blob: 90eaa43e77587aa8a1a6c48ce28191a2d4f2e83b [file] [log] [blame]
#!/usr/bin/env python
# Copyright (c) 2012 The Chromium Authors. All rights reserved.
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.
"""Start and stop Web Page Replay.
Of the public module names, the following one is key:
ReplayServer: a class to start/stop Web Page Replay.
"""
import logging
import os
import re
import signal
import subprocess
import sys
import time
import urllib
_CHROME_BASE_DIR = os.path.abspath(os.path.join(
os.path.dirname(__file__), os.pardir, os.pardir, os.pardir, os.pardir))
REPLAY_DIR = os.path.join(
_CHROME_BASE_DIR, 'src', 'third_party', 'webpagereplay')
LOG_PATH = os.path.join(
_CHROME_BASE_DIR, 'src', 'webpagereplay_logs', 'logs.txt')
# Chrome options to make it work with Web Page Replay.
def GetChromeFlags(replay_host, http_port, https_port):
assert replay_host and http_port and https_port, 'All arguments required'
return [
'--host-resolver-rules=MAP * %s,EXCLUDE localhost' % replay_host,
'--testing-fixed-http-port=%s' % http_port,
'--testing-fixed-https-port=%s' % https_port,
'--ignore-certificate-errors',
]
# Signal masks on Linux are inherited from parent processes. If anything
# invoking us accidentally masks SIGINT (e.g. by putting a process in the
# background from a shell script), sending a SIGINT to the child will fail
# to terminate it. Running this signal handler before execing should fix that
# problem.
def ResetInterruptHandler():
signal.signal(signal.SIGINT, signal.SIG_DFL)
class ReplayError(Exception):
"""Catch-all exception for the module."""
pass
class ReplayNotFoundError(ReplayError):
def __init__(self, label, path):
self.args = (label, path)
def __str__(self):
label, path = self.args
return 'Path does not exist for %s: %s' % (label, path)
class ReplayNotStartedError(ReplayError):
pass
class ReplayServer(object):
"""Start and Stop Web Page Replay.
Web Page Replay is a proxy that can record and "replay" web pages with
simulated network characteristics -- without having to edit the pages
by hand. With WPR, tests can use "real" web content, and catch
performance issues that may result from introducing network delays and
bandwidth throttling.
Example:
with ReplayServer(archive_path):
self.NavigateToURL(start_url)
self.WaitUntil(...)
Environment Variables (for development):
WPR_ARCHIVE_PATH: path to alternate archive file (e.g. '/tmp/foo.wpr').
WPR_RECORD: if set, puts Web Page Replay in record mode instead of replay.
WPR_REPLAY_DIR: path to alternate Web Page Replay source.
"""
def __init__(self, archive_path, replay_host, http_port, https_port,
replay_options=None, replay_dir=None,
log_path=None):
"""Initialize ReplayServer.
Args:
archive_path: a path to a specific WPR archive (required).
replay_host: the hostname to serve traffic.
http_port: an integer port on which to serve HTTP traffic. May be zero
to let the OS choose an available port.
https_port: an integer port on which to serve HTTPS traffic. May be zero
to let the OS choose an available port.
replay_options: an iterable of options strings to forward to replay.py.
replay_dir: directory that has replay.py and related modules.
log_path: a path to a log file.
"""
self.archive_path = os.environ.get('WPR_ARCHIVE_PATH', archive_path)
self.replay_options = list(replay_options or ())
self.replay_dir = os.environ.get('WPR_REPLAY_DIR', replay_dir or REPLAY_DIR)
self.log_path = log_path or LOG_PATH
self.http_port = http_port
self.https_port = https_port
self._replay_host = replay_host
if 'WPR_RECORD' in os.environ and '--record' not in self.replay_options:
self.replay_options.append('--record')
self.is_record_mode = '--record' in self.replay_options
self._AddDefaultReplayOptions()
self.replay_py = os.path.join(self.replay_dir, 'replay.py')
if self.is_record_mode:
self._CheckPath('archive directory', os.path.dirname(self.archive_path))
elif not os.path.exists(self.archive_path):
self._CheckPath('archive file', self.archive_path)
self._CheckPath('replay script', self.replay_py)
self.log_fh = None
self.replay_process = None
def _AddDefaultReplayOptions(self):
"""Set WPR command-line options. Can be overridden if needed."""
self.replay_options = [
'--host', str(self._replay_host),
'--port', str(self.http_port),
'--ssl_port', str(self.https_port),
'--use_closest_match',
'--no-dns_forwarding',
'--log_level', 'warning'
] + self.replay_options
def _CheckPath(self, label, path):
if not os.path.exists(path):
raise ReplayNotFoundError(label, path)
def _OpenLogFile(self):
log_dir = os.path.dirname(self.log_path)
if not os.path.exists(log_dir):
os.makedirs(log_dir)
return open(self.log_path, 'w')
def WaitForStart(self, timeout):
"""Checks to see if the server is up and running."""
port_re = re.compile(
'.*(?P<protocol>HTTPS?) server started on (?P<host>.*):(?P<port>\d+)')
start_time = time.time()
elapsed_time = 0
while elapsed_time < timeout:
if self.replay_process.poll() is not None:
break # The process has exited.
# Read the ports from the WPR log.
if not self.http_port or not self.https_port:
for line in open(self.log_path).readlines():
m = port_re.match(line.strip())
if m:
if not self.http_port and m.group('protocol') == 'HTTP':
self.http_port = int(m.group('port'))
elif not self.https_port and m.group('protocol') == 'HTTPS':
self.https_port = int(m.group('port'))
# Try to connect to the WPR ports.
if self.http_port and self.https_port:
try:
up_url = '%s://%s:%s/web-page-replay-generate-200'
http_up_url = up_url % ('http', self._replay_host, self.http_port)
https_up_url = up_url % ('https', self._replay_host, self.https_port)
if (200 == urllib.urlopen(http_up_url, None, {}).getcode() and
200 == urllib.urlopen(https_up_url, None, {}).getcode()):
return True
except IOError:
pass
poll_interval = min(max(elapsed_time / 10., .1), 5)
time.sleep(poll_interval)
elapsed_time = time.time() - start_time
return False
def StartServer(self):
"""Start Web Page Replay and verify that it started.
Raises:
ReplayNotStartedError: if Replay start-up fails.
"""
cmd_line = [sys.executable, self.replay_py]
cmd_line.extend(self.replay_options)
cmd_line.append(self.archive_path)
self.log_fh = self._OpenLogFile()
logging.debug('Starting Web-Page-Replay: %s', cmd_line)
kwargs = {'stdout': self.log_fh, 'stderr': subprocess.STDOUT}
if sys.platform.startswith('linux') or sys.platform == 'darwin':
kwargs['preexec_fn'] = ResetInterruptHandler
self.replay_process = subprocess.Popen(cmd_line, **kwargs)
if not self.WaitForStart(30):
log = open(self.log_path).read()
raise ReplayNotStartedError(
'Web Page Replay failed to start. Log output:\n%s' % log)
def StopServer(self):
"""Stop Web Page Replay."""
if self.replay_process:
logging.debug('Trying to stop Web-Page-Replay gracefully')
try:
url = 'http://localhost:%s/web-page-replay-command-exit'
urllib.urlopen(url % self.http_port, None, {})
except IOError:
# IOError is possible because the server might exit without response.
pass
start_time = time.time()
while time.time() - start_time < 10: # Timeout after 10 seconds.
if self.replay_process.poll() is not None:
break
time.sleep(1)
else:
try:
# Use a SIGINT so that it can do graceful cleanup.
self.replay_process.send_signal(signal.SIGINT)
except: # pylint: disable=W0702
# On Windows, we are left with no other option than terminate().
if 'no-dns_forwarding' not in self.replay_options:
logging.warning('DNS configuration might not be restored!')
try:
self.replay_process.terminate()
except: # pylint: disable=W0702
pass
self.replay_process.wait()
if self.log_fh:
self.log_fh.close()
def __enter__(self):
"""Add support for with-statement."""
self.StartServer()
return self
def __exit__(self, unused_exc_type, unused_exc_val, unused_exc_tb):
"""Add support for with-statement."""
self.StopServer()