Git-Mirrors/brozzler
1
0
Fork 0
mirror of https://github.com/internetarchive/brozzler.git synced 2025-08-03 12:06:28 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

Use Network.setBlockedUrls instead of Debugger to block URLs

Browse source
This commit is contained in:
Vangelis Banos 2018-02-01 08:45:56 +00:00 • committed by Noah Levitt
parent fc000ff515
commit f54d62ea40
1 changed files with 9 additions and 27 deletions
Download patch file Download diff file Expand all files Collapse all files

36
brozzler/browser.py
View file

@ -200,21 +200,6 @@ class WebsockReceiverThread(threading.Thread):
'uncaught exception in _handle_message message=%s', 'uncaught exception in _handle_message message=%s',
message, exc_info=True) message, exc_info=True)
def _debugger_paused(self, message):
# we hit the breakpoint set in start(), get rid of google analytics
self.logger.debug('debugger paused! message=%s', message)
scriptId = message['params']['callFrames'][0]['location']['scriptId']
# replace script
self.websock.send(
json.dumps(dict(
id=0, method='Debugger.setScriptSource',
params={'scriptId': scriptId,
'scriptSource': 'console.log("google analytics is no more!");'})))
# resume execution
self.websock.send(json.dumps(dict(id=0, method='Debugger.resume')))
def _network_response_received(self, message): def _network_response_received(self, message):
if (message['params']['response']['status'] == 420 if (message['params']['response']['status'] == 420
and 'Warcprox-Meta' in CaseInsensitiveDict( and 'Warcprox-Meta' in CaseInsensitiveDict(
@ -255,8 +240,6 @@ class WebsockReceiverThread(threading.Thread):
elif message['method'] == 'Network.requestWillBeSent': elif message['method'] == 'Network.requestWillBeSent':
if self.on_request: if self.on_request:
self.on_request(message) self.on_request(message)
elif message['method'] == 'Debugger.paused':
self._debugger_paused(message)
elif message['method'] == 'Page.interstitialShown': elif message['method'] == 'Page.interstitialShown':
# for AITFIVE-1529: handle http auth # for AITFIVE-1529: handle http auth
# for now, we should consider killing the browser when we receive Page.interstitialShown and # for now, we should consider killing the browser when we receive Page.interstitialShown and
@ -332,9 +315,10 @@ class Browser:
msg_id = next(self._command_id) msg_id = next(self._command_id)
kwargs['id'] = msg_id kwargs['id'] = msg_id
msg = json.dumps(kwargs) msg = json.dumps(kwargs)
logging.log( #logging.log(
brozzler.TRACE if suppress_logging else logging.DEBUG, # brozzler.TRACE if suppress_logging else logging.DEBUG,
'sending message to %s: %s', self.websock, msg) # 'sending message to %s: %s', self.websock, msg)
# print(msg)
self.websock.send(msg) self.websock.send(msg)
return msg_id return msg_id
@ -358,16 +342,14 @@ class Browser:
self.send_to_chrome(method='Network.enable') self.send_to_chrome(method='Network.enable')
self.send_to_chrome(method='Page.enable') self.send_to_chrome(method='Page.enable')
self.send_to_chrome(method='Console.enable') self.send_to_chrome(method='Console.enable')
self.send_to_chrome(method='Debugger.enable')
self.send_to_chrome(method='Runtime.enable') self.send_to_chrome(method='Runtime.enable')
# disable google analytics, see _handle_message() where breakpoint # disable google analytics
# is caught Debugger.paused
self.send_to_chrome( self.send_to_chrome(
method='Debugger.setBreakpointByUrl', method='Network.setBlockedURLs',
params={ params={'urls': ['http://www.google-analytics.com/analytics.js',
'lineNumber': 1, 'https://www.google-analytics.com/analytics.js']
'urlRegex': 'https?://www.google-analytics.com/analytics.js'}) )
def stop(self): def stop(self):
''' '''