def _process_redirect(self):
'''Update the Redirect Tracker.'''
_logger.debug('Handling redirect.')
if self._redirect_tracker.exceeded():
raise ProtocolError('Too many redirects.')
try:
url = self._redirect_tracker.next_location()
if not url:
raise ProtocolError('Redirect location missing.')
if self._redirect_tracker.is_repeat():
_logger.debug('Got redirect is repeat.')
request = self._original_request.copy()
request.url = url
else:
request = self._web_client.request_factory(url)
request.prepare_for_send()
except ValueError as error:
raise ProtocolError('Invalid redirect location.') from error
self._next_request = request
_logger.debug('Updated next redirect request to {0}.'.format(request))
def read_chunk_header(self):
'''Read a single chunk's header.
Returns:
tuple: 2-item tuple with the size of the content in the chunk and
the raw header byte string.
Coroutine.
'''
# _logger.debug('Reading chunk.')
try:
chunk_size_hex = yield from self._connection.readline()
except ValueError as error:
raise ProtocolError(
'Invalid chunk size: {0}'.format(error)) from error
if not chunk_size_hex.endswith(b'\n'):
raise NetworkError('Connection closed.')
try:
chunk_size = int(chunk_size_hex.split(b';', 1)[0].strip(), 16)
except ValueError as error:
raise ProtocolError(
'Invalid chunk size: {0}'.format(error)) from error
if chunk_size < 0:
raise ProtocolError('Chunk size cannot be negative.')
self._chunk_size = self._bytes_left = chunk_size
return chunk_size, chunk_size_hex
def _update_redirect_request(self):
'''Update the Redirect Tracker.'''
_logger.debug('Handling redirect.')
if self._redirect_tracker.exceeded():
raise ProtocolError('Too many redirects.')
url = self._redirect_tracker.next_location()
if not url:
raise ProtocolError('Redirect location missing.')
try:
request = self._rich_client.request_factory(url)
except ValueError as error:
raise ProtocolError('Invalid redirect location.') from error
if self._redirect_tracker.is_repeat():
_logger.debug('Got redirect is repeat.')
request.method = self._original_request.method
request.body = self._original_request.body
for name, value in self._original_request.fields.items():
if name not in request.fields:
request.fields.add(name, value)
self._next_request = request
_logger.debug('Updated next redirect request to {0}.'.format(request))
def _decompress_data(self, data):
'''Decompress the given data and return the uncompressed data.'''
if self._decompressor:
try:
return self._decompressor.decompress(data)
except zlib.error as error:
raise ProtocolError(
'zlib error: {0}.'.format(error)) from error
else:
return data
def _flush_decompressor(self):
'''Return any data left in the decompressor.'''
if self._decompressor:
try:
return self._decompressor.flush()
except zlib.error as error:
raise ProtocolError(
'zlib flush error: {0}.'.format(error)) from error
else:
return b''
def read_chunk(self):
'''Read a single chunk of the chunked transfer encoding.
Returns:
int: The size of the content in the chunk.
'''
_logger.debug('Reading chunk.')
chunk_size_hex = yield self._io_stream.read_until(b'\n')
self.data_event.fire(chunk_size_hex)
try:
chunk_size = int(chunk_size_hex.split(b';', 1)[0].strip(), 16)
except ValueError as error:
raise ProtocolError(error.args[0]) from error
_logger.debug('Getting chunk size={0}.'.format(chunk_size))
if not chunk_size:
raise tornado.gen.Return(chunk_size)
data_queue = self._io_stream.read_bytes_queue(chunk_size)
while True:
data = yield data_queue.get()
if data is None:
break
self.data_event.fire(data)
self.content_event.fire(data)
newline_data = yield self._io_stream.read_until(b'\n')
self.data_event.fire(newline_data)
if len(newline_data) > 2:
# Should be either CRLF or LF
# This could our problem or the server's problem
raise ProtocolError('Error reading newline after chunk.')
raise tornado.gen.Return(chunk_size)
def parse(self, data):
assert self.name is None
assert not self.argument
match = re.match(br'(\w+) ?([^\r\n]*)', data)
if not match:
raise ProtocolError('Failed to parse command.')
self.name = match.group(1).decode('utf-8', errors='surrogateescape')
self.argument = match.group(2).decode('utf-8', errors='surrogateescape')
def read_listing_content(self, file, duration_timeout=None):
'''Read file listings.
Returns:
.ftp.request.ListingResponse: A Response populated the
file listings
Be sure to call :meth:`fetch_file_listing` first.
Coroutine.
'''
yield From(
self.read_content(file=file,
rewind=False,
duration_timeout=duration_timeout))
try:
if self._response.body.tell() == 0:
listings = ()
elif self._listing_type == 'mlsd':
self._response.body.seek(0)
machine_listings = wpull.ftp.util.parse_machine_listing(
self._response.body.read().decode(
'utf-8', errors='surrogateescape'),
convert=True,
strict=False)
listings = list(
wpull.ftp.util.machine_listings_to_file_entries(
machine_listings))
else:
self._response.body.seek(0)
file = io.TextIOWrapper(self._response.body,
encoding='utf-8',
errors='surrogateescape')
listing_parser = ListingParser(file=file)
heuristics_result = listing_parser.run_heuristics()
_logger.debug('Listing detected as %s', heuristics_result)
listings = listing_parser.parse()
# We don't want the file to be closed when exiting this function
file.detach()
except (ListingError, ValueError) as error:
raise ProtocolError(*error.args) from error
self._response.files = listings
self._response.body.seek(0)
raise Return(self._response)
def read_response(self, response=None):
'''Read the response's HTTP status line and header fields.
Coroutine.
'''
_logger.debug('Reading header.')
if response is None:
response = Response()
header_lines = []
bytes_read = 0
while True:
try:
data = yield From(self._connection.readline())
except ValueError as error:
raise ProtocolError(
'Invalid header: {0}'.format(error)) from error
self._data_observer.notify('response', data)
if not data.endswith(b'\n'):
raise NetworkError('Connection closed.')
elif data in (b'\r\n', b'\n'):
break
header_lines.append(data)
assert data.endswith(b'\n')
bytes_read += len(data)
if bytes_read > 32768:
raise ProtocolError('Header too big.')
if not header_lines:
raise ProtocolError('No header received.')
response.parse(b''.join(header_lines))
raise Return(response)
def read_chunk_body(self):
'''Read a fragment of a single chunk.
Call :meth:`read_chunk_header` first.
Returns:
tuple: 2-item tuple with the content data and raw data.
First item is empty bytes string when chunk is fully read.
Coroutine.
'''
# chunk_size = self._chunk_size
bytes_left = self._bytes_left
# _logger.debug(__('Getting chunk size={0}, remain={1}.',
# chunk_size, bytes_left))
if bytes_left > 0:
size = min(bytes_left, self._read_size)
data = yield from self._connection.read(size)
self._bytes_left -= len(data)
return (data, data)
elif bytes_left < 0:
raise ProtocolError('Chunked-transfer overrun.')
elif bytes_left:
raise NetworkError('Connection closed.')
newline_data = yield from self._connection.readline()
if len(newline_data) > 2:
# Should be either CRLF or LF
# This could our problem or the server's problem
raise ProtocolError('Error reading newline after chunk.')
self._chunk_size = self._bytes_left = None
return (b'', newline_data)
def response_callback(request):
request.prepare_for_send()
self.assertTrue(request.url_info.url.endswith('robots.txt'))
response = Response(302, 'See else')
response.request = request
response.fields['Location'] = '/robots.txt'
nonlocal_dict['counter'] += 1
if nonlocal_dict['counter'] > 20:
raise ProtocolError('Mock redirect loop error.')
return response
def parse_status_line(cls, string):
'''Parse the status line bytes.
Returns:
tuple: An tuple representing the version, code, and reason.
'''
match = re.match(br'(HTTP/1\.[01])[ \t]+([0-9]{1,3})[ \t]*([^\r\n]*)',
string)
if match:
groups = match.groups()
if len(groups) == 3:
return wpull.string.to_str(
(groups[0], int(groups[1]), groups[2]),
encoding='latin-1',
)
raise ProtocolError("Error parsing status line '{0}'".format(string))
def parse(self, data):
for line in data.splitlines(False):
match = re.match(br'(\d{3}|^)([ -]?)(.*)', line)
if not match:
raise ProtocolError('Failed to parse reply.')
if match.group(1) and match.group(2) == b' ':
assert self.code is None
self.code = int(match.group(1))
if self.text is None:
self.text = match.group(3).decode('utf-8',
errors='surrogateescape')
else:
self.text += '\r\n{0}'.format(match.group(3).decode(
'utf-8', errors='surrogateescape'))
def parse_status_line(cls, data):
'''Parse the status line bytes.
Returns:
tuple: An tuple representing the version, code, and reason.
'''
match = re.match(br'(HTTP/\d+\.\d+)[ \t]+([0-9]{1,3})[ \t]*([^\r\n]*)',
data)
if match:
groups = match.groups()
if len(groups) == 3:
return wpull.string.to_str(
(groups[0], int(groups[1]), groups[2]),
encoding='latin-1',
)
raise ProtocolError(
'Error parsing status line {line}".'.format(line=ascii(data)))
def parse_status_line(cls, string):
'''Parse the status line bytes.
Returns:
tuple: An tuple representing the method, resource path, and
version.
'''
match = re.match(br'([a-zA-Z]+)[ \t]+([^ \t]+)[ \t]+(HTTP/1\.[01])',
string)
if match:
groups = match.groups()
if len(groups) == 3:
return wpull.string.to_str(
(groups[0], groups[1], groups[2]),
encoding='latin-1',
)
raise ProtocolError('Error parsing status line ‘{0}’'.format(string))
def parse_status_line(self, data):
'''Parse the status line bytes.
Returns:
tuple: An tuple representing the method, URI, and
version.
'''
match = re.match(br'([a-zA-Z]+)[ \t]+([^ \t]+)[ \t]+(HTTP/\d+\.\d+)',
data)
if match:
groups = match.groups()
if len(groups) == 3:
return wpull.string.to_str(
(groups[0], groups[1], groups[2]),
encoding=self.encoding,
)
raise ProtocolError('Error parsing status line.')
def passive_mode(self):
'''Enable passive mode.
Returns:
tuple: The address (IP address, port) of the passive port.
Coroutine.
'''
yield From(self._control_stream.write_command(Command('PASV')))
reply = yield From(self._control_stream.read_reply())
self.raise_if_not_match('Passive mode',
ReplyCodes.entering_passive_mode, reply)
try:
raise Return(wpull.ftp.util.parse_address(reply.text))
except ValueError as error:
raise ProtocolError(str(error)) from error
def _process_request(self, request, response_factory):
'''Fulfill a single request.
Returns:
Response
'''
yield self._connect()
request.address = self._resolved_address
self._events.pre_request(request)
if sys.version_info < (3, 3):
error_class = (socket.error, StreamClosedError, ssl.SSLError)
else:
error_class = (ConnectionError, StreamClosedError, ssl.SSLError)
if not self._params.keep_alive and 'Connection' not in request.fields:
request.fields['Connection'] = 'close'
try:
yield self._send_request_header(request)
yield self._send_request_body(request)
self._events.request.fire(request)
response = yield self._read_response_header(response_factory)
# TODO: handle 100 Continue
yield self._read_response_body(request, response)
except error_class as error:
raise NetworkError('Network error: {0}'.format(error)) from error
except BufferFullError as error:
raise ProtocolError(*error.args) from error
self._events.response.fire(response)
if self.should_close(request.version,
response.fields.get('Connection')):
_logger.debug('HTTP connection close.')
self.close()
else:
self._io_stream.monitor_for_close()
raise tornado.gen.Return(response)
def _read_request_header(self):
request = Request()
for dummy in range(100):
line = yield From(self._reader.readline())
_logger.debug(__('Got line {0}', line))
if line[-1:] != b'\n':
return
if not line.strip():
break
request.parse(line)
else:
raise ProtocolError('Request has too many headers.')
raise Return(request)
def handle_error(self, item_session: ItemSession,
error: BaseException) -> Actions:
'''Process an error.
Returns:
A value from :class:`.hook.Actions`.
'''
if not self._ssl_verification and \
isinstance(error, SSLVerificationError):
# Change it into a different error since the user doesn't care
# about verifying certificates
self._statistics.increment_error(ProtocolError())
else:
self._statistics.increment_error(error)
self._waiter.increment()
action = self.consult_error_hook(item_session, error)
if action == Actions.RETRY:
item_session.set_status(Status.error)
elif action == Actions.FINISH:
item_session.set_status(Status.done)
elif action == Actions.STOP:
raise HookStop('Script requested immediate stop.')
elif self._ssl_verification and isinstance(error,
SSLVerificationError):
raise
elif isinstance(error, ConnectionRefused) and \
not self.retry_connrefused:
item_session.set_status(Status.skipped)
elif isinstance(error, DNSNotFound) and \
not self.retry_dns_error:
item_session.set_status(Status.skipped)
else:
item_session.set_status(Status.error)
return action
def _stream_closed_callback(self):
_logger.debug('Stream closed. '
'active={0} connected={1} ' \
'closed={2} reading={3} writing={3}'.format(
self._active,
self.connected,
self._io_stream.closed(),
self._io_stream.reading(),
self._io_stream.writing())
)
if not self._active:
# We are likely in a context that's already dead
_logger.debug('Ignoring stream closed error={0}.'\
.format(self._io_stream.error))
return
if self._io_stream.error:
_logger.debug('Throwing error {0}.'.format(self._io_stream.error))
raise self._io_stream.error
if self._io_stream.buffer_full:
_logger.debug('Buffer full.')
raise ProtocolError('Buffer full.')
def download_listing(self, file: Optional[IO],
duration_timeout: Optional[float]=None) -> \
ListingResponse:
'''Read file listings.
Args:
file: A file object or asyncio stream.
duration_timeout: Maximum time in seconds of which the
entire file must be read.
Returns:
A Response populated the file listings
Be sure to call :meth:`start_file_listing` first.
Coroutine.
'''
if self._session_state != SessionState.directory_request_sent:
raise RuntimeError('File request not sent')
self._session_state = SessionState.file_request_sent
yield from self.download(file=file,
rewind=False,
duration_timeout=duration_timeout)
try:
if self._response.body.tell() == 0:
listings = ()
elif self._listing_type == 'mlsd':
self._response.body.seek(0)
machine_listings = wpull.protocol.ftp.util.parse_machine_listing(
self._response.body.read().decode(
'utf-8', errors='surrogateescape'),
convert=True,
strict=False)
listings = list(
wpull.protocol.ftp.util.machine_listings_to_file_entries(
machine_listings))
else:
self._response.body.seek(0)
file = io.TextIOWrapper(self._response.body,
encoding='utf-8',
errors='surrogateescape')
listing_parser = ListingParser(file=file)
listings = list(listing_parser.parse_input())
_logger.debug('Listing detected as %s', listing_parser.type)
# We don't want the file to be closed when exiting this function
file.detach()
except (ListingError, ValueError) as error:
raise ProtocolError(*error.args) from error
self._response.files = listings
self._response.body.seek(0)
self._session_state = SessionState.response_received
return self._response
