def wsgi_request(self, environ, start_response):
if self.wsgi_debug:
print 'wsgi_request start'
print environ
print '...'
outstream = StringIO()
response = Response(stdout=outstream, stderr=sys.stderr)
extra = {
'SESSION': self.app.REQUEST.SESSION,
'AUTHENTICATED_USER': self.app.REQUEST.AUTHENTICATED_USER,
}
publish_module('Zope2', environ=environ, response=response, extra=extra, stdin=environ['wsgi.input'])
output = outstream.getvalue()
headers, body = output.split(newline*2, 1)
headers = [header.split(': ', 1) for header in headers.split(newline)]
status = headers.pop(0)[1]
if self.wsgi_debug:
print 'wsgi_request done, status="%s"' % status
print ' ' + '\n'.join([': '.join(header) for header in headers])
headers = [ (header[0], ', '.join(header[1:])) for header in headers ]
if 'content-type' not in (header[0].lower() for header in headers):
headers.append( ('Content-Type', 'text/html; charset=utf-8') )
start_response(status, headers)
return [body]
def publish(self, path, basic=None, env=None, extra=None,
request_method='GET', stdin=None, handle_errors=True):
'''Publishes the object at 'path' returning a response object.'''
from StringIO import StringIO
from ZPublisher.Response import Response
from ZPublisher.Test import publish_module
from AccessControl.SecurityManagement import getSecurityManager
from AccessControl.SecurityManagement import setSecurityManager
# Save current security manager
sm = getSecurityManager()
# Commit the sandbox for good measure
transaction.commit()
if env is None:
env = {}
if extra is None:
extra = {}
request = self.app.REQUEST
env['SERVER_NAME'] = request['SERVER_NAME']
env['SERVER_PORT'] = request['SERVER_PORT']
env['REQUEST_METHOD'] = request_method
p = path.split('?')
if len(p) == 1:
env['PATH_INFO'] = p[0]
elif len(p) == 2:
[env['PATH_INFO'], env['QUERY_STRING']] = p
else:
raise TypeError, ''
if basic:
env['HTTP_AUTHORIZATION'] = "Basic %s" % base64.encodestring(basic)
if stdin is None:
stdin = StringIO()
outstream = StringIO()
response = Response(stdout=outstream, stderr=sys.stderr)
publish_module('Zope2',
response=response,
stdin=stdin,
environ=env,
extra=extra,
debug=not handle_errors,
)
# Restore security manager
setSecurityManager(sm)
return ResponseWrapper(response, outstream, path)
def publish(self, path, basic=None, env=None, extra=None,
request_method='GET', stdin=None, handle_errors=True):
'''Publishes the object at 'path' returning a response object.'''
from StringIO import StringIO
from ZPublisher.Response import Response
from ZPublisher.Test import publish_module
# Commit the sandbox for good measure
transaction.commit()
if env is None:
env = {}
if extra is None:
extra = {}
request = self.app.REQUEST
env['SERVER_NAME'] = request['SERVER_NAME']
env['SERVER_PORT'] = request['SERVER_PORT']
env['REQUEST_METHOD'] = request_method
p = path.split('?')
if len(p) == 1:
env['PATH_INFO'] = p[0]
elif len(p) == 2:
[env['PATH_INFO'], env['QUERY_STRING']] = p
else:
raise TypeError, ''
if basic:
env['HTTP_AUTHORIZATION'] = "Basic %s" % base64.encodestring(basic)
if stdin is None:
stdin = StringIO()
outstream = StringIO()
response = Response(stdout=outstream, stderr=sys.stderr)
publish_module('Zope2',
response=response,
stdin=stdin,
environ=env,
extra=extra,
debug=not handle_errors,
)
return ResponseWrapper(response, outstream, path)
def __call__(self, requestString, handle_errors=True):
from ZPublisher.Iterators import IStreamIterator
from ZPublisher.Response import Response
from ZPublisher.Test import publish_module
class TestResponse(Response):
def setBody(self, body, title='', is_error=0, **kw):
if IStreamIterator.providedBy(body):
body = ''.join(body)
Response.setBody(self, body, title, is_error, **kw)
# Discard leading white space to make call layout simpler
requestString = requestString.lstrip()
# Split off and parse the command line
l = requestString.find('\n')
commandLine = requestString[:l].rstrip()
requestString = requestString[l+1:]
method, url, protocol = commandLine.split()
instream = StringIO(requestString)
env = {"HTTP_HOST": 'localhost',
"HTTP_REFERER": 'localhost',
"REQUEST_METHOD": method,
"SERVER_PROTOCOL": protocol,
}
p = url.split('?', 1)
if len(p) == 1:
env['PATH_INFO'] = p[0]
elif len(p) == 2:
[env['PATH_INFO'], env['QUERY_STRING']] = p
else:
raise TypeError, ''
# If you followed closely, you notice that one part of the url
# gets unquoted (PATH_INFO) while the other (QUERY_STRING)
# doesn't That complies with what the ZSERVER does.
env['PATH_INFO'] = urllib.unquote(env['PATH_INFO'])
headers = [splitHeader(header) for header in rfc822.Message(instream).headers]
# Store request body without headers
instream = StringIO(instream.read())
for name, value in headers:
name = ('_'.join(name.upper().split('-')))
if name not in ('CONTENT_TYPE', 'CONTENT_LENGTH'):
name = 'HTTP_' + name
env[name] = value.rstrip()
if env.has_key('HTTP_AUTHORIZATION'):
env['HTTP_AUTHORIZATION'] = authHeader(env['HTTP_AUTHORIZATION'])
outstream = StringIO()
response = TestResponse(stdout=outstream, stderr=sys.stderr)
publish_module('Zope2',
response=response,
stdin=instream,
environ=env,
debug=not handle_errors,
)
self.app._p_jar.sync()
return response
def http(request_string, handle_errors=True):
"""Execute an HTTP request string via the publisher
This is used for HTTP doc tests.
"""
import urllib
import rfc822
from cStringIO import StringIO
from ZPublisher.Response import Response
from ZPublisher.Test import publish_module
from AccessControl.SecurityManagement import getSecurityManager
from AccessControl.SecurityManagement import setSecurityManager
# Save current Security Manager
old_sm = getSecurityManager()
# Commit work done by previous python code.
transaction.commit()
# Discard leading white space to make call layout simpler
request_string = request_string.lstrip()
# Split off and parse the command line
l = request_string.find('\n')
command_line = request_string[:l].rstrip()
request_string = request_string[l+1:]
method, path, protocol = command_line.split()
path = urllib.unquote(path)
instream = StringIO(request_string)
env = {"HTTP_HOST": 'localhost',
"HTTP_REFERER": 'localhost',
"REQUEST_METHOD": method,
"SERVER_PROTOCOL": protocol,
}
p = path.split('?')
if len(p) == 1:
env['PATH_INFO'] = p[0]
elif len(p) == 2:
[env['PATH_INFO'], env['QUERY_STRING']] = p
else:
raise TypeError, ''
header_output = HTTPHeaderOutput(
protocol, ('x-content-type-warning', 'x-powered-by',
'bobo-exception-type', 'bobo-exception-file',
'bobo-exception-value', 'bobo-exception-line'))
headers = [split_header(header)
for header in rfc822.Message(instream).headers]
# Store request body without headers
instream = StringIO(instream.read())
for name, value in headers:
name = ('_'.join(name.upper().split('-')))
if name not in ('CONTENT_TYPE', 'CONTENT_LENGTH'):
name = 'HTTP_' + name
env[name] = value.rstrip()
if env.has_key('HTTP_AUTHORIZATION'):
env['HTTP_AUTHORIZATION'] = auth_header(env['HTTP_AUTHORIZATION'])
outstream = StringIO()
response = Response(stdout=outstream, stderr=sys.stderr)
publish_module('Zope2',
response=response,
stdin=instream,
environ=env,
debug=not handle_errors,
)
header_output.setResponseStatus(response.getStatus(), response.errmsg)
header_output.setResponseHeaders(response.headers)
header_output.appendResponseHeaders(response._cookie_list())
header_output.appendResponseHeaders(response.accumulated_headers.splitlines())
# Restore previous security manager, which may have been changed
# by calling the publish method above
setSecurityManager(old_sm)
# Sync connection
sync()
return DocResponseWrapper(response, outstream, path, header_output)
def http(request_string, handle_errors=True):
"""Execute an HTTP request string via the publisher
This is used for HTTP doc tests.
"""
import urllib
import rfc822
from cStringIO import StringIO
from ZPublisher.Response import Response
from ZPublisher.Test import publish_module
# Commit work done by previous python code.
transaction.commit()
# Discard leading white space to make call layout simpler
request_string = request_string.lstrip()
# Split off and parse the command line
l = request_string.find('\n')
command_line = request_string[:l].rstrip()
request_string = request_string[l+1:]
method, path, protocol = command_line.split()
path = urllib.unquote(path)
instream = StringIO(request_string)
env = {"HTTP_HOST": 'localhost',
"HTTP_REFERER": 'localhost',
"REQUEST_METHOD": method,
"SERVER_PROTOCOL": protocol,
}
p = path.split('?')
if len(p) == 1:
env['PATH_INFO'] = p[0]
elif len(p) == 2:
[env['PATH_INFO'], env['QUERY_STRING']] = p
else:
raise TypeError, ''
header_output = HTTPHeaderOutput(
protocol, ('x-content-type-warning', 'x-powered-by',
'bobo-exception-type', 'bobo-exception-file',
'bobo-exception-value', 'bobo-exception-line'))
headers = [split_header(header)
for header in rfc822.Message(instream).headers]
# Store request body without headers
instream = StringIO(instream.read())
for name, value in headers:
name = ('_'.join(name.upper().split('-')))
if name not in ('CONTENT_TYPE', 'CONTENT_LENGTH'):
name = 'HTTP_' + name
env[name] = value.rstrip()
if env.has_key('HTTP_AUTHORIZATION'):
env['HTTP_AUTHORIZATION'] = auth_header(env['HTTP_AUTHORIZATION'])
outstream = StringIO()
response = Response(stdout=outstream, stderr=sys.stderr)
publish_module('Zope2',
response=response,
stdin=instream,
environ=env,
debug=not handle_errors,
)
header_output.setResponseStatus(response.getStatus(), response.errmsg)
header_output.setResponseHeaders(response.headers)
header_output.appendResponseHeaders(response._cookie_list())
header_output.appendResponseHeaders(response.accumulated_headers.splitlines())
sync()
return DocResponseWrapper(response, outstream, path, header_output)
def run(app, args, rate=5):
# Adjust root logging handler levels
level = getConfiguration().eventlog.getLowestHandlerLevel()
root = logging.getLogger()
for handler in root.handlers:
handler.setLevel(level)
logger = logging.getLogger("linkcheck.processor")
logger.setLevel(level)
logger.info("looking for sites...")
session = requests.Session(timeout=5)
counter = 0
sites = {}
# Enter runloop
while True:
errors = set()
for name, item in app.objectItems():
if name in sites:
continue
if IPloneSiteRoot.providedBy(item):
try:
tool = getToolByName(item, 'portal_linkcheck')
except AttributeError:
continue
logger.info("found site '%s'." % name)
registry = getUtility(IRegistry, context=item)
try:
settings = registry.forInterface(ISettings)
except KeyError:
logger.warn("settings not available; please reinstall.")
continue
responses = []
def worker():
while True:
url = q.get()
r = None
try:
r = session.get(url)
except requests.Timeout:
status_code = 504
except requests.RequestException as exc:
logger.warn(exc)
status_code = 503
except UnicodeError as exc:
logger.warn("Unable to decode string: %r (%s)." % (
url, exc))
status_code = 502
if r is None:
r = requests.Response()
r.status_code = status_code
r.url = url
responses.append(r)
q.task_done()
q = Queue()
for i in range(settings.concurrency):
t = threading.Thread(target=worker)
t.daemon = True
t.start()
logger.info(
"%d worker threads started." % settings.concurrency
)
sites[name] = (tool, settings, q, responses)
if not sites and not counter:
logger.info(
"no sites found; polling every %d second(s) ..." % rate
)
for tool, settings, queue, responses in sites.values():
# Synchronize database
tool._p_jar.sync()
if not tool.is_available():
logger.warn("Tool not available; please run update step.")
logger.info("Sleeping for 10 seconds...")
time.sleep(10)
break
if not counter % 3600:
now = datetime.datetime.now()
# This timestamp is the threshold for items that need an
# update.
needs_update = int(time.mktime(
(now - datetime.timedelta(hours=settings.interval)).\
timetuple()
))
# This timestamp is the threshold for items that are no
# longer active.
expired = int(time.mktime(
(now - datetime.timedelta(days=settings.expiration)).\
timetuple()
))
discard = set()
for url, entry in tool.checked.items():
if url in tool.queue:
continue
# Discard items that are expired
if entry[0] and entry[0] < expired:
discard.add(url)
# Enqueue items with an out of date timestamp.
elif entry[0] and entry[0] < needs_update:
tool.queue.put(url)
for url in discard:
del tool.checked[url]
# Fetch set of URLs to check (up to transaction size).
queued = tool.queue[:settings.transaction_size]
if not queued:
continue
urls = filter(None, map(tool.links.get, queued))
# This keeps track of status updates, which we'll apply at
# the end.
updates = []
# Distinguish between internal and external requests.
internal, external = partition(
lambda url: url.startswith('/'),
urls
)
# Must be HTTP or HTTPS
external, invalid = partition(
lambda url: url.startswith('http://') or \
url.startswith('https://'),
external
)
for url in external:
queue.put(url)
# Wait for responses
queue.join()
while responses:
response = responses.pop()
status = response.status_code
# This may be a redirect.
if response.history:
url = response.history[0].url
if response.history[0].status_code == 301:
status = 301
else:
url = response.url
updates.append((url, status))
for url in internal:
# For now, we simply ignore internal links if we're
# not publishing.
if not settings.use_publisher:
continue
stdout = StringIO()
stderr = StringIO()
env = {
'GATEWAY_INTERFACE': 'CGI/1.1 ',
'HTTP_ACCEPT': '*/*',
'HTTP_HOST': '127.0.0.1',
'HTTP_USER_AGENT': 'Bobo',
'REQUEST_METHOD': 'GET',
'SCRIPT_NAME': '',
'SERVER_HOSTNAME': 'bobo.server.host',
'SERVER_NAME': 'bobo.server',
'SERVER_PORT': '80',
'SERVER_PROTOCOL': 'HTTP/1.0 ',
}
env['PATH_INFO'] = "/" + tool.aq_parent.absolute_url() + url
try:
status = publish_module(
'Zope2', environ=env, stdout=stdout, stderr=stderr
)
except ConflictError:
status = 503
else:
# This is assumed to be a good response.
if status == 302:
status = 200
updates.append((url, status))
# Pull URLs out of queue, actually removing them.
unchanged = []
urls = set(urls)
while urls:
try:
i = tool.queue.pull()
except IndexError:
transaction.abort()
continue
try:
url = tool.links[i]
urls.remove(url)
except KeyError:
unchanged.append(i)
# This shouldn't happen to frequently.
for i in unchanged:
tool.queue.put(i)
url = tool.links[i]
logger.warn("putting back unprocessed url: %s." % url)
for url in invalid:
tool.update(url, 0)
errors.add(url)
# Apply status updates
for url, status in updates:
tool.update(url, status)
transaction.get().note('updated link validity')
try:
transaction.commit()
except ConflictError:
transaction.abort()
for url in errors:
logger.warn("error checking: %s." % url)
time.sleep(rate)
app._p_jar.sync()
counter += 1
