http1connection.py
Go to the documentation of this file.
00001 #!/usr/bin/env python
00002 #
00003 # Copyright 2014 Facebook
00004 #
00005 # Licensed under the Apache License, Version 2.0 (the "License"); you may
00006 # not use this file except in compliance with the License. You may obtain
00007 # a copy of the License at
00008 #
00009 #     http://www.apache.org/licenses/LICENSE-2.0
00010 #
00011 # Unless required by applicable law or agreed to in writing, software
00012 # distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
00013 # WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
00014 # License for the specific language governing permissions and limitations
00015 # under the License.
00016 
00017 """Client and server implementations of HTTP/1.x.
00018 
00019 .. versionadded:: 4.0
00020 """
00021 
00022 from __future__ import absolute_import, division, print_function, with_statement
00023 
00024 import re
00025 
00026 from tornado.concurrent import Future
00027 from tornado.escape import native_str, utf8
00028 from tornado import gen
00029 from tornado import httputil
00030 from tornado import iostream
00031 from tornado.log import gen_log, app_log
00032 from tornado import stack_context
00033 from tornado.util import GzipDecompressor
00034 
00035 
00036 class _QuietException(Exception):
00037     def __init__(self):
00038         pass
00039 
00040 class _ExceptionLoggingContext(object):
00041     """Used with the ``with`` statement when calling delegate methods to
00042     log any exceptions with the given logger.  Any exceptions caught are
00043     converted to _QuietException
00044     """
00045     def __init__(self, logger):
00046         self.logger = logger
00047 
00048     def __enter__(self):
00049         pass
00050 
00051     def __exit__(self, typ, value, tb):
00052         if value is not None:
00053             self.logger.error("Uncaught exception", exc_info=(typ, value, tb))
00054             raise _QuietException
00055 
00056 class HTTP1ConnectionParameters(object):
00057     """Parameters for `.HTTP1Connection` and `.HTTP1ServerConnection`.
00058     """
00059     def __init__(self, no_keep_alive=False, chunk_size=None,
00060                  max_header_size=None, header_timeout=None, max_body_size=None,
00061                  body_timeout=None, decompress=False):
00062         """
00063         :arg bool no_keep_alive: If true, always close the connection after
00064             one request.
00065         :arg int chunk_size: how much data to read into memory at once
00066         :arg int max_header_size:  maximum amount of data for HTTP headers
00067         :arg float header_timeout: how long to wait for all headers (seconds)
00068         :arg int max_body_size: maximum amount of data for body
00069         :arg float body_timeout: how long to wait while reading body (seconds)
00070         :arg bool decompress: if true, decode incoming
00071             ``Content-Encoding: gzip``
00072         """
00073         self.no_keep_alive = no_keep_alive
00074         self.chunk_size = chunk_size or 65536
00075         self.max_header_size = max_header_size or 65536
00076         self.header_timeout = header_timeout
00077         self.max_body_size = max_body_size
00078         self.body_timeout = body_timeout
00079         self.decompress = decompress
00080 
00081 
00082 class HTTP1Connection(httputil.HTTPConnection):
00083     """Implements the HTTP/1.x protocol.
00084 
00085     This class can be on its own for clients, or via `HTTP1ServerConnection`
00086     for servers.
00087     """
00088     def __init__(self, stream, is_client, params=None, context=None):
00089         """
00090         :arg stream: an `.IOStream`
00091         :arg bool is_client: client or server
00092         :arg params: a `.HTTP1ConnectionParameters` instance or ``None``
00093         :arg context: an opaque application-defined object that can be accessed
00094             as ``connection.context``.
00095         """
00096         self.is_client = is_client
00097         self.stream = stream
00098         if params is None:
00099             params = HTTP1ConnectionParameters()
00100         self.params = params
00101         self.context = context
00102         self.no_keep_alive = params.no_keep_alive
00103         # The body limits can be altered by the delegate, so save them
00104         # here instead of just referencing self.params later.
00105         self._max_body_size = (self.params.max_body_size or
00106                                self.stream.max_buffer_size)
00107         self._body_timeout = self.params.body_timeout
00108         # _write_finished is set to True when finish() has been called,
00109         # i.e. there will be no more data sent.  Data may still be in the
00110         # stream's write buffer.
00111         self._write_finished = False
00112         # True when we have read the entire incoming body.
00113         self._read_finished = False
00114         # _finish_future resolves when all data has been written and flushed
00115         # to the IOStream.
00116         self._finish_future = Future()
00117         # If true, the connection should be closed after this request
00118         # (after the response has been written in the server side,
00119         # and after it has been read in the client)
00120         self._disconnect_on_finish = False
00121         self._clear_callbacks()
00122         # Save the start lines after we read or write them; they
00123         # affect later processing (e.g. 304 responses and HEAD methods
00124         # have content-length but no bodies)
00125         self._request_start_line = None
00126         self._response_start_line = None
00127         self._request_headers = None
00128         # True if we are writing output with chunked encoding.
00129         self._chunking_output = None
00130         # While reading a body with a content-length, this is the
00131         # amount left to read.
00132         self._expected_content_remaining = None
00133         # A Future for our outgoing writes, returned by IOStream.write.
00134         self._pending_write = None
00135 
00136     def read_response(self, delegate):
00137         """Read a single HTTP response.
00138 
00139         Typical client-mode usage is to write a request using `write_headers`,
00140         `write`, and `finish`, and then call ``read_response``.
00141 
00142         :arg delegate: a `.HTTPMessageDelegate`
00143 
00144         Returns a `.Future` that resolves to None after the full response has
00145         been read.
00146         """
00147         if self.params.decompress:
00148             delegate = _GzipMessageDelegate(delegate, self.params.chunk_size)
00149         return self._read_message(delegate)
00150 
00151     @gen.coroutine
00152     def _read_message(self, delegate):
00153         need_delegate_close = False
00154         try:
00155             header_future = self.stream.read_until_regex(
00156                 b"\r?\n\r?\n",
00157                 max_bytes=self.params.max_header_size)
00158             if self.params.header_timeout is None:
00159                 header_data = yield header_future
00160             else:
00161                 try:
00162                     header_data = yield gen.with_timeout(
00163                         self.stream.io_loop.time() + self.params.header_timeout,
00164                         header_future,
00165                         io_loop=self.stream.io_loop)
00166                 except gen.TimeoutError:
00167                     self.close()
00168                     raise gen.Return(False)
00169             start_line, headers = self._parse_headers(header_data)
00170             if self.is_client:
00171                 start_line = httputil.parse_response_start_line(start_line)
00172                 self._response_start_line = start_line
00173             else:
00174                 start_line = httputil.parse_request_start_line(start_line)
00175                 self._request_start_line = start_line
00176                 self._request_headers = headers
00177 
00178             self._disconnect_on_finish = not self._can_keep_alive(
00179                 start_line, headers)
00180             need_delegate_close = True
00181             with _ExceptionLoggingContext(app_log):
00182                 header_future = delegate.headers_received(start_line, headers)
00183                 if header_future is not None:
00184                     yield header_future
00185             if self.stream is None:
00186                 # We've been detached.
00187                 need_delegate_close = False
00188                 raise gen.Return(False)
00189             skip_body = False
00190             if self.is_client:
00191                 if (self._request_start_line is not None and
00192                         self._request_start_line.method == 'HEAD'):
00193                     skip_body = True
00194                 code = start_line.code
00195                 if code == 304:
00196                     # 304 responses may include the content-length header
00197                     # but do not actually have a body.
00198                     # http://tools.ietf.org/html/rfc7230#section-3.3
00199                     skip_body = True
00200                 if code >= 100 and code < 200:
00201                     # 1xx responses should never indicate the presence of
00202                     # a body.
00203                     if ('Content-Length' in headers or
00204                         'Transfer-Encoding' in headers):
00205                         raise httputil.HTTPInputError(
00206                             "Response code %d cannot have body" % code)
00207                     # TODO: client delegates will get headers_received twice
00208                     # in the case of a 100-continue.  Document or change?
00209                     yield self._read_message(delegate)
00210             else:
00211                 if (headers.get("Expect") == "100-continue" and
00212                         not self._write_finished):
00213                     self.stream.write(b"HTTP/1.1 100 (Continue)\r\n\r\n")
00214             if not skip_body:
00215                 body_future = self._read_body(
00216                     start_line.code if self.is_client else 0, headers, delegate)
00217                 if body_future is not None:
00218                     if self._body_timeout is None:
00219                         yield body_future
00220                     else:
00221                         try:
00222                             yield gen.with_timeout(
00223                                 self.stream.io_loop.time() + self._body_timeout,
00224                                 body_future, self.stream.io_loop)
00225                         except gen.TimeoutError:
00226                             gen_log.info("Timeout reading body from %s",
00227                                          self.context)
00228                             self.stream.close()
00229                             raise gen.Return(False)
00230             self._read_finished = True
00231             if not self._write_finished or self.is_client:
00232                 need_delegate_close = False
00233                 with _ExceptionLoggingContext(app_log):
00234                     delegate.finish()
00235             # If we're waiting for the application to produce an asynchronous
00236             # response, and we're not detached, register a close callback
00237             # on the stream (we didn't need one while we were reading)
00238             if (not self._finish_future.done() and
00239                     self.stream is not None and
00240                     not self.stream.closed()):
00241                 self.stream.set_close_callback(self._on_connection_close)
00242                 yield self._finish_future
00243             if self.is_client and self._disconnect_on_finish:
00244                 self.close()
00245             if self.stream is None:
00246                 raise gen.Return(False)
00247         except httputil.HTTPInputError as e:
00248             gen_log.info("Malformed HTTP message from %s: %s",
00249                          self.context, e)
00250             self.close()
00251             raise gen.Return(False)
00252         finally:
00253             if need_delegate_close:
00254                 with _ExceptionLoggingContext(app_log):
00255                     delegate.on_connection_close()
00256             self._clear_callbacks()
00257         raise gen.Return(True)
00258 
00259     def _clear_callbacks(self):
00260         """Clears the callback attributes.
00261 
00262         This allows the request handler to be garbage collected more
00263         quickly in CPython by breaking up reference cycles.
00264         """
00265         self._write_callback = None
00266         self._write_future = None
00267         self._close_callback = None
00268         if self.stream is not None:
00269             self.stream.set_close_callback(None)
00270 
00271     def set_close_callback(self, callback):
00272         """Sets a callback that will be run when the connection is closed.
00273 
00274         .. deprecated:: 4.0
00275             Use `.HTTPMessageDelegate.on_connection_close` instead.
00276         """
00277         self._close_callback = stack_context.wrap(callback)
00278 
00279     def _on_connection_close(self):
00280         # Note that this callback is only registered on the IOStream
00281         # when we have finished reading the request and are waiting for
00282         # the application to produce its response.
00283         if self._close_callback is not None:
00284             callback = self._close_callback
00285             self._close_callback = None
00286             callback()
00287         if not self._finish_future.done():
00288             self._finish_future.set_result(None)
00289         self._clear_callbacks()
00290 
00291     def close(self):
00292         if self.stream is not None:
00293             self.stream.close()
00294         self._clear_callbacks()
00295         if not self._finish_future.done():
00296             self._finish_future.set_result(None)
00297 
00298     def detach(self):
00299         """Take control of the underlying stream.
00300 
00301         Returns the underlying `.IOStream` object and stops all further
00302         HTTP processing.  May only be called during
00303         `.HTTPMessageDelegate.headers_received`.  Intended for implementing
00304         protocols like websockets that tunnel over an HTTP handshake.
00305         """
00306         self._clear_callbacks()
00307         stream = self.stream
00308         self.stream = None
00309         return stream
00310 
00311     def set_body_timeout(self, timeout):
00312         """Sets the body timeout for a single request.
00313 
00314         Overrides the value from `.HTTP1ConnectionParameters`.
00315         """
00316         self._body_timeout = timeout
00317 
00318     def set_max_body_size(self, max_body_size):
00319         """Sets the body size limit for a single request.
00320 
00321         Overrides the value from `.HTTP1ConnectionParameters`.
00322         """
00323         self._max_body_size = max_body_size
00324 
00325     def write_headers(self, start_line, headers, chunk=None, callback=None):
00326         """Implements `.HTTPConnection.write_headers`."""
00327         if self.is_client:
00328             self._request_start_line = start_line
00329             # Client requests with a non-empty body must have either a
00330             # Content-Length or a Transfer-Encoding.
00331             self._chunking_output = (
00332                 start_line.method in ('POST', 'PUT', 'PATCH') and
00333                 'Content-Length' not in headers and
00334                 'Transfer-Encoding' not in headers)
00335         else:
00336             self._response_start_line = start_line
00337             self._chunking_output = (
00338                 # TODO: should this use
00339                 # self._request_start_line.version or
00340                 # start_line.version?
00341                 self._request_start_line.version == 'HTTP/1.1' and
00342                 # 304 responses have no body (not even a zero-length body), and so
00343                 # should not have either Content-Length or Transfer-Encoding.
00344                 # headers.
00345                 start_line.code != 304 and
00346                 # No need to chunk the output if a Content-Length is specified.
00347                 'Content-Length' not in headers and
00348                 # Applications are discouraged from touching Transfer-Encoding,
00349                 # but if they do, leave it alone.
00350                 'Transfer-Encoding' not in headers)
00351             # If a 1.0 client asked for keep-alive, add the header.
00352             if (self._request_start_line.version == 'HTTP/1.0' and
00353                 (self._request_headers.get('Connection', '').lower()
00354                  == 'keep-alive')):
00355                 headers['Connection'] = 'Keep-Alive'
00356         if self._chunking_output:
00357             headers['Transfer-Encoding'] = 'chunked'
00358         if (not self.is_client and
00359             (self._request_start_line.method == 'HEAD' or
00360              start_line.code == 304)):
00361             self._expected_content_remaining = 0
00362         elif 'Content-Length' in headers:
00363             self._expected_content_remaining = int(headers['Content-Length'])
00364         else:
00365             self._expected_content_remaining = None
00366         lines = [utf8("%s %s %s" % start_line)]
00367         lines.extend([utf8(n) + b": " + utf8(v) for n, v in headers.get_all()])
00368         for line in lines:
00369             if b'\n' in line:
00370                 raise ValueError('Newline in header: ' + repr(line))
00371         future = None
00372         if self.stream.closed():
00373             future = self._write_future = Future()
00374             future.set_exception(iostream.StreamClosedError())
00375         else:
00376             if callback is not None:
00377                 self._write_callback = stack_context.wrap(callback)
00378             else:
00379                 future = self._write_future = Future()
00380             data = b"\r\n".join(lines) + b"\r\n\r\n"
00381             if chunk:
00382                 data += self._format_chunk(chunk)
00383             self._pending_write = self.stream.write(data)
00384             self._pending_write.add_done_callback(self._on_write_complete)
00385         return future
00386 
00387     def _format_chunk(self, chunk):
00388         if self._expected_content_remaining is not None:
00389             self._expected_content_remaining -= len(chunk)
00390             if self._expected_content_remaining < 0:
00391                 # Close the stream now to stop further framing errors.
00392                 self.stream.close()
00393                 raise httputil.HTTPOutputError(
00394                     "Tried to write more data than Content-Length")
00395         if self._chunking_output and chunk:
00396             # Don't write out empty chunks because that means END-OF-STREAM
00397             # with chunked encoding
00398             return utf8("%x" % len(chunk)) + b"\r\n" + chunk + b"\r\n"
00399         else:
00400             return chunk
00401 
00402     def write(self, chunk, callback=None):
00403         """Implements `.HTTPConnection.write`.
00404 
00405         For backwards compatibility is is allowed but deprecated to
00406         skip `write_headers` and instead call `write()` with a
00407         pre-encoded header block.
00408         """
00409         future = None
00410         if self.stream.closed():
00411             future = self._write_future = Future()
00412             self._write_future.set_exception(iostream.StreamClosedError())
00413         else:
00414             if callback is not None:
00415                 self._write_callback = stack_context.wrap(callback)
00416             else:
00417                 future = self._write_future = Future()
00418             self._pending_write = self.stream.write(self._format_chunk(chunk))
00419             self._pending_write.add_done_callback(self._on_write_complete)
00420         return future
00421 
00422     def finish(self):
00423         """Implements `.HTTPConnection.finish`."""
00424         if (self._expected_content_remaining is not None and
00425                 self._expected_content_remaining != 0 and
00426                 not self.stream.closed()):
00427             self.stream.close()
00428             raise httputil.HTTPOutputError(
00429                 "Tried to write %d bytes less than Content-Length" %
00430                 self._expected_content_remaining)
00431         if self._chunking_output:
00432             if not self.stream.closed():
00433                 self._pending_write = self.stream.write(b"0\r\n\r\n")
00434                 self._pending_write.add_done_callback(self._on_write_complete)
00435         self._write_finished = True
00436         # If the app finished the request while we're still reading,
00437         # divert any remaining data away from the delegate and
00438         # close the connection when we're done sending our response.
00439         # Closing the connection is the only way to avoid reading the
00440         # whole input body.
00441         if not self._read_finished:
00442             self._disconnect_on_finish = True
00443         # No more data is coming, so instruct TCP to send any remaining
00444         # data immediately instead of waiting for a full packet or ack.
00445         self.stream.set_nodelay(True)
00446         if self._pending_write is None:
00447             self._finish_request(None)
00448         else:
00449             self._pending_write.add_done_callback(self._finish_request)
00450 
00451     def _on_write_complete(self, future):
00452         if self._write_callback is not None:
00453             callback = self._write_callback
00454             self._write_callback = None
00455             self.stream.io_loop.add_callback(callback)
00456         if self._write_future is not None:
00457             future = self._write_future
00458             self._write_future = None
00459             future.set_result(None)
00460 
00461     def _can_keep_alive(self, start_line, headers):
00462         if self.params.no_keep_alive:
00463             return False
00464         connection_header = headers.get("Connection")
00465         if connection_header is not None:
00466             connection_header = connection_header.lower()
00467         if start_line.version == "HTTP/1.1":
00468             return connection_header != "close"
00469         elif ("Content-Length" in headers
00470               or start_line.method in ("HEAD", "GET")):
00471             return connection_header == "keep-alive"
00472         return False
00473 
00474     def _finish_request(self, future):
00475         self._clear_callbacks()
00476         if not self.is_client and self._disconnect_on_finish:
00477             self.close()
00478             return
00479         # Turn Nagle's algorithm back on, leaving the stream in its
00480         # default state for the next request.
00481         self.stream.set_nodelay(False)
00482         if not self._finish_future.done():
00483             self._finish_future.set_result(None)
00484 
00485     def _parse_headers(self, data):
00486         data = native_str(data.decode('latin1'))
00487         eol = data.find("\r\n")
00488         start_line = data[:eol]
00489         try:
00490             headers = httputil.HTTPHeaders.parse(data[eol:])
00491         except ValueError:
00492             # probably form split() if there was no ':' in the line
00493             raise httputil.HTTPInputError("Malformed HTTP headers: %r" %
00494                                           data[eol:100])
00495         return start_line, headers
00496 
00497     def _read_body(self, code, headers, delegate):
00498         if "Content-Length" in headers:
00499             if "," in headers["Content-Length"]:
00500                 # Proxies sometimes cause Content-Length headers to get
00501                 # duplicated.  If all the values are identical then we can
00502                 # use them but if they differ it's an error.
00503                 pieces = re.split(r',\s*', headers["Content-Length"])
00504                 if any(i != pieces[0] for i in pieces):
00505                     raise httputil.HTTPInputError(
00506                         "Multiple unequal Content-Lengths: %r" %
00507                         headers["Content-Length"])
00508                 headers["Content-Length"] = pieces[0]
00509             content_length = int(headers["Content-Length"])
00510 
00511             if content_length > self._max_body_size:
00512                 raise httputil.HTTPInputError("Content-Length too long")
00513         else:
00514             content_length = None
00515 
00516         if code == 204:
00517             # This response code is not allowed to have a non-empty body,
00518             # and has an implicit length of zero instead of read-until-close.
00519             # http://www.w3.org/Protocols/rfc2616/rfc2616-sec4.html#sec4.3
00520             if ("Transfer-Encoding" in headers or
00521                     content_length not in (None, 0)):
00522                 raise httputil.HTTPInputError(
00523                     "Response with code %d should not have body" % code)
00524             content_length = 0
00525 
00526         if content_length is not None:
00527             return self._read_fixed_body(content_length, delegate)
00528         if headers.get("Transfer-Encoding") == "chunked":
00529             return self._read_chunked_body(delegate)
00530         if self.is_client:
00531             return self._read_body_until_close(delegate)
00532         return None
00533 
00534     @gen.coroutine
00535     def _read_fixed_body(self, content_length, delegate):
00536         while content_length > 0:
00537             body = yield self.stream.read_bytes(
00538                 min(self.params.chunk_size, content_length), partial=True)
00539             content_length -= len(body)
00540             if not self._write_finished or self.is_client:
00541                 with _ExceptionLoggingContext(app_log):
00542                     yield gen.maybe_future(delegate.data_received(body))
00543 
00544     @gen.coroutine
00545     def _read_chunked_body(self, delegate):
00546         # TODO: "chunk extensions" http://tools.ietf.org/html/rfc2616#section-3.6.1
00547         total_size = 0
00548         while True:
00549             chunk_len = yield self.stream.read_until(b"\r\n", max_bytes=64)
00550             chunk_len = int(chunk_len.strip(), 16)
00551             if chunk_len == 0:
00552                 return
00553             total_size += chunk_len
00554             if total_size > self._max_body_size:
00555                 raise httputil.HTTPInputError("chunked body too large")
00556             bytes_to_read = chunk_len
00557             while bytes_to_read:
00558                 chunk = yield self.stream.read_bytes(
00559                     min(bytes_to_read, self.params.chunk_size), partial=True)
00560                 bytes_to_read -= len(chunk)
00561                 if not self._write_finished or self.is_client:
00562                     with _ExceptionLoggingContext(app_log):
00563                         yield gen.maybe_future(delegate.data_received(chunk))
00564             # chunk ends with \r\n
00565             crlf = yield self.stream.read_bytes(2)
00566             assert crlf == b"\r\n"
00567 
00568     @gen.coroutine
00569     def _read_body_until_close(self, delegate):
00570         body = yield self.stream.read_until_close()
00571         if not self._write_finished or self.is_client:
00572             with _ExceptionLoggingContext(app_log):
00573                 delegate.data_received(body)
00574 
00575 
00576 class _GzipMessageDelegate(httputil.HTTPMessageDelegate):
00577     """Wraps an `HTTPMessageDelegate` to decode ``Content-Encoding: gzip``.
00578     """
00579     def __init__(self, delegate, chunk_size):
00580         self._delegate = delegate
00581         self._chunk_size = chunk_size
00582         self._decompressor = None
00583 
00584     def headers_received(self, start_line, headers):
00585         if headers.get("Content-Encoding") == "gzip":
00586             self._decompressor = GzipDecompressor()
00587             # Downstream delegates will only see uncompressed data,
00588             # so rename the content-encoding header.
00589             # (but note that curl_httpclient doesn't do this).
00590             headers.add("X-Consumed-Content-Encoding",
00591                         headers["Content-Encoding"])
00592             del headers["Content-Encoding"]
00593         return self._delegate.headers_received(start_line, headers)
00594 
00595     @gen.coroutine
00596     def data_received(self, chunk):
00597         if self._decompressor:
00598             compressed_data = chunk
00599             while compressed_data:
00600                 decompressed = self._decompressor.decompress(
00601                     compressed_data, self._chunk_size)
00602                 if decompressed:
00603                     yield gen.maybe_future(
00604                         self._delegate.data_received(decompressed))
00605                 compressed_data = self._decompressor.unconsumed_tail
00606         else:
00607             yield gen.maybe_future(self._delegate.data_received(chunk))
00608 
00609     def finish(self):
00610         if self._decompressor is not None:
00611             tail = self._decompressor.flush()
00612             if tail:
00613                 # I believe the tail will always be empty (i.e.
00614                 # decompress will return all it can).  The purpose
00615                 # of the flush call is to detect errors such
00616                 # as truncated input.  But in case it ever returns
00617                 # anything, treat it as an extra chunk
00618                 self._delegate.data_received(tail)
00619         return self._delegate.finish()
00620 
00621     def on_connection_close(self):
00622         return self._delegate.on_connection_close()
00623 
00624 
00625 class HTTP1ServerConnection(object):
00626     """An HTTP/1.x server."""
00627     def __init__(self, stream, params=None, context=None):
00628         """
00629         :arg stream: an `.IOStream`
00630         :arg params: a `.HTTP1ConnectionParameters` or None
00631         :arg context: an opaque application-defined object that is accessible
00632             as ``connection.context``
00633         """
00634         self.stream = stream
00635         if params is None:
00636             params = HTTP1ConnectionParameters()
00637         self.params = params
00638         self.context = context
00639         self._serving_future = None
00640 
00641     @gen.coroutine
00642     def close(self):
00643         """Closes the connection.
00644 
00645         Returns a `.Future` that resolves after the serving loop has exited.
00646         """
00647         self.stream.close()
00648         # Block until the serving loop is done, but ignore any exceptions
00649         # (start_serving is already responsible for logging them).
00650         try:
00651             yield self._serving_future
00652         except Exception:
00653             pass
00654 
00655     def start_serving(self, delegate):
00656         """Starts serving requests on this connection.
00657 
00658         :arg delegate: a `.HTTPServerConnectionDelegate`
00659         """
00660         assert isinstance(delegate, httputil.HTTPServerConnectionDelegate)
00661         self._serving_future = self._server_request_loop(delegate)
00662         # Register the future on the IOLoop so its errors get logged.
00663         self.stream.io_loop.add_future(self._serving_future,
00664                                        lambda f: f.result())
00665 
00666     @gen.coroutine
00667     def _server_request_loop(self, delegate):
00668         try:
00669             while True:
00670                 conn = HTTP1Connection(self.stream, False,
00671                                        self.params, self.context)
00672                 request_delegate = delegate.start_request(self, conn)
00673                 try:
00674                     ret = yield conn.read_response(request_delegate)
00675                 except (iostream.StreamClosedError,
00676                         iostream.UnsatisfiableReadError):
00677                     return
00678                 except _QuietException:
00679                     # This exception was already logged.
00680                     conn.close()
00681                     return
00682                 except Exception:
00683                     gen_log.error("Uncaught exception", exc_info=True)
00684                     conn.close()
00685                     return
00686                 if not ret:
00687                     return
00688                 yield gen.moment
00689         finally:
00690             delegate.on_close(self)


rosbridge_server
Author(s): Jonathan Mace
autogenerated on Thu Aug 27 2015 14:50:39