTED_SUFFIX = "-expected.txt"

test/message/testcfg.py

@@ -25,14 +25,13 @@
 # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
-import itertools
 import os
 import re
 
 from testrunner.local import testsuite
 from testrunner.local import utils
-from testrunner.objects import outproc
 from testrunner.objects import testcase
+from testrunner.outproc import message
 
 
 INVALID_FLAGS = ["--enable-slow-asserts"]
@@ -103,56 +102,9 @@ class TestCase(testcase.TestCase):
 
   @property
   def output_proc(self):
-    return OutProc(self.expected_outcomes,
-                   os.path.join(self.suite.root, self.path),
-                   self._expected_fail())
-
-
-class OutProc(outproc.OutProc):
-  def __init__(self, expected_outcomes, basepath, expected_fail):
-    super(OutProc, self).__init__(expected_outcomes)
-    self._basepath = basepath
-    self._expected_fail = expected_fail
-
-  def _is_failure_output(self, output):
-    fail = output.exit_code != 0
-    if fail != self._expected_fail:
-      return True
-
-    expected_lines = []
-    # Can't use utils.ReadLinesFrom() here because it strips whitespace.
-    with open(self._basepath + '.out') as f:
-      for line in f:
-        if line.startswith("#") or not line.strip():
-          continue
-        expected_lines.append(line)
-    raw_lines = output.stdout.splitlines()
-    actual_lines = [ s for s in raw_lines if not self._ignore_line(s) ]
-    if len(expected_lines) != len(actual_lines):
-      return True
-
-    env = {
-      'basename': os.path.basename(self._basepath + '.js'),
-    }
-    for (expected, actual) in itertools.izip_longest(
-        expected_lines, actual_lines, fillvalue=''):
-      pattern = re.escape(expected.rstrip() % env)
-      pattern = pattern.replace('\\*', '.*')
-      pattern = pattern.replace('\\{NUMBER\\}', '\d+(?:\.\d*)?')
-      pattern = '^%s$' % pattern
-      if not re.match(pattern, actual):
-        return True
-    return False
-
-  def _ignore_line(self, string):
-    """Ignore empty lines, valgrind output, Android output."""
-    return (
-      not string or
-      not string.strip() or
-      string.startswith("==") or
-      string.startswith("**") or
-      string.startswith("ANDROID")
-    )
+    return message.OutProc(self.expected_outcomes,
+                           os.path.join(self.suite.root, self.path),
+                           self._expected_fail())
 
 
 def GetSuite(name, root):

test/mkgrokdump/testcfg.py

@@ -3,11 +3,10 @@
 # found in the LICENSE file.
 
 import os
-import difflib
 
 from testrunner.local import testsuite
-from testrunner.objects import outproc
 from testrunner.objects import testcase
+from testrunner.outproc import mkgrokdump
 
 
 SHELL = 'mkgrokdump'
@@ -42,31 +41,7 @@ class TestCase(testcase.TestCase):
 
   @property
   def output_proc(self):
-    return OutProc(self.expected_outcomes, self.suite.expected_path)
-
-
-class OutProc(outproc.OutProc):
-  def __init__(self, expected_outcomes, expected_path):
-    super(OutProc, self).__init__(expected_outcomes)
-    self._expected_path = expected_path
-
-  def _is_failure_output(self, output):
-    with open(self._expected_path) as f:
-      expected = f.read()
-    expected_lines = expected.splitlines()
-    actual_lines = output.stdout.splitlines()
-    diff = difflib.unified_diff(expected_lines, actual_lines, lineterm="",
-                                fromfile="expected_path")
-    diffstring = '\n'.join(diff)
-    if diffstring is not "":
-      if "generated from a non-shipping build" in output.stdout:
-        return False
-      if not "generated from a shipping build" in output.stdout:
-        output.stdout = "Unexpected output:\n\n" + output.stdout
-        return True
-      output.stdout = diffstring
-      return True
-    return False
+    return mkgrokdump.OutProc(self.expected_outcomes, self.suite.expected_path)
 
 
 def GetSuite(name, root):

test/mozilla/testcfg.py

@@ -29,8 +29,8 @@
 import os
 
 from testrunner.local import testsuite
-from testrunner.objects import outproc
 from testrunner.objects import testcase
+from testrunner.outproc import mozilla
 
 EXCLUDED = ["CVS", ".svn"]
 
@@ -110,37 +110,13 @@ class TestCase(testcase.TestCase):
   def output_proc(self):
     if not self.expected_outcomes:
       if self.path.endswith('-n'):
-        return MOZILLA_PASS_NEGATIVE
-      return MOZILLA_PASS_DEFAULT
+        return mozilla.MOZILLA_PASS_NEGATIVE
+      return mozilla.MOZILLA_PASS_DEFAULT
     if self.path.endswith('-n'):
-      return NegOutProc(self.expected_outcomes)
-    return OutProc(self.expected_outcomes)
+      return mozilla.NegOutProc(self.expected_outcomes)
+    return mozilla.OutProc(self.expected_outcomes)
 
 
-def _is_failure_output(self, output):
-  return (
-    output.exit_code != 0 or
-    'FAILED!' in output.stdout
-  )
-
-
-class OutProc(outproc.OutProc):
-  """Optimized for positive tests."""
-OutProc._is_failure_output = _is_failure_output
-
-
-class PassOutProc(outproc.PassOutProc):
-  """Optimized for positive tests expected to PASS."""
-PassOutProc._is_failure_output = _is_failure_output
-
-
-NegOutProc = outproc.negative(OutProc)
-NegPassOutProc = outproc.negative(PassOutProc)
-
-
-MOZILLA_PASS_DEFAULT = PassOutProc()
-MOZILLA_PASS_NEGATIVE = NegPassOutProc()
-
 
 def GetSuite(name, root):
   return TestSuite(name, root)

test/test262/testcfg.py

@@ -33,12 +33,13 @@ import re
 import sys
 import tarfile
 
-
 from testrunner.local import statusfile
 from testrunner.local import testsuite
 from testrunner.local import utils
-from testrunner.objects import outproc
 from testrunner.objects import testcase
+from testrunner.outproc import base as outproc
+from testrunner.outproc import test262
+
 
 # TODO(littledan): move the flag mapping into the status file
 FEATURE_FLAGS = {
@@ -246,57 +247,13 @@ class TestCase(testcase.TestCase):
   @property
   def output_proc(self):
     if self._expected_exception is not None:
-      return ExceptionOutProc(self.expected_outcomes, self._expected_exception)
+      return test262.ExceptionOutProc(self.expected_outcomes,
+                                      self._expected_exception)
     if self.expected_outcomes == outproc.OUTCOMES_PASS:
-      return PASS_NO_EXCEPTION
-    return NoExceptionOutProc(self.expected_outcomes)
+      return test262.PASS_NO_EXCEPTION
+    return test262.NoExceptionOutProc(self.expected_outcomes)
 
 
-class ExceptionOutProc(outproc.OutProc):
-  """Output processor for tests with expected exception."""
-  def __init__(self, expected_outcomes, expected_exception=None):
-    super(ExceptionOutProc, self).__init__(expected_outcomes)
-    self._expected_exception = expected_exception
-
-  def _is_failure_output(self, output):
-    if output.exit_code != 0:
-      return True
-    if self._expected_exception != self._parse_exception(output.stdout):
-      return True
-    return 'FAILED!' in output.stdout
-
-  def _parse_exception(self, string):
-    # somefile:somelinenumber: someerror[: sometext]
-    # somefile might include an optional drive letter on windows e.g. "e:".
-    match = re.search(
-        '^(?:\w:)?[^:]*:[0-9]+: ([^: ]+?)($|: )', string, re.MULTILINE)
-    if match:
-      return match.group(1).strip()
-    else:
-      return None
-
-
-def _is_failure_output(self, output):
-  return (
-    output.exit_code != 0 or
-    'FAILED!' in output.stdout
-  )
-
-
-class NoExceptionOutProc(outproc.OutProc):
-  """Output processor optimized for tests without expected exception."""
-NoExceptionOutProc._is_failure_output = _is_failure_output
-
-
-class PassNoExceptionOutProc(outproc.PassOutProc):
-  """
-  Output processor optimized for tests expected to PASS without expected
-  exception.
-  """
-PassNoExceptionOutProc._is_failure_output = _is_failure_output
-
-
-PASS_NO_EXCEPTION = PassNoExceptionOutProc()
 
 
 def GetSuite(name, root):

test/webkit/testcfg.py

@@ -29,8 +29,8 @@ import os
 import re
 
 from testrunner.local import testsuite
-from testrunner.objects import outproc
 from testrunner.objects import testcase
+from testrunner.outproc import webkit
 
 FILES_PATTERN = re.compile(r"//\s+Files:(.*)")
 SELF_SCRIPT_PATTERN = re.compile(r"//\s+Env: TEST_FILE_NAME")
@@ -104,23 +104,10 @@ class TestCase(testcase.TestCase):
 
   @property
   def output_proc(self):
-    return OutProc(
+    return webkit.OutProc(
         self.expected_outcomes,
         os.path.join(self.suite.root, self.path) + '-expected.txt')
 
 
-class OutProc(outproc.ExpectedOutProc):
-  def _is_failure_output(self, output):
-    if output.exit_code != 0:
-      return True
-    return super(OutProc, self)._is_failure_output(output)
-
-  def _ignore_expected_line(self, line):
-    return (
-        line.startswith('#') or
-        super(OutProc, self)._ignore_expected_line(line)
-    )
-
-
 def GetSuite(name, root):
   return TestSuite(name, root)

tools/testrunner/deopt_fuzzer.py

@@ -264,7 +264,6 @@ class DeoptFuzzer(base_runner.BaseTestRunner):
                           True,  # No sorting of test cases.
                           0,  # Don't rerun failing tests.
                           0,  # No use of a rerun-failing-tests maximum.
-                          False,  # No predictable mode.
                           False,  # No no_harness mode.
                           False,  # Don't use perf data.
                           False)  # Coverage not supported.

tools/testrunner/gc_fuzzer.py

@@ -179,7 +179,6 @@ class GCFuzzer(base_runner.BaseTestRunner):
                           True,  # No sorting of test cases.
                           0,  # Don't rerun failing tests.
                           0,  # No use of a rerun-failing-tests maximum.
-                          False,  # No predictable mode.
                           False,  # No no_harness mode.
                           False,  # Don't use perf data.
                           False)  # Coverage not supported.
@@ -213,7 +212,7 @@ class GCFuzzer(base_runner.BaseTestRunner):
     for s in suites:
       for t in s.tests:
         # Skip failed tests.
-        if s.HasUnexpectedOutput(t, runner.outputs[t]):
+        if t.output_proc.has_unexpected_output(runner.outputs[t]):
           print '%s failed, skipping' % t.path
           continue
         max_limit = self._get_max_limit_reached(runner.outputs[t])

tools/testrunner/local/execution.py

@@ -31,6 +31,7 @@ import os
 import re
 import shutil
 import sys
+import traceback
 
 from . import command
 from . import perfdata
@@ -51,7 +52,7 @@ ProcessContext = collections.namedtuple(
 
 
 TestJobResult = collections.namedtuple(
-    'TestJobResult', ['id', 'output'])
+    'TestJobResult', ['id', 'outproc_result'])
 
 def MakeProcessContext(sancov_dir):
   return ProcessContext(sancov_dir)
@@ -74,9 +75,10 @@ class Job(object):
 
 
 class TestJob(Job):
-  def __init__(self, test_id, cmd, run_num):
+  def __init__(self, test_id, cmd, outproc, run_num):
     self.test_id = test_id
     self.cmd = cmd
+    self.outproc = outproc
     self.run_num = run_num
 
   def _rename_coverage_data(self, out, sancov_dir):
@@ -105,20 +107,21 @@ class TestJob(Job):
         os.rename(sancov_file, new_sancov_file)
 
   def run(self, context):
-    out = self.cmd.execute()
-    self._rename_coverage_data(out, context.sancov_dir)
-    return TestJobResult(self.test_id, out)
+    output = self.cmd.execute()
+    self._rename_coverage_data(output, context.sancov_dir)
+    return TestJobResult(self.test_id, self.outproc.process(output))
 
 
 class Runner(object):
 
-  def __init__(self, suites, progress_indicator, context):
+  def __init__(self, suites, progress_indicator, context, outproc_factory=None):
     self.datapath = os.path.join("out", "testrunner_data")
     self.perf_data_manager = perfdata.GetPerfDataManager(
         context, self.datapath)
     self.perfdata = self.perf_data_manager.GetStore(context.arch, context.mode)
     self.perf_failures = False
     self.printed_allocations = False
+    self.outproc_factory = outproc_factory or (lambda test: test.output_proc)
     self.tests = [t for s in suites for t in s.tests]
 
     # TODO(majeski): Pass dynamically instead of keeping them in the runner.
@@ -159,7 +162,7 @@ class Runner(object):
       print("PerfData exception: %s" % e)
       self.perf_failures = True
 
-  def _MaybeRerun(self, pool, test, job_result):
+  def _MaybeRerun(self, pool, test, result):
     if test.run <= self.context.rerun_failures_count:
       # Possibly rerun this test if its run count is below the maximum per
       # test. <= as the flag controls reruns not including the first run.
@@ -172,23 +175,24 @@ class Runner(object):
           # reached.
           return
       if (test.run >= 2 and
-          job_result.output.duration > self.context.timeout / 20.0):
+          result.output.duration > self.context.timeout / 20.0):
         # Rerun slow tests at most once.
         return
 
       # Rerun this test.
       test.run += 1
-      pool.add([TestJob(test.id, test.cmd, test.run)])
+      pool.add([
+          TestJob(test.id, test.cmd, self.outproc_factory(test), test.run)
+      ])
       self.remaining += 1
       self.total += 1
 
-  def _ProcessTest(self, test, job_result, pool):
-    self.outputs[test] = job_result.output
-    has_unexpected_output = test.suite.HasUnexpectedOutput(
-        test, job_result.output, self.context)
+  def _ProcessTest(self, test, result, pool):
+    self.outputs[test] = result.output
+    has_unexpected_output = result.has_unexpected_output
     if has_unexpected_output:
       self.failed.append(test)
-      if job_result.output.HasCrashed():
+      if result.output.HasCrashed():
         self.crashed += 1
     else:
       self.succeeded += 1
@@ -196,12 +200,12 @@ class Runner(object):
     # For the indicator, everything that happens after the first run is treated
     # as unexpected even if it flakily passes in order to include it in the
     # output.
-    self.indicator.HasRun(test, job_result.output,
+    self.indicator.HasRun(test, result.output,
                           has_unexpected_output or test.run > 1)
     if has_unexpected_output:
       # Rerun test failures after the indicator has processed the results.
       self._VerbosePrint("Attempting to rerun test after failure.")
-      self._MaybeRerun(pool, test, job_result)
+      self._MaybeRerun(pool, test, result)
     # Update the perf database if the test succeeded.
     return not has_unexpected_output
 
@@ -224,11 +228,13 @@ class Runner(object):
         assert test.id >= 0
         test_map[test.id] = test
         try:
-          yield [TestJob(test.id, test.cmd, test.run)]
+          yield [
+              TestJob(test.id, test.cmd, self.outproc_factory(test), test.run)
+          ]
         except Exception, e:
           # If this failed, save the exception and re-raise it later (after
           # all other tests have had a chance to run).
-          queued_exception[0] = e
+          queued_exception[0] = e, traceback.format_exc()
           continue
     try:
       it = pool.imap_unordered(
@@ -243,12 +249,19 @@ class Runner(object):
           continue
 
         job_result = result.value
+        test_id = job_result.id
+        outproc_result = job_result.outproc_result
 
-        test = test_map[job_result.id]
-        update_perf = self._ProcessTest(test, job_result, pool)
+        test = test_map[test_id]
+        update_perf = self._ProcessTest(test, outproc_result, pool)
         if update_perf:
           self._RunPerfSafe(lambda: self.perfdata.UpdatePerfData(
-              test, job_result.output.duration))
+              test, outproc_result.output.duration))
+    except KeyboardInterrupt:
+      raise
+    except:
+      traceback.print_exc()
+      raise
     finally:
       self._VerbosePrint("Closing process pool.")
       pool.terminate()
@@ -260,7 +273,9 @@ class Runner(object):
         print "Deleting perf test data due to db corruption."
         shutil.rmtree(self.datapath)
     if queued_exception[0]:
-      raise queued_exception[0]
+      e, stacktrace = queued_exception[0]
+      print stacktrace
+      raise e
 
   def _VerbosePrint(self, text):
     if self.context.verbose:

tools/testrunner/local/testsuite.py

@@ -171,20 +171,6 @@ class TestSuite(object):
           break
     self.tests = filtered
 
-  def HasUnexpectedOutput(self, test, output, ctx=None):
-    if ctx and ctx.predictable:
-      # Only check the exit code of the predictable_wrapper in
-      # verify-predictable mode. Negative tests are not supported as they
-      # usually also don't print allocation hashes. There are two versions of
-      # negative tests: one specified by the test, the other specified through
-      # the status file (e.g. known bugs).
-      return (
-          output.exit_code != 0 and
-          not test.output_proc.negative and
-          statusfile.FAIL not in test.expected_outcomes
-      )
-    return test.output_proc.has_unexpected_output(output)
-
   def _create_test(self, path, **kwargs):
     test = self._test_class()(self, path, self._path_to_name(path), **kwargs)
     return test

tools/testrunner/objects/context.py

@@ -29,8 +29,8 @@
 class Context():
   def __init__(self, arch, mode, shell_dir, mode_flags, verbose, timeout,
                isolates, command_prefix, extra_flags, noi18n, random_seed,
-               no_sorting, rerun_failures_count, rerun_failures_max,
-               predictable, no_harness, use_perf_data, sancov_dir):
+               no_sorting, rerun_failures_count, rerun_failures_max, no_harness,
+               use_perf_data, sancov_dir):
     self.arch = arch
     self.mode = mode
     self.shell_dir = shell_dir
@@ -45,7 +45,6 @@ class Context():
     self.no_sorting = no_sorting
     self.rerun_failures_count = rerun_failures_count
     self.rerun_failures_max = rerun_failures_max
-    self.predictable = predictable
     self.no_harness = no_harness
     self.use_perf_data = use_perf_data
     self.sancov_dir = sancov_dir

tools/testrunner/objects/predictable.py

@@ -0,0 +1,56 @@
+# Copyright 2017 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.
+
+from ..local import statusfile
+from ..outproc import base as outproc_base
+
+
+# Only check the exit code of the predictable_wrapper in
+# verify-predictable mode. Negative tests are not supported as they
+# usually also don't print allocation hashes. There are two versions of
+# negative tests: one specified by the test, the other specified through
+# the status file (e.g. known bugs).
+
+
+def get_outproc(test):
+  output_proc = test.output_proc
+  if output_proc.negative or statusfile.FAIL in test.expected_outcomes:
+    # TODO(majeski): Skip these tests instead of having special outproc.
+    return NeverUnexpectedOutputOutProc(output_proc)
+  return OutProc(output_proc)
+
+
+class OutProc(outproc_base.BaseOutProc):
+  """Output processor wrapper for predictable mode. It has custom
+  has_unexpected_output implementation, but for all other methods it simply
+  calls wrapped output processor.
+  """
+  def __init__(self, _outproc):
+    super(OutProc, self).__init__()
+    self._outproc = _outproc
+
+  def process(self, output):
+    return self._outproc.process(output)
+
+  def has_unexpected_output(self, output):
+    return output.exit_code != 0
+
+  def get_outcome(self, output):
+    return self._outproc.get_outcome(output)
+
+  @property
+  def negative(self):
+    return self._outproc.negative
+
+  @property
+  def expected_outcomes(self):
+    return self._outproc.expected_outcomes
+
+
+class NeverUnexpectedOutputOutProc(OutProc):
+  """Output processor wrapper for tests that we will return False for
+  has_unexpected_output in the predictable mode.
+  """
+  def has_unexpected_output(self, output):
+    return False

tools/testrunner/objects/testcase.py

@@ -30,7 +30,7 @@ import os
 import re
 import shlex
 
-from . import outproc
+from ..outproc import base as outproc
 from ..local import command
 from ..local import statusfile
 from ..local import utils

tools/testrunner/outproc/__init__.py

@@ -0,0 +1,3 @@
+# 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.

tools/testrunner/outproc/base.py

@@ -2,6 +2,7 @@
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
+import collections
 import itertools
 
 from ..local import statusfile
@@ -10,8 +11,13 @@ from ..local import statusfile
 OUTCOMES_PASS = [statusfile.PASS]
 OUTCOMES_FAIL = [statusfile.FAIL]
 
+Result = collections.namedtuple('Result', ['has_unexpected_output', 'output'])
+
 
 class BaseOutProc(object):
+  def process(self, output):
+    return Result(self.has_unexpected_output(output), output)
+
   def has_unexpected_output(self, output):
     return self.get_outcome(output) not in self.expected_outcomes
 
@@ -43,12 +49,10 @@ class BaseOutProc(object):
     raise NotImplementedError()
 
 
-def negative(cls):
-  class Neg(cls):
-    @property
-    def negative(self):
-      return True
-  return Neg
+class Negative(object):
+  @property
+  def negative(self):
+    return True
 
 
 class PassOutProc(BaseOutProc):

tools/testrunner/outproc/message.py

@@ -0,0 +1,56 @@
+# 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.
+
+import itertools
+import os
+import re
+
+from . import base
+
+
+class OutProc(base.OutProc):
+  def __init__(self, expected_outcomes, basepath, expected_fail):
+    super(OutProc, self).__init__(expected_outcomes)
+    self._basepath = basepath
+    self._expected_fail = expected_fail
+
+  def _is_failure_output(self, output):
+    fail = output.exit_code != 0
+    if fail != self._expected_fail:
+      return True
+
+    expected_lines = []
+    # Can't use utils.ReadLinesFrom() here because it strips whitespace.
+    with open(self._basepath + '.out') as f:
+      for line in f:
+        if line.startswith("#") or not line.strip():
+          continue
+        expected_lines.append(line)
+    raw_lines = output.stdout.splitlines()
+    actual_lines = [ s for s in raw_lines if not self._ignore_line(s) ]
+    if len(expected_lines) != len(actual_lines):
+      return True
+
+    env = {
+      'basename': os.path.basename(self._basepath + '.js'),
+    }
+    for (expected, actual) in itertools.izip_longest(
+        expected_lines, actual_lines, fillvalue=''):
+      pattern = re.escape(expected.rstrip() % env)
+      pattern = pattern.replace('\\*', '.*')
+      pattern = pattern.replace('\\{NUMBER\\}', '\d+(?:\.\d*)?')
+      pattern = '^%s$' % pattern
+      if not re.match(pattern, actual):
+        return True
+    return False
+
+  def _ignore_line(self, string):
+    """Ignore empty lines, valgrind output, Android output."""
+    return (
+      not string or
+      not string.strip() or
+      string.startswith("==") or
+      string.startswith("**") or
+      string.startswith("ANDROID")
+    )

tools/testrunner/outproc/mkgrokdump.py

@@ -0,0 +1,31 @@
+# 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.
+
+import difflib
+
+from . import base
+
+
+class OutProc(base.OutProc):
+  def __init__(self, expected_outcomes, expected_path):
+    super(OutProc, self).__init__(expected_outcomes)
+    self._expected_path = expected_path
+
+  def _is_failure_output(self, output):
+    with open(self._expected_path) as f:
+      expected = f.read()
+    expected_lines = expected.splitlines()
+    actual_lines = output.stdout.splitlines()
+    diff = difflib.unified_diff(expected_lines, actual_lines, lineterm="",
+                                fromfile="expected_path")
+    diffstring = '\n'.join(diff)
+    if diffstring is not "":
+      if "generated from a non-shipping build" in output.stdout:
+        return False
+      if not "generated from a shipping build" in output.stdout:
+        output.stdout = "Unexpected output:\n\n" + output.stdout
+        return True
+      output.stdout = diffstring
+      return True
+    return False

tools/testrunner/outproc/mozilla.py

@@ -0,0 +1,33 @@
+# 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.
+
+from . import base
+
+
+def _is_failure_output(self, output):
+  return (
+    output.exit_code != 0 or
+    'FAILED!' in output.stdout
+  )
+
+
+class OutProc(base.OutProc):
+  """Optimized for positive tests."""
+OutProc._is_failure_output = _is_failure_output
+
+
+class PassOutProc(base.PassOutProc):
+  """Optimized for positive tests expected to PASS."""
+PassOutProc._is_failure_output = _is_failure_output
+
+
+class NegOutProc(base.Negative, OutProc):
+  pass
+
+class NegPassOutProc(base.Negative, PassOutProc):
+  pass
+
+
+MOZILLA_PASS_DEFAULT = PassOutProc()
+MOZILLA_PASS_NEGATIVE = NegPassOutProc()

tools/testrunner/outproc/test262.py

@@ -0,0 +1,54 @@
+# 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.
+
+import re
+
+from . import base
+
+
+class ExceptionOutProc(base.OutProc):
+  """Output processor for tests with expected exception."""
+  def __init__(self, expected_outcomes, expected_exception=None):
+    super(ExceptionOutProc, self).__init__(expected_outcomes)
+    self._expected_exception = expected_exception
+
+  def _is_failure_output(self, output):
+    if output.exit_code != 0:
+      return True
+    if self._expected_exception != self._parse_exception(output.stdout):
+      return True
+    return 'FAILED!' in output.stdout
+
+  def _parse_exception(self, string):
+    # somefile:somelinenumber: someerror[: sometext]
+    # somefile might include an optional drive letter on windows e.g. "e:".
+    match = re.search(
+        '^(?:\w:)?[^:]*:[0-9]+: ([^: ]+?)($|: )', string, re.MULTILINE)
+    if match:
+      return match.group(1).strip()
+    else:
+      return None
+
+
+def _is_failure_output(self, output):
+  return (
+    output.exit_code != 0 or
+    'FAILED!' in output.stdout
+  )
+
+
+class NoExceptionOutProc(base.OutProc):
+  """Output processor optimized for tests without expected exception."""
+NoExceptionOutProc._is_failure_output = _is_failure_output
+
+
+class PassNoExceptionOutProc(base.PassOutProc):
+  """
+  Output processor optimized for tests expected to PASS without expected
+  exception.
+  """
+PassNoExceptionOutProc._is_failure_output = _is_failure_output
+
+
+PASS_NO_EXCEPTION = PassNoExceptionOutProc()

tools/testrunner/outproc/webkit.py

@@ -0,0 +1,18 @@
+# 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.
+
+from . import base
+
+
+class OutProc(base.ExpectedOutProc):
+  def _is_failure_output(self, output):
+    if output.exit_code != 0:
+      return True
+    return super(OutProc, self)._is_failure_output(output)
+
+  def _ignore_expected_line(self, line):
+    return (
+        line.startswith('#') or
+        super(OutProc, self)._ignore_expected_line(line)
+    )

tools/testrunner/standard_runner.py

@@ -25,6 +25,7 @@ from testrunner.local import utils
 from testrunner.local import verbose
 from testrunner.local.variants import ALL_VARIANTS
 from testrunner.objects import context
+from testrunner.objects import predictable
 
 
 TIMEOUT_DEFAULT = 60
@@ -377,7 +378,6 @@ class StandardTestRunner(base_runner.BaseTestRunner):
                             options.no_sorting,
                             options.rerun_failures_count,
                             options.rerun_failures_max,
-                            self.build_config.predictable,
                             options.no_harness,
                             use_perf_data=not options.swarming,
                             sancov_dir=self.sancov_dir)
@@ -495,7 +495,12 @@ class StandardTestRunner(base_runner.BaseTestRunner):
         progress_indicator.Register(progress.FlakinessTestProgressIndicator(
             options.flakiness_results))
 
-      runner = execution.Runner(suites, progress_indicator, ctx)
+      if self.build_config.predictable:
+        outproc_factory = predictable.get_outproc
+      else:
+        outproc_factory = None
+      runner = execution.Runner(suites, progress_indicator, ctx,
+                                outproc_factory)
       exit_code = runner.Run(options.j)
       overall_duration = time.time() - start_time