blob: 6de4b0e73cf5a17aa7635cf275c3589cbf1adf4e [file] [log] [blame]
Greg Stein5e0fa402000-06-26 08:28:01 +00001"""HTTP/1.1 client library
Guido van Rossum41999c11997-12-09 00:12:23 +00002
Greg Stein5e0fa402000-06-26 08:28:01 +00003<intro stuff goes here>
4<other stuff, too>
Guido van Rossum41999c11997-12-09 00:12:23 +00005
Thomas Wouters0e3f5912006-08-11 14:57:12 +00006HTTPConnection goes through a number of "states", which define when a client
Greg Stein5e0fa402000-06-26 08:28:01 +00007may legally make another request or fetch the response for a particular
8request. This diagram details these state transitions:
Guido van Rossum41999c11997-12-09 00:12:23 +00009
Greg Stein5e0fa402000-06-26 08:28:01 +000010 (null)
11 |
12 | HTTPConnection()
13 v
14 Idle
15 |
16 | putrequest()
17 v
18 Request-started
19 |
20 | ( putheader() )* endheaders()
21 v
22 Request-sent
23 |
24 | response = getresponse()
25 v
26 Unread-response [Response-headers-read]
27 |\____________________
Tim Peters5ceadc82001-01-13 19:16:21 +000028 | |
29 | response.read() | putrequest()
30 v v
31 Idle Req-started-unread-response
32 ______/|
33 / |
34 response.read() | | ( putheader() )* endheaders()
35 v v
36 Request-started Req-sent-unread-response
37 |
38 | response.read()
39 v
40 Request-sent
Greg Stein5e0fa402000-06-26 08:28:01 +000041
42This diagram presents the following rules:
43 -- a second request may not be started until {response-headers-read}
44 -- a response [object] cannot be retrieved until {request-sent}
45 -- there is no differentiation between an unread response body and a
46 partially read response body
47
48Note: this enforcement is applied by the HTTPConnection class. The
49 HTTPResponse class does not enforce this state machine, which
50 implies sophisticated clients may accelerate the request/response
51 pipeline. Caution should be taken, though: accelerating the states
52 beyond the above pattern may imply knowledge of the server's
53 connection-close behavior for certain requests. For example, it
54 is impossible to tell whether the server will close the connection
55 UNTIL the response headers have been read; this means that further
56 requests cannot be placed into the pipeline until it is known that
57 the server will NOT be closing the connection.
58
59Logical State __state __response
60------------- ------- ----------
61Idle _CS_IDLE None
62Request-started _CS_REQ_STARTED None
63Request-sent _CS_REQ_SENT None
64Unread-response _CS_IDLE <response_class>
65Req-started-unread-response _CS_REQ_STARTED <response_class>
66Req-sent-unread-response _CS_REQ_SENT <response_class>
Guido van Rossum41999c11997-12-09 00:12:23 +000067"""
Guido van Rossum23acc951994-02-21 16:36:04 +000068
Barry Warsaw820c1202008-06-12 04:06:45 +000069import email.parser
70import email.message
Jeremy Hylton636950f2009-03-28 04:34:21 +000071import io
72import os
73import socket
Senthil Kumaran7bc0d872010-12-19 10:49:52 +000074import collections
Jeremy Hylton1afc1692008-06-18 20:49:58 +000075from urllib.parse import urlsplit
Guido van Rossum23acc951994-02-21 16:36:04 +000076
Thomas Wouters47b49bf2007-08-30 22:15:33 +000077__all__ = ["HTTPResponse", "HTTPConnection",
Skip Montanaro951a8842001-06-01 16:25:38 +000078 "HTTPException", "NotConnected", "UnknownProtocol",
Jeremy Hylton7c75c992002-06-28 23:38:14 +000079 "UnknownTransferEncoding", "UnimplementedFileMode",
80 "IncompleteRead", "InvalidURL", "ImproperConnectionState",
81 "CannotSendRequest", "CannotSendHeader", "ResponseNotReady",
Georg Brandl6aab16e2006-02-17 19:17:25 +000082 "BadStatusLine", "error", "responses"]
Skip Montanaro2dd42762001-01-23 15:35:05 +000083
Guido van Rossum23acc951994-02-21 16:36:04 +000084HTTP_PORT = 80
Guido van Rossum09c8b6c1999-12-07 21:37:17 +000085HTTPS_PORT = 443
86
Greg Stein5e0fa402000-06-26 08:28:01 +000087_UNKNOWN = 'UNKNOWN'
88
89# connection states
90_CS_IDLE = 'Idle'
91_CS_REQ_STARTED = 'Request-started'
92_CS_REQ_SENT = 'Request-sent'
93
Martin v. Löwis39a31782004-09-18 09:03:49 +000094# status codes
95# informational
96CONTINUE = 100
97SWITCHING_PROTOCOLS = 101
98PROCESSING = 102
99
100# successful
101OK = 200
102CREATED = 201
103ACCEPTED = 202
104NON_AUTHORITATIVE_INFORMATION = 203
105NO_CONTENT = 204
106RESET_CONTENT = 205
107PARTIAL_CONTENT = 206
108MULTI_STATUS = 207
109IM_USED = 226
110
111# redirection
112MULTIPLE_CHOICES = 300
113MOVED_PERMANENTLY = 301
114FOUND = 302
115SEE_OTHER = 303
116NOT_MODIFIED = 304
117USE_PROXY = 305
118TEMPORARY_REDIRECT = 307
119
120# client error
121BAD_REQUEST = 400
122UNAUTHORIZED = 401
123PAYMENT_REQUIRED = 402
124FORBIDDEN = 403
125NOT_FOUND = 404
126METHOD_NOT_ALLOWED = 405
127NOT_ACCEPTABLE = 406
128PROXY_AUTHENTICATION_REQUIRED = 407
129REQUEST_TIMEOUT = 408
130CONFLICT = 409
131GONE = 410
132LENGTH_REQUIRED = 411
133PRECONDITION_FAILED = 412
134REQUEST_ENTITY_TOO_LARGE = 413
135REQUEST_URI_TOO_LONG = 414
136UNSUPPORTED_MEDIA_TYPE = 415
137REQUESTED_RANGE_NOT_SATISFIABLE = 416
138EXPECTATION_FAILED = 417
139UNPROCESSABLE_ENTITY = 422
140LOCKED = 423
141FAILED_DEPENDENCY = 424
142UPGRADE_REQUIRED = 426
Hynek Schlawack51b2ed52012-05-16 09:51:07 +0200143PRECONDITION_REQUIRED = 428
144TOO_MANY_REQUESTS = 429
145REQUEST_HEADER_FIELDS_TOO_LARGE = 431
Martin v. Löwis39a31782004-09-18 09:03:49 +0000146
147# server error
148INTERNAL_SERVER_ERROR = 500
149NOT_IMPLEMENTED = 501
150BAD_GATEWAY = 502
151SERVICE_UNAVAILABLE = 503
152GATEWAY_TIMEOUT = 504
153HTTP_VERSION_NOT_SUPPORTED = 505
154INSUFFICIENT_STORAGE = 507
155NOT_EXTENDED = 510
Hynek Schlawack51b2ed52012-05-16 09:51:07 +0200156NETWORK_AUTHENTICATION_REQUIRED = 511
Martin v. Löwis39a31782004-09-18 09:03:49 +0000157
Georg Brandl6aab16e2006-02-17 19:17:25 +0000158# Mapping status codes to official W3C names
159responses = {
160 100: 'Continue',
161 101: 'Switching Protocols',
162
163 200: 'OK',
164 201: 'Created',
165 202: 'Accepted',
166 203: 'Non-Authoritative Information',
167 204: 'No Content',
168 205: 'Reset Content',
169 206: 'Partial Content',
170
171 300: 'Multiple Choices',
172 301: 'Moved Permanently',
173 302: 'Found',
174 303: 'See Other',
175 304: 'Not Modified',
176 305: 'Use Proxy',
177 306: '(Unused)',
178 307: 'Temporary Redirect',
179
180 400: 'Bad Request',
181 401: 'Unauthorized',
182 402: 'Payment Required',
183 403: 'Forbidden',
184 404: 'Not Found',
185 405: 'Method Not Allowed',
186 406: 'Not Acceptable',
187 407: 'Proxy Authentication Required',
188 408: 'Request Timeout',
189 409: 'Conflict',
190 410: 'Gone',
191 411: 'Length Required',
192 412: 'Precondition Failed',
193 413: 'Request Entity Too Large',
194 414: 'Request-URI Too Long',
195 415: 'Unsupported Media Type',
196 416: 'Requested Range Not Satisfiable',
197 417: 'Expectation Failed',
Hynek Schlawack51b2ed52012-05-16 09:51:07 +0200198 428: 'Precondition Required',
199 429: 'Too Many Requests',
200 431: 'Request Header Fields Too Large',
Georg Brandl6aab16e2006-02-17 19:17:25 +0000201
202 500: 'Internal Server Error',
203 501: 'Not Implemented',
204 502: 'Bad Gateway',
205 503: 'Service Unavailable',
206 504: 'Gateway Timeout',
207 505: 'HTTP Version Not Supported',
Hynek Schlawack51b2ed52012-05-16 09:51:07 +0200208 511: 'Network Authentication Required',
Georg Brandl6aab16e2006-02-17 19:17:25 +0000209}
210
Georg Brandl80ba8e82005-09-29 20:16:07 +0000211# maximal amount of data to read at one time in _safe_read
212MAXAMOUNT = 1048576
213
Senthil Kumaran5466bf12010-12-18 16:55:23 +0000214# maximal line length when calling readline().
215_MAXLINE = 65536
Georg Brandlbf3f8eb2013-10-27 07:34:48 +0100216_MAXHEADERS = 100
217
Senthil Kumaran5466bf12010-12-18 16:55:23 +0000218
Barry Warsaw820c1202008-06-12 04:06:45 +0000219class HTTPMessage(email.message.Message):
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000220 # XXX The only usage of this method is in
221 # http.server.CGIHTTPRequestHandler. Maybe move the code there so
222 # that it doesn't need to be part of the public API. The API has
223 # never been defined so this could cause backwards compatibility
224 # issues.
225
Barry Warsaw820c1202008-06-12 04:06:45 +0000226 def getallmatchingheaders(self, name):
227 """Find all header lines matching a given header name.
Jeremy Hylton6d0a4c72002-07-07 16:51:37 +0000228
Barry Warsaw820c1202008-06-12 04:06:45 +0000229 Look through the list of headers and find all lines matching a given
230 header name (and their continuation lines). A list of the lines is
231 returned, without interpretation. If the header does not occur, an
232 empty list is returned. If the header occurs multiple times, all
233 occurrences are returned. Case is not important in the header name.
Jeremy Hylton6d0a4c72002-07-07 16:51:37 +0000234
Jeremy Hylton6d0a4c72002-07-07 16:51:37 +0000235 """
Barry Warsaw820c1202008-06-12 04:06:45 +0000236 name = name.lower() + ':'
237 n = len(name)
238 lst = []
239 hit = 0
240 for line in self.keys():
241 if line[:n].lower() == name:
242 hit = 1
243 elif not line[:1].isspace():
244 hit = 0
245 if hit:
246 lst.append(line)
247 return lst
Jeremy Hylton6d0a4c72002-07-07 16:51:37 +0000248
Jeremy Hylton98eb6c22009-03-27 18:31:36 +0000249def parse_headers(fp, _class=HTTPMessage):
Barry Warsaw820c1202008-06-12 04:06:45 +0000250 """Parses only RFC2822 headers from a file pointer.
251
252 email Parser wants to see strings rather than bytes.
253 But a TextIOWrapper around self.rfile would buffer too many bytes
254 from the stream, bytes which we later need to read as bytes.
255 So we read the correct bytes here, as bytes, for email Parser
256 to parse.
257
258 """
Barry Warsaw820c1202008-06-12 04:06:45 +0000259 headers = []
260 while True:
Senthil Kumaran5466bf12010-12-18 16:55:23 +0000261 line = fp.readline(_MAXLINE + 1)
262 if len(line) > _MAXLINE:
263 raise LineTooLong("header line")
Barry Warsaw820c1202008-06-12 04:06:45 +0000264 headers.append(line)
Georg Brandlbf3f8eb2013-10-27 07:34:48 +0100265 if len(headers) > _MAXHEADERS:
266 raise HTTPException("got more than %d headers" % _MAXHEADERS)
Barry Warsaw820c1202008-06-12 04:06:45 +0000267 if line in (b'\r\n', b'\n', b''):
268 break
269 hstring = b''.join(headers).decode('iso-8859-1')
Jeremy Hylton98eb6c22009-03-27 18:31:36 +0000270 return email.parser.Parser(_class=_class).parsestr(hstring)
Greg Stein5e0fa402000-06-26 08:28:01 +0000271
Jeremy Hyltond46aa372002-07-06 18:48:07 +0000272
Antoine Pitrou988dbd72010-12-17 17:35:56 +0000273class HTTPResponse(io.RawIOBase):
Jeremy Hyltond46aa372002-07-06 18:48:07 +0000274
275 # See RFC 2616 sec 19.6 and RFC 1945 sec 6 for details.
276
Jeremy Hylton811fc142007-08-03 13:30:02 +0000277 # The bytes from the socket object are iso-8859-1 strings.
278 # See RFC 2616 sec 2.2 which notes an exception for MIME-encoded
279 # text following RFC 2047. The basic status line parsing only
280 # accepts iso-8859-1.
281
Senthil Kumaran052ddb02013-03-18 14:11:41 -0700282 def __init__(self, sock, debuglevel=0, method=None, url=None):
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000283 # If the response includes a content-length header, we need to
284 # make sure that the client doesn't read more than the
Jeremy Hylton39b198d2007-08-04 19:22:00 +0000285 # specified number of bytes. If it does, it will block until
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000286 # the server times out and closes the connection. This will
287 # happen if a self.fp.read() is done (without a size) whether
288 # self.fp is buffered or not. So, no self.fp.read() by
289 # clients unless they know what they are doing.
Benjamin Petersonf72d9fb2009-02-08 00:29:20 +0000290 self.fp = sock.makefile("rb")
Jeremy Hylton30f86742000-09-18 22:50:38 +0000291 self.debuglevel = debuglevel
Jeremy Hyltonc1b2cb92003-05-05 16:13:58 +0000292 self._method = method
Greg Stein5e0fa402000-06-26 08:28:01 +0000293
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000294 # The HTTPResponse object is returned via urllib. The clients
295 # of http and urllib expect different attributes for the
296 # headers. headers is used here and supports urllib. msg is
297 # provided as a backwards compatibility layer for http
298 # clients.
299
300 self.headers = self.msg = None
Greg Stein5e0fa402000-06-26 08:28:01 +0000301
Greg Steindd6eefb2000-07-18 09:09:48 +0000302 # from the Status-Line of the response
Tim Peters07e99cb2001-01-14 23:47:14 +0000303 self.version = _UNKNOWN # HTTP-Version
304 self.status = _UNKNOWN # Status-Code
305 self.reason = _UNKNOWN # Reason-Phrase
Greg Stein5e0fa402000-06-26 08:28:01 +0000306
Tim Peters07e99cb2001-01-14 23:47:14 +0000307 self.chunked = _UNKNOWN # is "chunked" being used?
308 self.chunk_left = _UNKNOWN # bytes left to read in current chunk
309 self.length = _UNKNOWN # number of bytes left in response
310 self.will_close = _UNKNOWN # conn will close at end of response
Greg Stein5e0fa402000-06-26 08:28:01 +0000311
Jeremy Hyltonbe4fcf12002-06-28 22:38:01 +0000312 def _read_status(self):
Senthil Kumaran5466bf12010-12-18 16:55:23 +0000313 line = str(self.fp.readline(_MAXLINE + 1), "iso-8859-1")
314 if len(line) > _MAXLINE:
315 raise LineTooLong("status line")
Jeremy Hylton30f86742000-09-18 22:50:38 +0000316 if self.debuglevel > 0:
Guido van Rossumbe19ed72007-02-09 05:37:30 +0000317 print("reply:", repr(line))
Jeremy Hyltonb6769522003-06-29 17:55:05 +0000318 if not line:
319 # Presumably, the server closed the connection before
320 # sending a valid response.
321 raise BadStatusLine(line)
Greg Steindd6eefb2000-07-18 09:09:48 +0000322 try:
Antoine Pitrou988dbd72010-12-17 17:35:56 +0000323 version, status, reason = line.split(None, 2)
Greg Steindd6eefb2000-07-18 09:09:48 +0000324 except ValueError:
325 try:
Antoine Pitrou988dbd72010-12-17 17:35:56 +0000326 version, status = line.split(None, 1)
Greg Steindd6eefb2000-07-18 09:09:48 +0000327 reason = ""
328 except ValueError:
Antoine Pitrou988dbd72010-12-17 17:35:56 +0000329 # empty version will cause next test to fail.
Jeremy Hyltond46aa372002-07-06 18:48:07 +0000330 version = ""
Jeremy Hyltone5d0e842007-08-03 13:45:24 +0000331 if not version.startswith("HTTP/"):
Serhiy Storchakab5b9c8c2013-02-06 10:31:57 +0200332 self._close_conn()
Antoine Pitrou988dbd72010-12-17 17:35:56 +0000333 raise BadStatusLine(line)
Greg Stein5e0fa402000-06-26 08:28:01 +0000334
Jeremy Hylton23d40472001-04-13 14:57:08 +0000335 # The status code is a three-digit number
336 try:
Jeremy Hyltonbe4fcf12002-06-28 22:38:01 +0000337 status = int(status)
Jeremy Hylton23d40472001-04-13 14:57:08 +0000338 if status < 100 or status > 999:
339 raise BadStatusLine(line)
340 except ValueError:
341 raise BadStatusLine(line)
Jeremy Hyltonbe4fcf12002-06-28 22:38:01 +0000342 return version, status, reason
Greg Stein5e0fa402000-06-26 08:28:01 +0000343
Jeremy Hylton39c03802002-07-12 14:04:09 +0000344 def begin(self):
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000345 if self.headers is not None:
Jeremy Hyltonbe4fcf12002-06-28 22:38:01 +0000346 # we've already started reading the response
347 return
348
349 # read until we get a non-100 response
Raymond Hettingerb2e0b922003-02-26 22:45:18 +0000350 while True:
Jeremy Hyltonbe4fcf12002-06-28 22:38:01 +0000351 version, status, reason = self._read_status()
Martin v. Löwis39a31782004-09-18 09:03:49 +0000352 if status != CONTINUE:
Jeremy Hyltonbe4fcf12002-06-28 22:38:01 +0000353 break
354 # skip the header from the 100 response
Raymond Hettingerb2e0b922003-02-26 22:45:18 +0000355 while True:
Senthil Kumaran5466bf12010-12-18 16:55:23 +0000356 skip = self.fp.readline(_MAXLINE + 1)
357 if len(skip) > _MAXLINE:
358 raise LineTooLong("header line")
359 skip = skip.strip()
Jeremy Hyltonbe4fcf12002-06-28 22:38:01 +0000360 if not skip:
361 break
362 if self.debuglevel > 0:
Guido van Rossumbe19ed72007-02-09 05:37:30 +0000363 print("header:", skip)
Tim Petersc411dba2002-07-16 21:35:23 +0000364
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000365 self.code = self.status = status
Jeremy Hyltonbe4fcf12002-06-28 22:38:01 +0000366 self.reason = reason.strip()
Antoine Pitrou988dbd72010-12-17 17:35:56 +0000367 if version in ("HTTP/1.0", "HTTP/0.9"):
368 # Some servers might still return "0.9", treat it as 1.0 anyway
Greg Steindd6eefb2000-07-18 09:09:48 +0000369 self.version = 10
Jeremy Hyltone5d0e842007-08-03 13:45:24 +0000370 elif version.startswith("HTTP/1."):
Tim Peters07e99cb2001-01-14 23:47:14 +0000371 self.version = 11 # use HTTP/1.1 code for HTTP/1.x where x>=1
Greg Steindd6eefb2000-07-18 09:09:48 +0000372 else:
373 raise UnknownProtocol(version)
Greg Stein5e0fa402000-06-26 08:28:01 +0000374
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000375 self.headers = self.msg = parse_headers(self.fp)
Barry Warsaw820c1202008-06-12 04:06:45 +0000376
Jeremy Hylton30f86742000-09-18 22:50:38 +0000377 if self.debuglevel > 0:
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000378 for hdr in self.headers:
Jeremy Hyltone5d0e842007-08-03 13:45:24 +0000379 print("header:", hdr, end=" ")
Greg Stein5e0fa402000-06-26 08:28:01 +0000380
Greg Steindd6eefb2000-07-18 09:09:48 +0000381 # are we using the chunked-style of transfer encoding?
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000382 tr_enc = self.headers.get("transfer-encoding")
Jeremy Hyltond229b3a2002-09-03 19:24:24 +0000383 if tr_enc and tr_enc.lower() == "chunked":
Jeremy Hylton236156f2008-12-15 03:00:50 +0000384 self.chunked = True
Greg Steindd6eefb2000-07-18 09:09:48 +0000385 self.chunk_left = None
386 else:
Jeremy Hylton236156f2008-12-15 03:00:50 +0000387 self.chunked = False
Greg Stein5e0fa402000-06-26 08:28:01 +0000388
Greg Steindd6eefb2000-07-18 09:09:48 +0000389 # will the connection close at the end of the response?
Jeremy Hylton22b3a492002-11-13 17:27:43 +0000390 self.will_close = self._check_close()
Greg Stein5e0fa402000-06-26 08:28:01 +0000391
Greg Steindd6eefb2000-07-18 09:09:48 +0000392 # do we have a Content-Length?
393 # NOTE: RFC 2616, S4.4, #3 says we ignore this if tr_enc is "chunked"
Jeremy Hylton0ee5eeb2007-08-04 03:25:17 +0000394 self.length = None
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000395 length = self.headers.get("content-length")
Barry Warsaw820c1202008-06-12 04:06:45 +0000396
397 # are we using the chunked-style of transfer encoding?
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000398 tr_enc = self.headers.get("transfer-encoding")
Greg Steindd6eefb2000-07-18 09:09:48 +0000399 if length and not self.chunked:
Jeremy Hylton30a81812000-09-14 20:34:27 +0000400 try:
401 self.length = int(length)
402 except ValueError:
Christian Heimesa612dc02008-02-24 13:08:18 +0000403 self.length = None
404 else:
405 if self.length < 0: # ignore nonsensical negative lengths
406 self.length = None
407 else:
408 self.length = None
Greg Stein5e0fa402000-06-26 08:28:01 +0000409
Greg Steindd6eefb2000-07-18 09:09:48 +0000410 # does the body have a fixed length? (of zero)
Martin v. Löwis39a31782004-09-18 09:03:49 +0000411 if (status == NO_CONTENT or status == NOT_MODIFIED or
Jeremy Hyltonc1b2cb92003-05-05 16:13:58 +0000412 100 <= status < 200 or # 1xx codes
Jeremy Hyltone5d0e842007-08-03 13:45:24 +0000413 self._method == "HEAD"):
Greg Steindd6eefb2000-07-18 09:09:48 +0000414 self.length = 0
Greg Stein5e0fa402000-06-26 08:28:01 +0000415
Greg Steindd6eefb2000-07-18 09:09:48 +0000416 # if the connection remains open, and we aren't using chunked, and
417 # a content-length was not provided, then assume that the connection
418 # WILL close.
Jeremy Hylton0ee5eeb2007-08-04 03:25:17 +0000419 if (not self.will_close and
420 not self.chunked and
421 self.length is None):
Jeremy Hylton236156f2008-12-15 03:00:50 +0000422 self.will_close = True
Greg Stein5e0fa402000-06-26 08:28:01 +0000423
Jeremy Hylton22b3a492002-11-13 17:27:43 +0000424 def _check_close(self):
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000425 conn = self.headers.get("connection")
Jeremy Hylton22b3a492002-11-13 17:27:43 +0000426 if self.version == 11:
427 # An HTTP/1.1 proxy is assumed to stay open unless
428 # explicitly closed.
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000429 conn = self.headers.get("connection")
Raymond Hettingerbac788a2004-05-04 09:21:43 +0000430 if conn and "close" in conn.lower():
Jeremy Hylton22b3a492002-11-13 17:27:43 +0000431 return True
432 return False
433
Jeremy Hylton2c178252004-08-07 16:28:14 +0000434 # Some HTTP/1.0 implementations have support for persistent
435 # connections, using rules different than HTTP/1.1.
Jeremy Hylton22b3a492002-11-13 17:27:43 +0000436
Christian Heimes895627f2007-12-08 17:28:33 +0000437 # For older HTTP, Keep-Alive indicates persistent connection.
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000438 if self.headers.get("keep-alive"):
Jeremy Hylton22b3a492002-11-13 17:27:43 +0000439 return False
Tim Peters77c06fb2002-11-24 02:35:35 +0000440
Jeremy Hylton2c178252004-08-07 16:28:14 +0000441 # At least Akamai returns a "Connection: Keep-Alive" header,
442 # which was supposed to be sent by the client.
443 if conn and "keep-alive" in conn.lower():
444 return False
445
Jeremy Hylton22b3a492002-11-13 17:27:43 +0000446 # Proxy-Connection is a netscape hack.
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000447 pconn = self.headers.get("proxy-connection")
Raymond Hettingerbac788a2004-05-04 09:21:43 +0000448 if pconn and "keep-alive" in pconn.lower():
Jeremy Hylton22b3a492002-11-13 17:27:43 +0000449 return False
450
451 # otherwise, assume it will close
452 return True
453
Serhiy Storchakab5b9c8c2013-02-06 10:31:57 +0200454 def _close_conn(self):
455 fp = self.fp
456 self.fp = None
457 fp.close()
458
Greg Steindd6eefb2000-07-18 09:09:48 +0000459 def close(self):
Serhiy Storchakab5b9c8c2013-02-06 10:31:57 +0200460 super().close() # set "closed" flag
Greg Steindd6eefb2000-07-18 09:09:48 +0000461 if self.fp:
Serhiy Storchakab5b9c8c2013-02-06 10:31:57 +0200462 self._close_conn()
Greg Stein5e0fa402000-06-26 08:28:01 +0000463
Jeremy Hyltondf5f6b52007-08-08 17:36:33 +0000464 # These implementations are for the benefit of io.BufferedReader.
465
466 # XXX This class should probably be revised to act more like
467 # the "raw stream" that BufferedReader expects.
468
Jeremy Hyltondf5f6b52007-08-08 17:36:33 +0000469 def flush(self):
Serhiy Storchakab5b9c8c2013-02-06 10:31:57 +0200470 super().flush()
471 if self.fp:
472 self.fp.flush()
Jeremy Hyltondf5f6b52007-08-08 17:36:33 +0000473
Jeremy Hyltona7cff022009-04-01 02:35:56 +0000474 def readable(self):
475 return True
476
Jeremy Hyltondf5f6b52007-08-08 17:36:33 +0000477 # End of "raw stream" methods
478
Greg Steindd6eefb2000-07-18 09:09:48 +0000479 def isclosed(self):
Serhiy Storchakab5b9c8c2013-02-06 10:31:57 +0200480 """True if the connection is closed."""
Greg Steindd6eefb2000-07-18 09:09:48 +0000481 # NOTE: it is possible that we will not ever call self.close(). This
482 # case occurs when will_close is TRUE, length is None, and we
483 # read up to the last byte, but NOT past it.
484 #
485 # IMPLIES: if will_close is FALSE, then self.close() will ALWAYS be
486 # called, meaning self.isclosed() is meaningful.
487 return self.fp is None
488
489 def read(self, amt=None):
490 if self.fp is None:
Guido van Rossum8ce8a782007-11-01 19:42:39 +0000491 return b""
Greg Steindd6eefb2000-07-18 09:09:48 +0000492
Senthil Kumaran71fb6c82010-04-28 17:39:48 +0000493 if self._method == "HEAD":
Serhiy Storchakab5b9c8c2013-02-06 10:31:57 +0200494 self._close_conn()
Senthil Kumaran71fb6c82010-04-28 17:39:48 +0000495 return b""
496
Antoine Pitrou38d96432011-12-06 22:33:57 +0100497 if amt is not None:
498 # Amount is given, so call base class version
499 # (which is implemented in terms of self.readinto)
500 return super(HTTPResponse, self).read(amt)
501 else:
502 # Amount is not given (unbounded read) so we must check self.length
503 # and self.chunked
Tim Peters230a60c2002-11-09 05:08:07 +0000504
Antoine Pitrou38d96432011-12-06 22:33:57 +0100505 if self.chunked:
506 return self._readall_chunked()
507
Jeremy Hyltondef9d2a2004-11-07 16:13:49 +0000508 if self.length is None:
Greg Steindd6eefb2000-07-18 09:09:48 +0000509 s = self.fp.read()
510 else:
Antoine Pitroubeec61a2013-02-02 22:49:34 +0100511 try:
512 s = self._safe_read(self.length)
513 except IncompleteRead:
Serhiy Storchakab5b9c8c2013-02-06 10:31:57 +0200514 self._close_conn()
Antoine Pitroubeec61a2013-02-02 22:49:34 +0100515 raise
Jeremy Hyltondef9d2a2004-11-07 16:13:49 +0000516 self.length = 0
Serhiy Storchakab5b9c8c2013-02-06 10:31:57 +0200517 self._close_conn() # we read everything
Greg Steindd6eefb2000-07-18 09:09:48 +0000518 return s
519
Antoine Pitrou38d96432011-12-06 22:33:57 +0100520 def readinto(self, b):
521 if self.fp is None:
522 return 0
523
524 if self._method == "HEAD":
Serhiy Storchakab6c86fd2013-02-06 10:35:40 +0200525 self._close_conn()
Antoine Pitrou38d96432011-12-06 22:33:57 +0100526 return 0
527
528 if self.chunked:
529 return self._readinto_chunked(b)
530
Greg Steindd6eefb2000-07-18 09:09:48 +0000531 if self.length is not None:
Antoine Pitrou38d96432011-12-06 22:33:57 +0100532 if len(b) > self.length:
Greg Steindd6eefb2000-07-18 09:09:48 +0000533 # clip the read to the "end of response"
Antoine Pitrou38d96432011-12-06 22:33:57 +0100534 b = memoryview(b)[0:self.length]
Greg Steindd6eefb2000-07-18 09:09:48 +0000535
536 # we do not use _safe_read() here because this may be a .will_close
537 # connection, and the user is reading more bytes than will be provided
538 # (for example, reading in 1k chunks)
Antoine Pitrou38d96432011-12-06 22:33:57 +0100539 n = self.fp.readinto(b)
Serhiy Storchaka1c84ac12013-12-17 21:50:02 +0200540 if not n and b:
Antoine Pitroubeec61a2013-02-02 22:49:34 +0100541 # Ideally, we would raise IncompleteRead if the content-length
542 # wasn't satisfied, but it might break compatibility.
Serhiy Storchakab5b9c8c2013-02-06 10:31:57 +0200543 self._close_conn()
Antoine Pitrou6a35e182013-02-02 23:04:56 +0100544 elif self.length is not None:
Antoine Pitrou38d96432011-12-06 22:33:57 +0100545 self.length -= n
Guido van Rossum8ce8a782007-11-01 19:42:39 +0000546 if not self.length:
Serhiy Storchakab5b9c8c2013-02-06 10:31:57 +0200547 self._close_conn()
Antoine Pitrou38d96432011-12-06 22:33:57 +0100548 return n
Greg Steindd6eefb2000-07-18 09:09:48 +0000549
Antoine Pitrou38d96432011-12-06 22:33:57 +0100550 def _read_next_chunk_size(self):
551 # Read the next chunk size from the file
552 line = self.fp.readline(_MAXLINE + 1)
553 if len(line) > _MAXLINE:
554 raise LineTooLong("chunk size")
555 i = line.find(b";")
556 if i >= 0:
557 line = line[:i] # strip chunk-extensions
558 try:
559 return int(line, 16)
560 except ValueError:
561 # close the connection as protocol synchronisation is
562 # probably lost
Serhiy Storchakab6c86fd2013-02-06 10:35:40 +0200563 self._close_conn()
Antoine Pitrou38d96432011-12-06 22:33:57 +0100564 raise
Jeremy Hyltond4c472c2002-09-03 20:49:06 +0000565
Antoine Pitrou38d96432011-12-06 22:33:57 +0100566 def _read_and_discard_trailer(self):
Jeremy Hyltond4c472c2002-09-03 20:49:06 +0000567 # read and discard trailer up to the CRLF terminator
568 ### note: we shouldn't have any trailers!
Raymond Hettingerb2e0b922003-02-26 22:45:18 +0000569 while True:
Senthil Kumaran5466bf12010-12-18 16:55:23 +0000570 line = self.fp.readline(_MAXLINE + 1)
571 if len(line) > _MAXLINE:
572 raise LineTooLong("trailer line")
Christian Heimes0bd4e112008-02-12 22:59:25 +0000573 if not line:
574 # a vanishingly small number of sites EOF without
575 # sending the trailer
576 break
Senthil Kumaran7e70a5c2012-04-29 10:39:49 +0800577 if line in (b'\r\n', b'\n', b''):
Jeremy Hyltond4c472c2002-09-03 20:49:06 +0000578 break
579
Antoine Pitrou38d96432011-12-06 22:33:57 +0100580 def _readall_chunked(self):
581 assert self.chunked != _UNKNOWN
582 chunk_left = self.chunk_left
583 value = []
584 while True:
585 if chunk_left is None:
586 try:
587 chunk_left = self._read_next_chunk_size()
588 if chunk_left == 0:
589 break
590 except ValueError:
591 raise IncompleteRead(b''.join(value))
592 value.append(self._safe_read(chunk_left))
593
594 # we read the whole chunk, get another
595 self._safe_read(2) # toss the CRLF at the end of the chunk
596 chunk_left = None
597
598 self._read_and_discard_trailer()
599
Jeremy Hyltond4c472c2002-09-03 20:49:06 +0000600 # we read everything; close the "file"
Serhiy Storchakab5b9c8c2013-02-06 10:31:57 +0200601 self._close_conn()
Jeremy Hyltond4c472c2002-09-03 20:49:06 +0000602
Chris Withers9c33cbf2009-09-04 17:04:16 +0000603 return b''.join(value)
Tim Peters230a60c2002-11-09 05:08:07 +0000604
Antoine Pitrou38d96432011-12-06 22:33:57 +0100605 def _readinto_chunked(self, b):
606 assert self.chunked != _UNKNOWN
607 chunk_left = self.chunk_left
608
609 total_bytes = 0
610 mvb = memoryview(b)
611 while True:
612 if chunk_left is None:
613 try:
614 chunk_left = self._read_next_chunk_size()
615 if chunk_left == 0:
616 break
617 except ValueError:
618 raise IncompleteRead(bytes(b[0:total_bytes]))
Antoine Pitrou4ce6aa42011-12-06 22:34:36 +0100619
Antoine Pitrou38d96432011-12-06 22:33:57 +0100620 if len(mvb) < chunk_left:
621 n = self._safe_readinto(mvb)
622 self.chunk_left = chunk_left - n
Antoine Pitrouf7e78182012-01-04 18:57:22 +0100623 return total_bytes + n
Antoine Pitrou38d96432011-12-06 22:33:57 +0100624 elif len(mvb) == chunk_left:
625 n = self._safe_readinto(mvb)
626 self._safe_read(2) # toss the CRLF at the end of the chunk
627 self.chunk_left = None
Antoine Pitrouf7e78182012-01-04 18:57:22 +0100628 return total_bytes + n
Antoine Pitrou38d96432011-12-06 22:33:57 +0100629 else:
630 temp_mvb = mvb[0:chunk_left]
631 n = self._safe_readinto(temp_mvb)
632 mvb = mvb[n:]
633 total_bytes += n
634
635 # we read the whole chunk, get another
636 self._safe_read(2) # toss the CRLF at the end of the chunk
637 chunk_left = None
638
639 self._read_and_discard_trailer()
640
641 # we read everything; close the "file"
Serhiy Storchakab6c86fd2013-02-06 10:35:40 +0200642 self._close_conn()
Antoine Pitrou38d96432011-12-06 22:33:57 +0100643
644 return total_bytes
645
Greg Steindd6eefb2000-07-18 09:09:48 +0000646 def _safe_read(self, amt):
647 """Read the number of bytes requested, compensating for partial reads.
648
649 Normally, we have a blocking socket, but a read() can be interrupted
650 by a signal (resulting in a partial read).
651
652 Note that we cannot distinguish between EOF and an interrupt when zero
653 bytes have been read. IncompleteRead() will be raised in this
654 situation.
655
656 This function should be used when <amt> bytes "should" be present for
657 reading. If the bytes are truly not available (due to EOF), then the
658 IncompleteRead exception can be used to detect the problem.
659 """
Georg Brandl80ba8e82005-09-29 20:16:07 +0000660 s = []
Greg Steindd6eefb2000-07-18 09:09:48 +0000661 while amt > 0:
Georg Brandl80ba8e82005-09-29 20:16:07 +0000662 chunk = self.fp.read(min(amt, MAXAMOUNT))
Greg Steindd6eefb2000-07-18 09:09:48 +0000663 if not chunk:
Benjamin Peterson6accb982009-03-02 22:50:25 +0000664 raise IncompleteRead(b''.join(s), amt)
Georg Brandl80ba8e82005-09-29 20:16:07 +0000665 s.append(chunk)
Raymond Hettingerb2e0b922003-02-26 22:45:18 +0000666 amt -= len(chunk)
Guido van Rossuma00f1232007-09-12 19:43:09 +0000667 return b"".join(s)
Greg Steindd6eefb2000-07-18 09:09:48 +0000668
Antoine Pitrou38d96432011-12-06 22:33:57 +0100669 def _safe_readinto(self, b):
670 """Same as _safe_read, but for reading into a buffer."""
671 total_bytes = 0
672 mvb = memoryview(b)
673 while total_bytes < len(b):
674 if MAXAMOUNT < len(mvb):
675 temp_mvb = mvb[0:MAXAMOUNT]
676 n = self.fp.readinto(temp_mvb)
677 else:
678 n = self.fp.readinto(mvb)
679 if not n:
680 raise IncompleteRead(bytes(mvb[0:total_bytes]), len(b))
681 mvb = mvb[n:]
682 total_bytes += n
683 return total_bytes
684
Antoine Pitroub353c122009-02-11 00:39:14 +0000685 def fileno(self):
686 return self.fp.fileno()
687
Greg Steindd6eefb2000-07-18 09:09:48 +0000688 def getheader(self, name, default=None):
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000689 if self.headers is None:
Greg Steindd6eefb2000-07-18 09:09:48 +0000690 raise ResponseNotReady()
Senthil Kumaran9f8dc442010-08-02 11:04:58 +0000691 headers = self.headers.get_all(name) or default
692 if isinstance(headers, str) or not hasattr(headers, '__iter__'):
693 return headers
694 else:
695 return ', '.join(headers)
Greg Stein5e0fa402000-06-26 08:28:01 +0000696
Martin v. Löwisdeacce22004-08-18 12:46:26 +0000697 def getheaders(self):
698 """Return list of (header, value) tuples."""
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000699 if self.headers is None:
Martin v. Löwisdeacce22004-08-18 12:46:26 +0000700 raise ResponseNotReady()
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000701 return list(self.headers.items())
Martin v. Löwisdeacce22004-08-18 12:46:26 +0000702
Antoine Pitroub353c122009-02-11 00:39:14 +0000703 # We override IOBase.__iter__ so that it doesn't check for closed-ness
704
705 def __iter__(self):
706 return self
707
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000708 # For compatibility with old-style urllib responses.
709
710 def info(self):
711 return self.headers
712
713 def geturl(self):
714 return self.url
715
716 def getcode(self):
717 return self.status
Greg Stein5e0fa402000-06-26 08:28:01 +0000718
719class HTTPConnection:
720
Greg Steindd6eefb2000-07-18 09:09:48 +0000721 _http_vsn = 11
722 _http_vsn_str = 'HTTP/1.1'
Greg Stein5e0fa402000-06-26 08:28:01 +0000723
Greg Steindd6eefb2000-07-18 09:09:48 +0000724 response_class = HTTPResponse
725 default_port = HTTP_PORT
726 auto_open = 1
Jeremy Hylton30f86742000-09-18 22:50:38 +0000727 debuglevel = 0
Antoine Pitrou90e47742013-01-02 22:10:47 +0100728 # TCP Maximum Segment Size (MSS) is determined by the TCP stack on
729 # a per-connection basis. There is no simple and efficient
730 # platform independent mechanism for determining the MSS, so
731 # instead a reasonable estimate is chosen. The getsockopt()
732 # interface using the TCP_MAXSEG parameter may be a suitable
733 # approach on some operating systems. A value of 16KiB is chosen
734 # as a reasonable estimate of the maximum MSS.
735 mss = 16384
Greg Stein5e0fa402000-06-26 08:28:01 +0000736
Senthil Kumaran052ddb02013-03-18 14:11:41 -0700737 def __init__(self, host, port=None, timeout=socket._GLOBAL_DEFAULT_TIMEOUT,
738 source_address=None):
Guido van Rossumd8faa362007-04-27 19:54:29 +0000739 self.timeout = timeout
Gregory P. Smithb4066372010-01-03 03:28:29 +0000740 self.source_address = source_address
Greg Steindd6eefb2000-07-18 09:09:48 +0000741 self.sock = None
Jeremy Hylton8531b1b2002-07-16 21:21:11 +0000742 self._buffer = []
Greg Steindd6eefb2000-07-18 09:09:48 +0000743 self.__response = None
744 self.__state = _CS_IDLE
Jeremy Hyltonc1b2cb92003-05-05 16:13:58 +0000745 self._method = None
Senthil Kumaran97f0c6b2009-07-25 04:24:38 +0000746 self._tunnel_host = None
747 self._tunnel_port = None
Georg Brandlc7c199f2011-01-22 22:06:24 +0000748 self._tunnel_headers = {}
Tim Petersc411dba2002-07-16 21:35:23 +0000749
Senthil Kumaran9da047b2014-04-14 13:07:56 -0400750 (self.host, self.port) = self._get_hostport(host, port)
751
752 # This is stored as an instance variable to allow unit
753 # tests to replace it with a suitable mockup
754 self._create_connection = socket.create_connection
Greg Stein5e0fa402000-06-26 08:28:01 +0000755
Senthil Kumaran47fff872009-12-20 07:10:31 +0000756 def set_tunnel(self, host, port=None, headers=None):
Senthil Kumaran9da047b2014-04-14 13:07:56 -0400757 """Set up host and port for HTTP CONNECT tunnelling.
Senthil Kumaran47fff872009-12-20 07:10:31 +0000758
Senthil Kumaran9da047b2014-04-14 13:07:56 -0400759 In a connection that uses HTTP CONNECT tunneling, the host passed to the
760 constructor is used as a proxy server that relays all communication to
761 the endpoint passed to `set_tunnel`. This done by sending an HTTP
762 CONNECT request to the proxy server when the connection is established.
763
764 This method must be called before the HTML connection has been
765 established.
766
767 The headers argument should be a mapping of extra HTTP headers to send
768 with the CONNECT request.
Senthil Kumaran47fff872009-12-20 07:10:31 +0000769 """
Senthil Kumaran9da047b2014-04-14 13:07:56 -0400770
771 if self.sock:
772 raise RuntimeError("Can't set up tunnel for established connection")
773
Serhiy Storchaka4ac7ed92014-12-12 09:29:15 +0200774 self._tunnel_host, self._tunnel_port = self._get_hostport(host, port)
Senthil Kumaran47fff872009-12-20 07:10:31 +0000775 if headers:
776 self._tunnel_headers = headers
777 else:
778 self._tunnel_headers.clear()
Senthil Kumaran97f0c6b2009-07-25 04:24:38 +0000779
Senthil Kumaran9da047b2014-04-14 13:07:56 -0400780 def _get_hostport(self, host, port):
Greg Steindd6eefb2000-07-18 09:09:48 +0000781 if port is None:
Skip Montanaro10e6e0e2004-09-14 16:32:02 +0000782 i = host.rfind(':')
Skip Montanarocae14d22004-09-14 17:55:21 +0000783 j = host.rfind(']') # ipv6 addresses have [...]
784 if i > j:
Skip Montanaro9d389972002-03-24 16:53:50 +0000785 try:
786 port = int(host[i+1:])
787 except ValueError:
Łukasz Langaa5a9a9c2011-10-18 21:17:39 +0200788 if host[i+1:] == "": # http://foo.com:/ == http://foo.com/
789 port = self.default_port
790 else:
791 raise InvalidURL("nonnumeric port: '%s'" % host[i+1:])
Greg Steindd6eefb2000-07-18 09:09:48 +0000792 host = host[:i]
793 else:
794 port = self.default_port
Raymond Hettinger4d037912004-10-14 15:23:38 +0000795 if host and host[0] == '[' and host[-1] == ']':
Brett Cannon0a1af4a2004-09-15 23:26:23 +0000796 host = host[1:-1]
Senthil Kumaran9da047b2014-04-14 13:07:56 -0400797
798 return (host, port)
Greg Stein5e0fa402000-06-26 08:28:01 +0000799
Jeremy Hylton30f86742000-09-18 22:50:38 +0000800 def set_debuglevel(self, level):
801 self.debuglevel = level
802
Senthil Kumaran97f0c6b2009-07-25 04:24:38 +0000803 def _tunnel(self):
Serhiy Storchaka4ac7ed92014-12-12 09:29:15 +0200804 connect_str = "CONNECT %s:%d HTTP/1.0\r\n" % (self._tunnel_host,
805 self._tunnel_port)
Senthil Kumaran97f0c6b2009-07-25 04:24:38 +0000806 connect_bytes = connect_str.encode("ascii")
807 self.send(connect_bytes)
Georg Brandlc7c199f2011-01-22 22:06:24 +0000808 for header, value in self._tunnel_headers.items():
Senthil Kumaran47fff872009-12-20 07:10:31 +0000809 header_str = "%s: %s\r\n" % (header, value)
Marc-André Lemburg8f36af72011-02-25 15:42:01 +0000810 header_bytes = header_str.encode("latin-1")
Senthil Kumaran47fff872009-12-20 07:10:31 +0000811 self.send(header_bytes)
Georg Brandlc7c199f2011-01-22 22:06:24 +0000812 self.send(b'\r\n')
Senthil Kumaran47fff872009-12-20 07:10:31 +0000813
Georg Brandlc7c199f2011-01-22 22:06:24 +0000814 response = self.response_class(self.sock, method=self._method)
Senthil Kumaran97f0c6b2009-07-25 04:24:38 +0000815 (version, code, message) = response._read_status()
Senthil Kumaran47fff872009-12-20 07:10:31 +0000816
Senthil Kumaran97f0c6b2009-07-25 04:24:38 +0000817 if code != 200:
818 self.close()
Andrew Svetlov0832af62012-12-18 23:10:48 +0200819 raise OSError("Tunnel connection failed: %d %s" % (code,
820 message.strip()))
Senthil Kumaran97f0c6b2009-07-25 04:24:38 +0000821 while True:
Senthil Kumaran5466bf12010-12-18 16:55:23 +0000822 line = response.fp.readline(_MAXLINE + 1)
823 if len(line) > _MAXLINE:
824 raise LineTooLong("header line")
Senthil Kumaranb12771a2012-04-23 23:50:07 +0800825 if not line:
826 # for sites which EOF without sending a trailer
827 break
Senthil Kumaran7e70a5c2012-04-29 10:39:49 +0800828 if line in (b'\r\n', b'\n', b''):
Senthil Kumaran97f0c6b2009-07-25 04:24:38 +0000829 break
830
Greg Steindd6eefb2000-07-18 09:09:48 +0000831 def connect(self):
832 """Connect to the host and port specified in __init__."""
Senthil Kumaran9da047b2014-04-14 13:07:56 -0400833 self.sock = self._create_connection((self.host,self.port),
834 self.timeout, self.source_address)
835
Senthil Kumaran97f0c6b2009-07-25 04:24:38 +0000836 if self._tunnel_host:
837 self._tunnel()
Greg Stein5e0fa402000-06-26 08:28:01 +0000838
Greg Steindd6eefb2000-07-18 09:09:48 +0000839 def close(self):
840 """Close the connection to the HTTP server."""
841 if self.sock:
Tim Peters07e99cb2001-01-14 23:47:14 +0000842 self.sock.close() # close it manually... there may be other refs
Greg Steindd6eefb2000-07-18 09:09:48 +0000843 self.sock = None
844 if self.__response:
845 self.__response.close()
846 self.__response = None
847 self.__state = _CS_IDLE
Greg Stein5e0fa402000-06-26 08:28:01 +0000848
Senthil Kumaran5a2da3b2010-10-02 10:33:13 +0000849 def send(self, data):
Senthil Kumaran7bc0d872010-12-19 10:49:52 +0000850 """Send `data' to the server.
851 ``data`` can be a string object, a bytes object, an array object, a
852 file-like object that supports a .read() method, or an iterable object.
853 """
854
Greg Steindd6eefb2000-07-18 09:09:48 +0000855 if self.sock is None:
856 if self.auto_open:
857 self.connect()
858 else:
859 raise NotConnected()
Greg Stein5e0fa402000-06-26 08:28:01 +0000860
Jeremy Hylton30f86742000-09-18 22:50:38 +0000861 if self.debuglevel > 0:
Senthil Kumaran5a2da3b2010-10-02 10:33:13 +0000862 print("send:", repr(data))
Jeremy Hylton636950f2009-03-28 04:34:21 +0000863 blocksize = 8192
Senthil Kumaran5a2da3b2010-10-02 10:33:13 +0000864 if hasattr(data, "read") :
Jeremy Hylton636950f2009-03-28 04:34:21 +0000865 if self.debuglevel > 0:
866 print("sendIng a read()able")
867 encode = False
Antoine Pitrouead1d622009-09-29 18:44:53 +0000868 try:
Senthil Kumaran5a2da3b2010-10-02 10:33:13 +0000869 mode = data.mode
Antoine Pitrouead1d622009-09-29 18:44:53 +0000870 except AttributeError:
871 # io.BytesIO and other file-like objects don't have a `mode`
872 # attribute.
873 pass
874 else:
875 if "b" not in mode:
876 encode = True
877 if self.debuglevel > 0:
878 print("encoding file using iso-8859-1")
Jeremy Hylton636950f2009-03-28 04:34:21 +0000879 while 1:
Senthil Kumaran5a2da3b2010-10-02 10:33:13 +0000880 datablock = data.read(blocksize)
881 if not datablock:
Jeremy Hylton636950f2009-03-28 04:34:21 +0000882 break
883 if encode:
Senthil Kumaran5a2da3b2010-10-02 10:33:13 +0000884 datablock = datablock.encode("iso-8859-1")
885 self.sock.sendall(datablock)
Andrew Svetlov7b2c8bb2013-04-12 22:49:19 +0300886 return
Senthil Kumaran7bc0d872010-12-19 10:49:52 +0000887 try:
Senthil Kumaran5a2da3b2010-10-02 10:33:13 +0000888 self.sock.sendall(data)
Senthil Kumaran7bc0d872010-12-19 10:49:52 +0000889 except TypeError:
890 if isinstance(data, collections.Iterable):
891 for d in data:
892 self.sock.sendall(d)
893 else:
Georg Brandl25e2cd12011-08-03 08:27:00 +0200894 raise TypeError("data should be a bytes-like object "
895 "or an iterable, got %r" % type(data))
Greg Stein5e0fa402000-06-26 08:28:01 +0000896
Jeremy Hylton8531b1b2002-07-16 21:21:11 +0000897 def _output(self, s):
898 """Add a line of output to the current request buffer.
Tim Peters469cdad2002-08-08 20:19:19 +0000899
Jeremy Hyltone3252ec2002-07-16 21:41:43 +0000900 Assumes that the line does *not* end with \\r\\n.
Jeremy Hylton8531b1b2002-07-16 21:21:11 +0000901 """
902 self._buffer.append(s)
903
Benjamin Peterson1742e402008-11-30 22:15:29 +0000904 def _send_output(self, message_body=None):
Jeremy Hylton8531b1b2002-07-16 21:21:11 +0000905 """Send the currently buffered request and clear the buffer.
906
Jeremy Hyltone3252ec2002-07-16 21:41:43 +0000907 Appends an extra \\r\\n to the buffer.
Benjamin Peterson822b21c2009-01-18 00:04:57 +0000908 A message_body may be specified, to be appended to the request.
Jeremy Hylton8531b1b2002-07-16 21:21:11 +0000909 """
Martin v. Löwisdd5a8602007-06-30 09:22:09 +0000910 self._buffer.extend((b"", b""))
911 msg = b"\r\n".join(self._buffer)
Jeremy Hylton8531b1b2002-07-16 21:21:11 +0000912 del self._buffer[:]
Benjamin Peterson1742e402008-11-30 22:15:29 +0000913 # If msg and message_body are sent in a single send() call,
914 # it will avoid performance problems caused by the interaction
Antoine Pitrou90e47742013-01-02 22:10:47 +0100915 # between delayed ack and the Nagle algorithm. However,
916 # there is no performance gain if the message is larger
917 # than MSS (and there is a memory penalty for the message
918 # copy).
919 if isinstance(message_body, bytes) and len(message_body) < self.mss:
Benjamin Peterson1742e402008-11-30 22:15:29 +0000920 msg += message_body
Benjamin Peterson822b21c2009-01-18 00:04:57 +0000921 message_body = None
Jeremy Hylton8531b1b2002-07-16 21:21:11 +0000922 self.send(msg)
Benjamin Peterson822b21c2009-01-18 00:04:57 +0000923 if message_body is not None:
Jeremy Hylton236654b2009-03-27 20:24:34 +0000924 # message_body was not a string (i.e. it is a file), and
925 # we must run the risk of Nagle.
Benjamin Peterson822b21c2009-01-18 00:04:57 +0000926 self.send(message_body)
Jeremy Hylton8531b1b2002-07-16 21:21:11 +0000927
Martin v. Löwisaf7dc8d2003-11-19 19:51:55 +0000928 def putrequest(self, method, url, skip_host=0, skip_accept_encoding=0):
Greg Steindd6eefb2000-07-18 09:09:48 +0000929 """Send a request to the server.
Greg Stein5e0fa402000-06-26 08:28:01 +0000930
Greg Steindd6eefb2000-07-18 09:09:48 +0000931 `method' specifies an HTTP request method, e.g. 'GET'.
932 `url' specifies the object being requested, e.g. '/index.html'.
Martin v. Löwisaf7dc8d2003-11-19 19:51:55 +0000933 `skip_host' if True does not add automatically a 'Host:' header
934 `skip_accept_encoding' if True does not add automatically an
935 'Accept-Encoding:' header
Greg Steindd6eefb2000-07-18 09:09:48 +0000936 """
Greg Stein5e0fa402000-06-26 08:28:01 +0000937
Greg Stein616a58d2003-06-24 06:35:19 +0000938 # if a prior response has been completed, then forget about it.
Greg Steindd6eefb2000-07-18 09:09:48 +0000939 if self.__response and self.__response.isclosed():
940 self.__response = None
Greg Stein5e0fa402000-06-26 08:28:01 +0000941
Tim Peters58eb11c2004-01-18 20:29:55 +0000942
Greg Steindd6eefb2000-07-18 09:09:48 +0000943 # in certain cases, we cannot issue another request on this connection.
944 # this occurs when:
945 # 1) we are in the process of sending a request. (_CS_REQ_STARTED)
946 # 2) a response to a previous request has signalled that it is going
947 # to close the connection upon completion.
948 # 3) the headers for the previous response have not been read, thus
949 # we cannot determine whether point (2) is true. (_CS_REQ_SENT)
950 #
951 # if there is no prior response, then we can request at will.
952 #
953 # if point (2) is true, then we will have passed the socket to the
954 # response (effectively meaning, "there is no prior response"), and
955 # will open a new one when a new request is made.
956 #
957 # Note: if a prior response exists, then we *can* start a new request.
958 # We are not allowed to begin fetching the response to this new
959 # request, however, until that prior response is complete.
960 #
961 if self.__state == _CS_IDLE:
962 self.__state = _CS_REQ_STARTED
963 else:
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +0000964 raise CannotSendRequest(self.__state)
Greg Stein5e0fa402000-06-26 08:28:01 +0000965
Jeremy Hyltonc1b2cb92003-05-05 16:13:58 +0000966 # Save the method we use, we need it later in the response phase
967 self._method = method
Greg Steindd6eefb2000-07-18 09:09:48 +0000968 if not url:
969 url = '/'
Martin v. Löwisdd5a8602007-06-30 09:22:09 +0000970 request = '%s %s %s' % (method, url, self._http_vsn_str)
Greg Stein5e0fa402000-06-26 08:28:01 +0000971
Martin v. Löwisdd5a8602007-06-30 09:22:09 +0000972 # Non-ASCII characters should have been eliminated earlier
973 self._output(request.encode('ascii'))
Greg Stein5e0fa402000-06-26 08:28:01 +0000974
Greg Steindd6eefb2000-07-18 09:09:48 +0000975 if self._http_vsn == 11:
976 # Issue some standard headers for better HTTP/1.1 compliance
Greg Stein5e0fa402000-06-26 08:28:01 +0000977
Jeremy Hylton3921ff62002-03-09 06:07:23 +0000978 if not skip_host:
979 # this header is issued *only* for HTTP/1.1
980 # connections. more specifically, this means it is
981 # only issued when the client uses the new
982 # HTTPConnection() class. backwards-compat clients
983 # will be using HTTP/1.0 and those clients may be
984 # issuing this header themselves. we should NOT issue
985 # it twice; some web servers (such as Apache) barf
986 # when they see two Host: headers
Guido van Rossumf6922aa2001-01-14 21:03:01 +0000987
Jeremy Hylton3921ff62002-03-09 06:07:23 +0000988 # If we need a non-standard port,include it in the
989 # header. If the request is going through a proxy,
990 # but the host of the actual URL, not the host of the
991 # proxy.
Jeremy Hylton8acf1e02002-03-08 19:35:51 +0000992
Jeremy Hylton3921ff62002-03-09 06:07:23 +0000993 netloc = ''
994 if url.startswith('http'):
995 nil, netloc, nil, nil, nil = urlsplit(url)
996
997 if netloc:
Thomas Wouters477c8d52006-05-27 19:21:47 +0000998 try:
999 netloc_enc = netloc.encode("ascii")
1000 except UnicodeEncodeError:
1001 netloc_enc = netloc.encode("idna")
1002 self.putheader('Host', netloc_enc)
Jeremy Hylton3921ff62002-03-09 06:07:23 +00001003 else:
Senthil Kumaran9da047b2014-04-14 13:07:56 -04001004 if self._tunnel_host:
1005 host = self._tunnel_host
1006 port = self._tunnel_port
1007 else:
1008 host = self.host
1009 port = self.port
1010
Thomas Wouters477c8d52006-05-27 19:21:47 +00001011 try:
Senthil Kumaran9da047b2014-04-14 13:07:56 -04001012 host_enc = host.encode("ascii")
Thomas Wouters477c8d52006-05-27 19:21:47 +00001013 except UnicodeEncodeError:
Senthil Kumaran9da047b2014-04-14 13:07:56 -04001014 host_enc = host.encode("idna")
Senthil Kumaran74ebd9e2010-11-13 12:27:49 +00001015
1016 # As per RFC 273, IPv6 address should be wrapped with []
1017 # when used as Host header
1018
Senthil Kumaran9da047b2014-04-14 13:07:56 -04001019 if host.find(':') >= 0:
Senthil Kumaran74ebd9e2010-11-13 12:27:49 +00001020 host_enc = b'[' + host_enc + b']'
1021
Senthil Kumaran9da047b2014-04-14 13:07:56 -04001022 if port == self.default_port:
Thomas Wouters477c8d52006-05-27 19:21:47 +00001023 self.putheader('Host', host_enc)
1024 else:
Guido van Rossum98297ee2007-11-06 21:34:58 +00001025 host_enc = host_enc.decode("ascii")
Senthil Kumaran9da047b2014-04-14 13:07:56 -04001026 self.putheader('Host', "%s:%s" % (host_enc, port))
Greg Stein5e0fa402000-06-26 08:28:01 +00001027
Greg Steindd6eefb2000-07-18 09:09:48 +00001028 # note: we are assuming that clients will not attempt to set these
1029 # headers since *this* library must deal with the
1030 # consequences. this also means that when the supporting
1031 # libraries are updated to recognize other forms, then this
1032 # code should be changed (removed or updated).
Greg Stein5e0fa402000-06-26 08:28:01 +00001033
Greg Steindd6eefb2000-07-18 09:09:48 +00001034 # we only want a Content-Encoding of "identity" since we don't
1035 # support encodings such as x-gzip or x-deflate.
Martin v. Löwisaf7dc8d2003-11-19 19:51:55 +00001036 if not skip_accept_encoding:
1037 self.putheader('Accept-Encoding', 'identity')
Greg Stein5e0fa402000-06-26 08:28:01 +00001038
Greg Steindd6eefb2000-07-18 09:09:48 +00001039 # we can accept "chunked" Transfer-Encodings, but no others
1040 # NOTE: no TE header implies *only* "chunked"
1041 #self.putheader('TE', 'chunked')
Greg Stein5e0fa402000-06-26 08:28:01 +00001042
Greg Steindd6eefb2000-07-18 09:09:48 +00001043 # if TE is supplied in the header, then it must appear in a
1044 # Connection header.
1045 #self.putheader('Connection', 'TE')
Greg Stein5e0fa402000-06-26 08:28:01 +00001046
Greg Steindd6eefb2000-07-18 09:09:48 +00001047 else:
1048 # For HTTP/1.0, the server will assume "not chunked"
1049 pass
Greg Stein5e0fa402000-06-26 08:28:01 +00001050
Benjamin Petersonf608c612008-11-16 18:33:53 +00001051 def putheader(self, header, *values):
Greg Steindd6eefb2000-07-18 09:09:48 +00001052 """Send a request header line to the server.
Greg Stein5e0fa402000-06-26 08:28:01 +00001053
Greg Steindd6eefb2000-07-18 09:09:48 +00001054 For example: h.putheader('Accept', 'text/html')
1055 """
1056 if self.__state != _CS_REQ_STARTED:
1057 raise CannotSendHeader()
Greg Stein5e0fa402000-06-26 08:28:01 +00001058
Guido van Rossum98297ee2007-11-06 21:34:58 +00001059 if hasattr(header, 'encode'):
1060 header = header.encode('ascii')
Benjamin Petersonf608c612008-11-16 18:33:53 +00001061 values = list(values)
1062 for i, one_value in enumerate(values):
1063 if hasattr(one_value, 'encode'):
Marc-André Lemburg8f36af72011-02-25 15:42:01 +00001064 values[i] = one_value.encode('latin-1')
Senthil Kumaran58d5dbf2010-10-03 18:22:42 +00001065 elif isinstance(one_value, int):
1066 values[i] = str(one_value).encode('ascii')
Benjamin Petersonf608c612008-11-16 18:33:53 +00001067 value = b'\r\n\t'.join(values)
Guido van Rossum98297ee2007-11-06 21:34:58 +00001068 header = header + b': ' + value
1069 self._output(header)
Greg Stein5e0fa402000-06-26 08:28:01 +00001070
Benjamin Peterson1742e402008-11-30 22:15:29 +00001071 def endheaders(self, message_body=None):
1072 """Indicate that the last header line has been sent to the server.
Greg Stein5e0fa402000-06-26 08:28:01 +00001073
Senthil Kumaranad87fa62011-10-05 23:26:49 +08001074 This method sends the request to the server. The optional message_body
1075 argument can be used to pass a message body associated with the
1076 request. The message body will be sent in the same packet as the
1077 message headers if it is a string, otherwise it is sent as a separate
1078 packet.
Benjamin Peterson1742e402008-11-30 22:15:29 +00001079 """
Greg Steindd6eefb2000-07-18 09:09:48 +00001080 if self.__state == _CS_REQ_STARTED:
1081 self.__state = _CS_REQ_SENT
1082 else:
1083 raise CannotSendHeader()
Benjamin Peterson1742e402008-11-30 22:15:29 +00001084 self._send_output(message_body)
Greg Stein5e0fa402000-06-26 08:28:01 +00001085
Greg Steindd6eefb2000-07-18 09:09:48 +00001086 def request(self, method, url, body=None, headers={}):
1087 """Send a complete request to the server."""
Jeremy Hylton636950f2009-03-28 04:34:21 +00001088 self._send_request(method, url, body, headers)
Greg Stein5e0fa402000-06-26 08:28:01 +00001089
Benjamin Peterson1742e402008-11-30 22:15:29 +00001090 def _set_content_length(self, body):
1091 # Set the content-length based on the body.
1092 thelen = None
1093 try:
1094 thelen = str(len(body))
1095 except TypeError as te:
1096 # If this is a file-like object, try to
1097 # fstat its file descriptor
Benjamin Peterson1742e402008-11-30 22:15:29 +00001098 try:
1099 thelen = str(os.fstat(body.fileno()).st_size)
1100 except (AttributeError, OSError):
1101 # Don't send a length if this failed
1102 if self.debuglevel > 0: print("Cannot stat!!")
1103
1104 if thelen is not None:
1105 self.putheader('Content-Length', thelen)
1106
Greg Steindd6eefb2000-07-18 09:09:48 +00001107 def _send_request(self, method, url, body, headers):
Jeremy Hylton636950f2009-03-28 04:34:21 +00001108 # Honor explicitly requested Host: and Accept-Encoding: headers.
Jeremy Hylton2c178252004-08-07 16:28:14 +00001109 header_names = dict.fromkeys([k.lower() for k in headers])
1110 skips = {}
1111 if 'host' in header_names:
1112 skips['skip_host'] = 1
1113 if 'accept-encoding' in header_names:
1114 skips['skip_accept_encoding'] = 1
Greg Stein5e0fa402000-06-26 08:28:01 +00001115
Jeremy Hylton2c178252004-08-07 16:28:14 +00001116 self.putrequest(method, url, **skips)
1117
Senthil Kumaran5fa4a892012-05-19 16:58:09 +08001118 if body is not None and ('content-length' not in header_names):
Benjamin Peterson1742e402008-11-30 22:15:29 +00001119 self._set_content_length(body)
Guido van Rossumcc2b0162007-02-11 06:12:03 +00001120 for hdr, value in headers.items():
Greg Steindd6eefb2000-07-18 09:09:48 +00001121 self.putheader(hdr, value)
Jeremy Hyltonef9f48e2009-03-26 22:04:05 +00001122 if isinstance(body, str):
Jeremy Hylton236654b2009-03-27 20:24:34 +00001123 # RFC 2616 Section 3.7.1 says that text default has a
1124 # default charset of iso-8859-1.
1125 body = body.encode('iso-8859-1')
Jeremy Hyltonef9f48e2009-03-26 22:04:05 +00001126 self.endheaders(body)
Greg Stein5e0fa402000-06-26 08:28:01 +00001127
Greg Steindd6eefb2000-07-18 09:09:48 +00001128 def getresponse(self):
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +00001129 """Get the response from the server.
1130
1131 If the HTTPConnection is in the correct state, returns an
1132 instance of HTTPResponse or of whatever object is returned by
1133 class the response_class variable.
1134
1135 If a request has not been sent or if a previous response has
1136 not be handled, ResponseNotReady is raised. If the HTTP
1137 response indicates that the connection should be closed, then
1138 it will be closed before the response is returned. When the
1139 connection is closed, the underlying socket is closed.
1140 """
Greg Stein5e0fa402000-06-26 08:28:01 +00001141
Greg Stein616a58d2003-06-24 06:35:19 +00001142 # if a prior response has been completed, then forget about it.
Greg Steindd6eefb2000-07-18 09:09:48 +00001143 if self.__response and self.__response.isclosed():
1144 self.__response = None
Greg Stein5e0fa402000-06-26 08:28:01 +00001145
Greg Steindd6eefb2000-07-18 09:09:48 +00001146 # if a prior response exists, then it must be completed (otherwise, we
1147 # cannot read this response's header to determine the connection-close
1148 # behavior)
1149 #
1150 # note: if a prior response existed, but was connection-close, then the
1151 # socket and response were made independent of this HTTPConnection
1152 # object since a new request requires that we open a whole new
1153 # connection
1154 #
1155 # this means the prior response had one of two states:
1156 # 1) will_close: this connection was reset and the prior socket and
1157 # response operate independently
1158 # 2) persistent: the response was retained and we await its
1159 # isclosed() status to become true.
1160 #
1161 if self.__state != _CS_REQ_SENT or self.__response:
Jeremy Hylton6c5e28c2009-03-31 14:35:53 +00001162 raise ResponseNotReady(self.__state)
Greg Stein5e0fa402000-06-26 08:28:01 +00001163
Jeremy Hylton30f86742000-09-18 22:50:38 +00001164 if self.debuglevel > 0:
Jeremy Hyltond46aa372002-07-06 18:48:07 +00001165 response = self.response_class(self.sock, self.debuglevel,
Jeremy Hyltonc1b2cb92003-05-05 16:13:58 +00001166 method=self._method)
Jeremy Hylton30f86742000-09-18 22:50:38 +00001167 else:
Antoine Pitrou988dbd72010-12-17 17:35:56 +00001168 response = self.response_class(self.sock, method=self._method)
Greg Stein5e0fa402000-06-26 08:28:01 +00001169
Serhiy Storchakab491e052014-12-01 13:07:45 +02001170 try:
1171 response.begin()
1172 assert response.will_close != _UNKNOWN
1173 self.__state = _CS_IDLE
Greg Stein5e0fa402000-06-26 08:28:01 +00001174
Serhiy Storchakab491e052014-12-01 13:07:45 +02001175 if response.will_close:
1176 # this effectively passes the connection to the response
1177 self.close()
1178 else:
1179 # remember this, so we can tell when it is complete
1180 self.__response = response
Greg Stein5e0fa402000-06-26 08:28:01 +00001181
Serhiy Storchakab491e052014-12-01 13:07:45 +02001182 return response
1183 except:
1184 response.close()
1185 raise
Greg Stein5e0fa402000-06-26 08:28:01 +00001186
Thomas Wouters47b49bf2007-08-30 22:15:33 +00001187try:
1188 import ssl
Brett Cannoncd171c82013-07-04 17:43:24 -04001189except ImportError:
Thomas Wouters47b49bf2007-08-30 22:15:33 +00001190 pass
1191else:
1192 class HTTPSConnection(HTTPConnection):
1193 "This class allows communication via SSL."
Jeremy Hylton29d27ac2002-07-09 21:22:36 +00001194
Thomas Wouters47b49bf2007-08-30 22:15:33 +00001195 default_port = HTTPS_PORT
Jeremy Hylton29d27ac2002-07-09 21:22:36 +00001196
Antoine Pitrou803e6d62010-10-13 10:36:15 +00001197 # XXX Should key_file and cert_file be deprecated in favour of context?
1198
Thomas Wouters47b49bf2007-08-30 22:15:33 +00001199 def __init__(self, host, port=None, key_file=None, cert_file=None,
Senthil Kumaran052ddb02013-03-18 14:11:41 -07001200 timeout=socket._GLOBAL_DEFAULT_TIMEOUT,
1201 source_address=None, *, context=None,
1202 check_hostname=None):
1203 super(HTTPSConnection, self).__init__(host, port, timeout,
Gregory P. Smithb4066372010-01-03 03:28:29 +00001204 source_address)
Thomas Wouters47b49bf2007-08-30 22:15:33 +00001205 self.key_file = key_file
1206 self.cert_file = cert_file
Antoine Pitrou803e6d62010-10-13 10:36:15 +00001207 if context is None:
Benjamin Peterson4ffb0752014-11-03 14:29:33 -05001208 context = ssl._create_default_https_context()
Antoine Pitrou803e6d62010-10-13 10:36:15 +00001209 will_verify = context.verify_mode != ssl.CERT_NONE
1210 if check_hostname is None:
Benjamin Petersona090f012014-12-07 13:18:25 -05001211 check_hostname = context.check_hostname
1212 if check_hostname and not will_verify:
Antoine Pitrou803e6d62010-10-13 10:36:15 +00001213 raise ValueError("check_hostname needs a SSL context with "
1214 "either CERT_OPTIONAL or CERT_REQUIRED")
1215 if key_file or cert_file:
Antoine Pitrou80603992010-10-13 10:45:16 +00001216 context.load_cert_chain(cert_file, key_file)
Antoine Pitrou803e6d62010-10-13 10:36:15 +00001217 self._context = context
1218 self._check_hostname = check_hostname
Jeremy Hylton29d27ac2002-07-09 21:22:36 +00001219
Thomas Wouters47b49bf2007-08-30 22:15:33 +00001220 def connect(self):
1221 "Connect to a host on a given (SSL) port."
Jeremy Hylton29d27ac2002-07-09 21:22:36 +00001222
Senthil Kumaran9da047b2014-04-14 13:07:56 -04001223 super().connect()
Senthil Kumaran97f0c6b2009-07-25 04:24:38 +00001224
1225 if self._tunnel_host:
Senthil Kumaran9da047b2014-04-14 13:07:56 -04001226 server_hostname = self._tunnel_host
1227 else:
1228 server_hostname = self.host
Senthil Kumaran97f0c6b2009-07-25 04:24:38 +00001229
Senthil Kumaran9da047b2014-04-14 13:07:56 -04001230 self.sock = self._context.wrap_socket(self.sock,
Benjamin Peterson7243b572014-11-23 17:04:34 -06001231 server_hostname=server_hostname)
Christian Heimesf723c712013-12-05 07:51:17 +01001232 if not self._context.check_hostname and self._check_hostname:
1233 try:
Senthil Kumaran9da047b2014-04-14 13:07:56 -04001234 ssl.match_hostname(self.sock.getpeercert(), server_hostname)
Christian Heimesf723c712013-12-05 07:51:17 +01001235 except Exception:
1236 self.sock.shutdown(socket.SHUT_RDWR)
1237 self.sock.close()
1238 raise
Jeremy Hylton29d27ac2002-07-09 21:22:36 +00001239
Thomas Wouters47b49bf2007-08-30 22:15:33 +00001240 __all__.append("HTTPSConnection")
Greg Stein5e0fa402000-06-26 08:28:01 +00001241
Greg Stein5e0fa402000-06-26 08:28:01 +00001242class HTTPException(Exception):
Jeremy Hylton12f4f352002-07-06 18:55:01 +00001243 # Subclasses that define an __init__ must call Exception.__init__
1244 # or define self.args. Otherwise, str() will fail.
Greg Steindd6eefb2000-07-18 09:09:48 +00001245 pass
Greg Stein5e0fa402000-06-26 08:28:01 +00001246
1247class NotConnected(HTTPException):
Greg Steindd6eefb2000-07-18 09:09:48 +00001248 pass
Greg Stein5e0fa402000-06-26 08:28:01 +00001249
Skip Montanaro9d389972002-03-24 16:53:50 +00001250class InvalidURL(HTTPException):
1251 pass
1252
Greg Stein5e0fa402000-06-26 08:28:01 +00001253class UnknownProtocol(HTTPException):
Greg Steindd6eefb2000-07-18 09:09:48 +00001254 def __init__(self, version):
Jeremy Hylton12f4f352002-07-06 18:55:01 +00001255 self.args = version,
Greg Steindd6eefb2000-07-18 09:09:48 +00001256 self.version = version
Greg Stein5e0fa402000-06-26 08:28:01 +00001257
1258class UnknownTransferEncoding(HTTPException):
Greg Steindd6eefb2000-07-18 09:09:48 +00001259 pass
Greg Stein5e0fa402000-06-26 08:28:01 +00001260
Greg Stein5e0fa402000-06-26 08:28:01 +00001261class UnimplementedFileMode(HTTPException):
Greg Steindd6eefb2000-07-18 09:09:48 +00001262 pass
Greg Stein5e0fa402000-06-26 08:28:01 +00001263
1264class IncompleteRead(HTTPException):
Benjamin Peterson6accb982009-03-02 22:50:25 +00001265 def __init__(self, partial, expected=None):
Jeremy Hylton12f4f352002-07-06 18:55:01 +00001266 self.args = partial,
Greg Steindd6eefb2000-07-18 09:09:48 +00001267 self.partial = partial
Benjamin Peterson6accb982009-03-02 22:50:25 +00001268 self.expected = expected
1269 def __repr__(self):
1270 if self.expected is not None:
1271 e = ', %i more expected' % self.expected
1272 else:
1273 e = ''
1274 return 'IncompleteRead(%i bytes read%s)' % (len(self.partial), e)
1275 def __str__(self):
1276 return repr(self)
Greg Stein5e0fa402000-06-26 08:28:01 +00001277
1278class ImproperConnectionState(HTTPException):
Greg Steindd6eefb2000-07-18 09:09:48 +00001279 pass
Greg Stein5e0fa402000-06-26 08:28:01 +00001280
1281class CannotSendRequest(ImproperConnectionState):
Greg Steindd6eefb2000-07-18 09:09:48 +00001282 pass
Greg Stein5e0fa402000-06-26 08:28:01 +00001283
1284class CannotSendHeader(ImproperConnectionState):
Greg Steindd6eefb2000-07-18 09:09:48 +00001285 pass
Greg Stein5e0fa402000-06-26 08:28:01 +00001286
1287class ResponseNotReady(ImproperConnectionState):
Greg Steindd6eefb2000-07-18 09:09:48 +00001288 pass
Greg Stein5e0fa402000-06-26 08:28:01 +00001289
1290class BadStatusLine(HTTPException):
Greg Steindd6eefb2000-07-18 09:09:48 +00001291 def __init__(self, line):
Benjamin Peterson11dbfd42010-03-21 22:50:04 +00001292 if not line:
1293 line = repr(line)
Jeremy Hylton12f4f352002-07-06 18:55:01 +00001294 self.args = line,
Greg Steindd6eefb2000-07-18 09:09:48 +00001295 self.line = line
Greg Stein5e0fa402000-06-26 08:28:01 +00001296
Senthil Kumaran5466bf12010-12-18 16:55:23 +00001297class LineTooLong(HTTPException):
1298 def __init__(self, line_type):
1299 HTTPException.__init__(self, "got more than %d bytes when reading %s"
1300 % (_MAXLINE, line_type))
1301
Greg Stein5e0fa402000-06-26 08:28:01 +00001302# for backwards compatibility
1303error = HTTPException