Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(1351)

Unified Diff: mojo/tools/mopy/gtest.py

Issue 971083002: Create an apptesting framework for dart. (Closed) Base URL: https://github.com/domokit/mojo.git@master
Patch Set: Update upload_binaries.py to add the apptest.dartzip artifact. Created 5 years, 10 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View side-by-side diff with in-line comments
Download patch
Index: mojo/tools/mopy/gtest.py
diff --git a/mojo/tools/mopy/gtest.py b/mojo/tools/mopy/gtest.py
index 4663c729a14315134003684bc7295a4663a6fe18..d876a9a0a0447a07e0230d8f7997d997ce01afd3 100644
--- a/mojo/tools/mopy/gtest.py
+++ b/mojo/tools/mopy/gtest.py
@@ -5,12 +5,12 @@
import logging
import os
import re
-import subprocess
import sys
_logging = logging.getLogger()
-from mopy import android
+from mopy import test_util
+from mopy.background_app_group import BackgroundAppGroup
from mopy.config import Config
from mopy.paths import Paths
from mopy.print_process_error import print_process_error
@@ -23,6 +23,36 @@ def set_color():
_logging.debug("Setting GTEST_COLOR=yes")
os.environ["GTEST_COLOR"] = "yes"
+def run_fixtures(config, apptest_dict, apptest, test_args, shell_args,
+ launched_services):
+ """Run the gtest fixtures in isolation."""
+
+ mojo_paths = Paths(config)
+
+ # List the apptest fixtures so they can be run independently for isolation.
+ # TODO(msw): Run some apptests without fixture isolation?
+ fixtures = get_fixtures(config, shell_args, apptest)
+
+ if not fixtures:
+ return "Failed with no tests found."
+
+ if any(not mojo_paths.IsValidAppUrl(url) for url in launched_services):
+ return ("Failed with malformed launched-services: %r" % launched_services)
+
+ apptest_result = "Succeeded"
+ for fixture in fixtures:
+ apptest_args = test_args + ["--gtest_filter=%s" % fixture]
+ if launched_services:
+ success = RunApptestInLauncher(config, mojo_paths, apptest, apptest_args,
+ shell_args, launched_services)
+ else:
+ success = RunApptestInShell(config, apptest, apptest_args, shell_args)
+
+ if not success:
+ apptest_result = "Failed test(s) in %r" % apptest_dict
+
+ return apptest_result
+
def run_test(config, shell_args, apps_and_args=None, run_launcher=False):
"""Runs a command line and checks the output for signs of gtest failure.
@@ -35,7 +65,8 @@ def run_test(config, shell_args, apps_and_args=None, run_launcher=False):
run_launcher: |True| is mojo_launcher must be used instead of mojo_shell.
"""
apps_and_args = apps_and_args or {}
- output = _try_run_test(config, shell_args, apps_and_args, run_launcher)
+ output = test_util.try_run_test(config, shell_args, apps_and_args,
+ run_launcher)
# Fail on output with gtest's "[ FAILED ]" or a lack of "[ PASSED ]".
# The latter condition ensures failure on broken command lines or output.
# Check output instead of exit codes because mojo_shell always exits with 0.
@@ -43,7 +74,8 @@ def run_test(config, shell_args, apps_and_args=None, run_launcher=False):
(output.find("[ FAILED ]") != -1 or output.find("[ PASSED ]") == -1)):
print "Failed test:"
print_process_error(
- _build_command_line(config, shell_args, apps_and_args, run_launcher),
+ test_util.build_command_line(config, shell_args, apps_and_args,
+ run_launcher),
output)
return False
_logging.debug("Succeeded with output:\n%s" % output)
@@ -64,79 +96,16 @@ def get_fixtures(config, shell_args, apptest):
"""
try:
apps_and_args = {apptest: ["--gtest_list_tests"]}
- list_output = _run_test(config, shell_args, apps_and_args)
+ list_output = test_util.run_test(config, shell_args, apps_and_args)
_logging.debug("Tests listed:\n%s" % list_output)
return _gtest_list_tests(list_output)
except Exception as e:
print "Failed to get test fixtures:"
print_process_error(
- _build_command_line(config, shell_args, apps_and_args), e)
+ test_util.build_command_line(config, shell_args, apps_and_args), e)
return []
-def build_shell_arguments(shell_args, apps_and_args=None):
- """Build the list of arguments for the shell. |shell_args| are the base
- arguments, |apps_and_args| is a dictionary that associates each application to
- its specific arguments|. Each app included will be run by the shell.
- """
- result = shell_args[:]
- if apps_and_args:
- for (application, args) in apps_and_args.items():
- result += ["--args-for=%s %s" % (application, " ".join(args))]
- result += apps_and_args.keys()
- return result
-
-
-def _get_shell_executable(config, run_launcher):
- paths = Paths(config=config)
- if config.target_os == Config.OS_ANDROID:
- return os.path.join(paths.src_root, "mojo", "tools",
- "android_mojo_shell.py")
- elif run_launcher:
- return paths.mojo_launcher_path
- else:
- return paths.mojo_shell_path
-
-
-def _build_command_line(config, shell_args, apps_and_args, run_launcher=False):
- executable = _get_shell_executable(config, run_launcher)
- return "%s %s" % (executable, " ".join(["%r" % x for x in
- build_shell_arguments(
- shell_args, apps_and_args)]))
-
-
-def _run_test_android(shell_args, apps_and_args):
- """Run the given test on the single/default android device."""
- (r, w) = os.pipe()
- with os.fdopen(r, "r") as rf:
- with os.fdopen(w, "w") as wf:
- arguments = build_shell_arguments(shell_args, apps_and_args)
- android.StartShell(arguments, wf, wf.close)
- return rf.read()
-
-
-def _run_test(config, shell_args, apps_and_args, run_launcher=False):
- """Run the given test, using mojo_launcher if |run_launcher| is True."""
- if (config.target_os == Config.OS_ANDROID):
- return _run_test_android(shell_args, apps_and_args)
- else:
- executable = _get_shell_executable(config, run_launcher)
- command = ([executable] + build_shell_arguments(shell_args, apps_and_args))
- return subprocess.check_output(command, stderr=subprocess.STDOUT)
-
-
-def _try_run_test(config, shell_args, apps_and_args, run_launcher):
- """Returns the output of a command line or an empty string on error."""
- command_line = _build_command_line(config, shell_args, apps_and_args,
- run_launcher=run_launcher)
- _logging.debug("Running command line: %s" % command_line)
- try:
- return _run_test(config, shell_args, apps_and_args, run_launcher)
- except Exception as e:
- print_process_error(command_line, e)
- return None
-
-
def _gtest_list_tests(gtest_list_tests_output):
"""Returns a list of strings formatted as TestSuite.TestFixture from the
output of running --gtest_list_tests on a GTEST application."""
@@ -161,3 +130,20 @@ def _gtest_list_tests(gtest_list_tests_output):
test_list.append(suite + line.strip())
return test_list
+
+
+def RunApptestInShell(config, application, application_args, shell_args):
+ return run_test(config, shell_args, {application: application_args})
+
+
+def RunApptestInLauncher(config, mojo_paths, application, application_args,
+ shell_args, launched_services):
+ with BackgroundAppGroup(
+ mojo_paths, launched_services,
+ test_util.build_shell_arguments(shell_args)) as apps:
+ launcher_args = [
+ '--shell-path=' + apps.socket_path,
+ '--app-url=' + application,
+ '--app-path=' + mojo_paths.FileFromUrl(application),
+ '--app-args=' + " ".join(application_args)]
+ return run_test(config, launcher_args, run_launcher=True)

Powered by Google App Engine
This is Rietveld 408576698