Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1 | """HTTP/1.1 client library |
Guido van Rossum | 41999c1 | 1997-12-09 00:12:23 +0000 | [diff] [blame] | 2 | |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 3 | <intro stuff goes here> |
| 4 | <other stuff, too> |
Guido van Rossum | 41999c1 | 1997-12-09 00:12:23 +0000 | [diff] [blame] | 5 | |
Thomas Wouters | 0e3f591 | 2006-08-11 14:57:12 +0000 | [diff] [blame] | 6 | HTTPConnection goes through a number of "states", which define when a client |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 7 | may legally make another request or fetch the response for a particular |
| 8 | request. This diagram details these state transitions: |
Guido van Rossum | 41999c1 | 1997-12-09 00:12:23 +0000 | [diff] [blame] | 9 | |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 10 | (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 Peters | 5ceadc8 | 2001-01-13 19:16:21 +0000 | [diff] [blame] | 28 | | | |
| 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 Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 41 | |
| 42 | This 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 | |
| 48 | Note: 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 | |
| 59 | Logical State __state __response |
| 60 | ------------- ------- ---------- |
| 61 | Idle _CS_IDLE None |
| 62 | Request-started _CS_REQ_STARTED None |
| 63 | Request-sent _CS_REQ_SENT None |
| 64 | Unread-response _CS_IDLE <response_class> |
| 65 | Req-started-unread-response _CS_REQ_STARTED <response_class> |
| 66 | Req-sent-unread-response _CS_REQ_SENT <response_class> |
Guido van Rossum | 41999c1 | 1997-12-09 00:12:23 +0000 | [diff] [blame] | 67 | """ |
Guido van Rossum | 23acc95 | 1994-02-21 16:36:04 +0000 | [diff] [blame] | 68 | |
Barry Warsaw | 820c120 | 2008-06-12 04:06:45 +0000 | [diff] [blame] | 69 | import email.parser |
| 70 | import email.message |
Jeremy Hylton | 636950f | 2009-03-28 04:34:21 +0000 | [diff] [blame] | 71 | import io |
| 72 | import os |
| 73 | import socket |
Jeremy Hylton | 1afc169 | 2008-06-18 20:49:58 +0000 | [diff] [blame] | 74 | from urllib.parse import urlsplit |
Thomas Wouters | 89d996e | 2007-09-08 17:39:28 +0000 | [diff] [blame] | 75 | import warnings |
Guido van Rossum | 23acc95 | 1994-02-21 16:36:04 +0000 | [diff] [blame] | 76 | |
Thomas Wouters | 47b49bf | 2007-08-30 22:15:33 +0000 | [diff] [blame] | 77 | __all__ = ["HTTPResponse", "HTTPConnection", |
Skip Montanaro | 951a884 | 2001-06-01 16:25:38 +0000 | [diff] [blame] | 78 | "HTTPException", "NotConnected", "UnknownProtocol", |
Jeremy Hylton | 7c75c99 | 2002-06-28 23:38:14 +0000 | [diff] [blame] | 79 | "UnknownTransferEncoding", "UnimplementedFileMode", |
| 80 | "IncompleteRead", "InvalidURL", "ImproperConnectionState", |
| 81 | "CannotSendRequest", "CannotSendHeader", "ResponseNotReady", |
Georg Brandl | 6aab16e | 2006-02-17 19:17:25 +0000 | [diff] [blame] | 82 | "BadStatusLine", "error", "responses"] |
Skip Montanaro | 2dd4276 | 2001-01-23 15:35:05 +0000 | [diff] [blame] | 83 | |
Guido van Rossum | 23acc95 | 1994-02-21 16:36:04 +0000 | [diff] [blame] | 84 | HTTP_PORT = 80 |
Guido van Rossum | 09c8b6c | 1999-12-07 21:37:17 +0000 | [diff] [blame] | 85 | HTTPS_PORT = 443 |
| 86 | |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 87 | _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. Löwis | 39a3178 | 2004-09-18 09:03:49 +0000 | [diff] [blame] | 94 | # status codes |
| 95 | # informational |
| 96 | CONTINUE = 100 |
| 97 | SWITCHING_PROTOCOLS = 101 |
| 98 | PROCESSING = 102 |
| 99 | |
| 100 | # successful |
| 101 | OK = 200 |
| 102 | CREATED = 201 |
| 103 | ACCEPTED = 202 |
| 104 | NON_AUTHORITATIVE_INFORMATION = 203 |
| 105 | NO_CONTENT = 204 |
| 106 | RESET_CONTENT = 205 |
| 107 | PARTIAL_CONTENT = 206 |
| 108 | MULTI_STATUS = 207 |
| 109 | IM_USED = 226 |
| 110 | |
| 111 | # redirection |
| 112 | MULTIPLE_CHOICES = 300 |
| 113 | MOVED_PERMANENTLY = 301 |
| 114 | FOUND = 302 |
| 115 | SEE_OTHER = 303 |
| 116 | NOT_MODIFIED = 304 |
| 117 | USE_PROXY = 305 |
| 118 | TEMPORARY_REDIRECT = 307 |
| 119 | |
| 120 | # client error |
| 121 | BAD_REQUEST = 400 |
| 122 | UNAUTHORIZED = 401 |
| 123 | PAYMENT_REQUIRED = 402 |
| 124 | FORBIDDEN = 403 |
| 125 | NOT_FOUND = 404 |
| 126 | METHOD_NOT_ALLOWED = 405 |
| 127 | NOT_ACCEPTABLE = 406 |
| 128 | PROXY_AUTHENTICATION_REQUIRED = 407 |
| 129 | REQUEST_TIMEOUT = 408 |
| 130 | CONFLICT = 409 |
| 131 | GONE = 410 |
| 132 | LENGTH_REQUIRED = 411 |
| 133 | PRECONDITION_FAILED = 412 |
| 134 | REQUEST_ENTITY_TOO_LARGE = 413 |
| 135 | REQUEST_URI_TOO_LONG = 414 |
| 136 | UNSUPPORTED_MEDIA_TYPE = 415 |
| 137 | REQUESTED_RANGE_NOT_SATISFIABLE = 416 |
| 138 | EXPECTATION_FAILED = 417 |
| 139 | UNPROCESSABLE_ENTITY = 422 |
| 140 | LOCKED = 423 |
| 141 | FAILED_DEPENDENCY = 424 |
| 142 | UPGRADE_REQUIRED = 426 |
| 143 | |
| 144 | # server error |
| 145 | INTERNAL_SERVER_ERROR = 500 |
| 146 | NOT_IMPLEMENTED = 501 |
| 147 | BAD_GATEWAY = 502 |
| 148 | SERVICE_UNAVAILABLE = 503 |
| 149 | GATEWAY_TIMEOUT = 504 |
| 150 | HTTP_VERSION_NOT_SUPPORTED = 505 |
| 151 | INSUFFICIENT_STORAGE = 507 |
| 152 | NOT_EXTENDED = 510 |
| 153 | |
Georg Brandl | 6aab16e | 2006-02-17 19:17:25 +0000 | [diff] [blame] | 154 | # Mapping status codes to official W3C names |
| 155 | responses = { |
| 156 | 100: 'Continue', |
| 157 | 101: 'Switching Protocols', |
| 158 | |
| 159 | 200: 'OK', |
| 160 | 201: 'Created', |
| 161 | 202: 'Accepted', |
| 162 | 203: 'Non-Authoritative Information', |
| 163 | 204: 'No Content', |
| 164 | 205: 'Reset Content', |
| 165 | 206: 'Partial Content', |
| 166 | |
| 167 | 300: 'Multiple Choices', |
| 168 | 301: 'Moved Permanently', |
| 169 | 302: 'Found', |
| 170 | 303: 'See Other', |
| 171 | 304: 'Not Modified', |
| 172 | 305: 'Use Proxy', |
| 173 | 306: '(Unused)', |
| 174 | 307: 'Temporary Redirect', |
| 175 | |
| 176 | 400: 'Bad Request', |
| 177 | 401: 'Unauthorized', |
| 178 | 402: 'Payment Required', |
| 179 | 403: 'Forbidden', |
| 180 | 404: 'Not Found', |
| 181 | 405: 'Method Not Allowed', |
| 182 | 406: 'Not Acceptable', |
| 183 | 407: 'Proxy Authentication Required', |
| 184 | 408: 'Request Timeout', |
| 185 | 409: 'Conflict', |
| 186 | 410: 'Gone', |
| 187 | 411: 'Length Required', |
| 188 | 412: 'Precondition Failed', |
| 189 | 413: 'Request Entity Too Large', |
| 190 | 414: 'Request-URI Too Long', |
| 191 | 415: 'Unsupported Media Type', |
| 192 | 416: 'Requested Range Not Satisfiable', |
| 193 | 417: 'Expectation Failed', |
| 194 | |
| 195 | 500: 'Internal Server Error', |
| 196 | 501: 'Not Implemented', |
| 197 | 502: 'Bad Gateway', |
| 198 | 503: 'Service Unavailable', |
| 199 | 504: 'Gateway Timeout', |
| 200 | 505: 'HTTP Version Not Supported', |
| 201 | } |
| 202 | |
Georg Brandl | 80ba8e8 | 2005-09-29 20:16:07 +0000 | [diff] [blame] | 203 | # maximal amount of data to read at one time in _safe_read |
| 204 | MAXAMOUNT = 1048576 |
| 205 | |
Antoine Pitrou | ff1bbba | 2010-12-18 18:04:38 +0000 | [diff] [blame] | 206 | # maximal line length when calling readline(). |
| 207 | _MAXLINE = 65536 |
| 208 | |
Barry Warsaw | 820c120 | 2008-06-12 04:06:45 +0000 | [diff] [blame] | 209 | class HTTPMessage(email.message.Message): |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 210 | # XXX The only usage of this method is in |
| 211 | # http.server.CGIHTTPRequestHandler. Maybe move the code there so |
| 212 | # that it doesn't need to be part of the public API. The API has |
| 213 | # never been defined so this could cause backwards compatibility |
| 214 | # issues. |
| 215 | |
Barry Warsaw | 820c120 | 2008-06-12 04:06:45 +0000 | [diff] [blame] | 216 | def getallmatchingheaders(self, name): |
| 217 | """Find all header lines matching a given header name. |
Jeremy Hylton | 6d0a4c7 | 2002-07-07 16:51:37 +0000 | [diff] [blame] | 218 | |
Barry Warsaw | 820c120 | 2008-06-12 04:06:45 +0000 | [diff] [blame] | 219 | Look through the list of headers and find all lines matching a given |
| 220 | header name (and their continuation lines). A list of the lines is |
| 221 | returned, without interpretation. If the header does not occur, an |
| 222 | empty list is returned. If the header occurs multiple times, all |
| 223 | occurrences are returned. Case is not important in the header name. |
Jeremy Hylton | 6d0a4c7 | 2002-07-07 16:51:37 +0000 | [diff] [blame] | 224 | |
Jeremy Hylton | 6d0a4c7 | 2002-07-07 16:51:37 +0000 | [diff] [blame] | 225 | """ |
Barry Warsaw | 820c120 | 2008-06-12 04:06:45 +0000 | [diff] [blame] | 226 | name = name.lower() + ':' |
| 227 | n = len(name) |
| 228 | lst = [] |
| 229 | hit = 0 |
| 230 | for line in self.keys(): |
| 231 | if line[:n].lower() == name: |
| 232 | hit = 1 |
| 233 | elif not line[:1].isspace(): |
| 234 | hit = 0 |
| 235 | if hit: |
| 236 | lst.append(line) |
| 237 | return lst |
Jeremy Hylton | 6d0a4c7 | 2002-07-07 16:51:37 +0000 | [diff] [blame] | 238 | |
Jeremy Hylton | 98eb6c2 | 2009-03-27 18:31:36 +0000 | [diff] [blame] | 239 | def parse_headers(fp, _class=HTTPMessage): |
Barry Warsaw | 820c120 | 2008-06-12 04:06:45 +0000 | [diff] [blame] | 240 | """Parses only RFC2822 headers from a file pointer. |
| 241 | |
| 242 | email Parser wants to see strings rather than bytes. |
| 243 | But a TextIOWrapper around self.rfile would buffer too many bytes |
| 244 | from the stream, bytes which we later need to read as bytes. |
| 245 | So we read the correct bytes here, as bytes, for email Parser |
| 246 | to parse. |
| 247 | |
| 248 | """ |
Barry Warsaw | 820c120 | 2008-06-12 04:06:45 +0000 | [diff] [blame] | 249 | headers = [] |
| 250 | while True: |
Antoine Pitrou | ff1bbba | 2010-12-18 18:04:38 +0000 | [diff] [blame] | 251 | line = fp.readline(_MAXLINE + 1) |
| 252 | if len(line) > _MAXLINE: |
| 253 | raise LineTooLong("header line") |
Barry Warsaw | 820c120 | 2008-06-12 04:06:45 +0000 | [diff] [blame] | 254 | headers.append(line) |
| 255 | if line in (b'\r\n', b'\n', b''): |
| 256 | break |
| 257 | hstring = b''.join(headers).decode('iso-8859-1') |
Jeremy Hylton | 98eb6c2 | 2009-03-27 18:31:36 +0000 | [diff] [blame] | 258 | return email.parser.Parser(_class=_class).parsestr(hstring) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 259 | |
Antoine Pitrou | b353c12 | 2009-02-11 00:39:14 +0000 | [diff] [blame] | 260 | class HTTPResponse(io.RawIOBase): |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 261 | |
| 262 | # strict: If true, raise BadStatusLine if the status line can't be |
| 263 | # parsed as a valid HTTP/1.0 or 1.1 status line. By default it is |
Skip Montanaro | 186bec2 | 2002-07-25 16:10:38 +0000 | [diff] [blame] | 264 | # false because it prevents clients from talking to HTTP/0.9 |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 265 | # servers. Note that a response with a sufficiently corrupted |
| 266 | # status line will look like an HTTP/0.9 response. |
| 267 | |
| 268 | # See RFC 2616 sec 19.6 and RFC 1945 sec 6 for details. |
| 269 | |
Jeremy Hylton | 811fc14 | 2007-08-03 13:30:02 +0000 | [diff] [blame] | 270 | # The bytes from the socket object are iso-8859-1 strings. |
| 271 | # See RFC 2616 sec 2.2 which notes an exception for MIME-encoded |
| 272 | # text following RFC 2047. The basic status line parsing only |
| 273 | # accepts iso-8859-1. |
| 274 | |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 275 | def __init__(self, sock, debuglevel=0, strict=0, method=None, url=None): |
| 276 | # If the response includes a content-length header, we need to |
| 277 | # make sure that the client doesn't read more than the |
Jeremy Hylton | 39b198d | 2007-08-04 19:22:00 +0000 | [diff] [blame] | 278 | # specified number of bytes. If it does, it will block until |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 279 | # the server times out and closes the connection. This will |
| 280 | # happen if a self.fp.read() is done (without a size) whether |
| 281 | # self.fp is buffered or not. So, no self.fp.read() by |
| 282 | # clients unless they know what they are doing. |
Benjamin Peterson | f72d9fb | 2009-02-08 00:29:20 +0000 | [diff] [blame] | 283 | self.fp = sock.makefile("rb") |
Jeremy Hylton | 30f8674 | 2000-09-18 22:50:38 +0000 | [diff] [blame] | 284 | self.debuglevel = debuglevel |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 285 | self.strict = strict |
Jeremy Hylton | c1b2cb9 | 2003-05-05 16:13:58 +0000 | [diff] [blame] | 286 | self._method = method |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 287 | |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 288 | # The HTTPResponse object is returned via urllib. The clients |
| 289 | # of http and urllib expect different attributes for the |
| 290 | # headers. headers is used here and supports urllib. msg is |
| 291 | # provided as a backwards compatibility layer for http |
| 292 | # clients. |
| 293 | |
| 294 | self.headers = self.msg = None |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 295 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 296 | # from the Status-Line of the response |
Tim Peters | 07e99cb | 2001-01-14 23:47:14 +0000 | [diff] [blame] | 297 | self.version = _UNKNOWN # HTTP-Version |
| 298 | self.status = _UNKNOWN # Status-Code |
| 299 | self.reason = _UNKNOWN # Reason-Phrase |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 300 | |
Tim Peters | 07e99cb | 2001-01-14 23:47:14 +0000 | [diff] [blame] | 301 | self.chunked = _UNKNOWN # is "chunked" being used? |
| 302 | self.chunk_left = _UNKNOWN # bytes left to read in current chunk |
| 303 | self.length = _UNKNOWN # number of bytes left in response |
| 304 | self.will_close = _UNKNOWN # conn will close at end of response |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 305 | |
Jeremy Hylton | be4fcf1 | 2002-06-28 22:38:01 +0000 | [diff] [blame] | 306 | def _read_status(self): |
Jeremy Hylton | 04319c7 | 2007-08-04 03:41:19 +0000 | [diff] [blame] | 307 | # Initialize with Simple-Response defaults. |
Jeremy Hylton | 811fc14 | 2007-08-03 13:30:02 +0000 | [diff] [blame] | 308 | line = str(self.fp.readline(), "iso-8859-1") |
Jeremy Hylton | 30f8674 | 2000-09-18 22:50:38 +0000 | [diff] [blame] | 309 | if self.debuglevel > 0: |
Guido van Rossum | be19ed7 | 2007-02-09 05:37:30 +0000 | [diff] [blame] | 310 | print("reply:", repr(line)) |
Jeremy Hylton | b676952 | 2003-06-29 17:55:05 +0000 | [diff] [blame] | 311 | if not line: |
| 312 | # Presumably, the server closed the connection before |
| 313 | # sending a valid response. |
| 314 | raise BadStatusLine(line) |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 315 | try: |
Guido van Rossum | 34735a6 | 2000-12-15 15:09:42 +0000 | [diff] [blame] | 316 | [version, status, reason] = line.split(None, 2) |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 317 | except ValueError: |
| 318 | try: |
Guido van Rossum | 34735a6 | 2000-12-15 15:09:42 +0000 | [diff] [blame] | 319 | [version, status] = line.split(None, 1) |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 320 | reason = "" |
| 321 | except ValueError: |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 322 | # empty version will cause next test to fail and status |
| 323 | # will be treated as 0.9 response. |
| 324 | version = "" |
Jeremy Hylton | e5d0e84 | 2007-08-03 13:45:24 +0000 | [diff] [blame] | 325 | if not version.startswith("HTTP/"): |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 326 | if self.strict: |
| 327 | self.close() |
| 328 | raise BadStatusLine(line) |
| 329 | else: |
Jeremy Hylton | 04319c7 | 2007-08-04 03:41:19 +0000 | [diff] [blame] | 330 | # Assume it's a Simple-Response from an 0.9 server. |
| 331 | # We have to convert the first line back to raw bytes |
| 332 | # because self.fp.readline() needs to return bytes. |
Guido van Rossum | 70d0dda | 2007-08-29 01:53:26 +0000 | [diff] [blame] | 333 | self.fp = LineAndFileWrapper(bytes(line, "ascii"), self.fp) |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 334 | return "HTTP/0.9", 200, "" |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 335 | |
Jeremy Hylton | 23d4047 | 2001-04-13 14:57:08 +0000 | [diff] [blame] | 336 | # The status code is a three-digit number |
| 337 | try: |
Jeremy Hylton | be4fcf1 | 2002-06-28 22:38:01 +0000 | [diff] [blame] | 338 | status = int(status) |
Jeremy Hylton | 23d4047 | 2001-04-13 14:57:08 +0000 | [diff] [blame] | 339 | if status < 100 or status > 999: |
| 340 | raise BadStatusLine(line) |
| 341 | except ValueError: |
| 342 | raise BadStatusLine(line) |
Jeremy Hylton | be4fcf1 | 2002-06-28 22:38:01 +0000 | [diff] [blame] | 343 | return version, status, reason |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 344 | |
Jeremy Hylton | 39c0380 | 2002-07-12 14:04:09 +0000 | [diff] [blame] | 345 | def begin(self): |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 346 | if self.headers is not None: |
Jeremy Hylton | be4fcf1 | 2002-06-28 22:38:01 +0000 | [diff] [blame] | 347 | # we've already started reading the response |
| 348 | return |
| 349 | |
| 350 | # read until we get a non-100 response |
Raymond Hettinger | b2e0b92 | 2003-02-26 22:45:18 +0000 | [diff] [blame] | 351 | while True: |
Jeremy Hylton | be4fcf1 | 2002-06-28 22:38:01 +0000 | [diff] [blame] | 352 | version, status, reason = self._read_status() |
Martin v. Löwis | 39a3178 | 2004-09-18 09:03:49 +0000 | [diff] [blame] | 353 | if status != CONTINUE: |
Jeremy Hylton | be4fcf1 | 2002-06-28 22:38:01 +0000 | [diff] [blame] | 354 | break |
| 355 | # skip the header from the 100 response |
Raymond Hettinger | b2e0b92 | 2003-02-26 22:45:18 +0000 | [diff] [blame] | 356 | while True: |
Antoine Pitrou | ff1bbba | 2010-12-18 18:04:38 +0000 | [diff] [blame] | 357 | skip = self.fp.readline(_MAXLINE + 1) |
| 358 | if len(skip) > _MAXLINE: |
| 359 | raise LineTooLong("header line") |
| 360 | skip = skip.strip() |
Jeremy Hylton | be4fcf1 | 2002-06-28 22:38:01 +0000 | [diff] [blame] | 361 | if not skip: |
| 362 | break |
| 363 | if self.debuglevel > 0: |
Guido van Rossum | be19ed7 | 2007-02-09 05:37:30 +0000 | [diff] [blame] | 364 | print("header:", skip) |
Tim Peters | c411dba | 2002-07-16 21:35:23 +0000 | [diff] [blame] | 365 | |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 366 | self.code = self.status = status |
Jeremy Hylton | be4fcf1 | 2002-06-28 22:38:01 +0000 | [diff] [blame] | 367 | self.reason = reason.strip() |
Jeremy Hylton | e5d0e84 | 2007-08-03 13:45:24 +0000 | [diff] [blame] | 368 | if version == "HTTP/1.0": |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 369 | self.version = 10 |
Jeremy Hylton | e5d0e84 | 2007-08-03 13:45:24 +0000 | [diff] [blame] | 370 | elif version.startswith("HTTP/1."): |
Tim Peters | 07e99cb | 2001-01-14 23:47:14 +0000 | [diff] [blame] | 371 | self.version = 11 # use HTTP/1.1 code for HTTP/1.x where x>=1 |
Jeremy Hylton | e5d0e84 | 2007-08-03 13:45:24 +0000 | [diff] [blame] | 372 | elif version == "HTTP/0.9": |
Jeremy Hylton | 110941a | 2000-10-12 19:58:36 +0000 | [diff] [blame] | 373 | self.version = 9 |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 374 | else: |
| 375 | raise UnknownProtocol(version) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 376 | |
Jeremy Hylton | 110941a | 2000-10-12 19:58:36 +0000 | [diff] [blame] | 377 | if self.version == 9: |
Georg Brandl | 0aade9a | 2005-06-26 22:06:54 +0000 | [diff] [blame] | 378 | self.length = None |
Jeremy Hylton | 236156f | 2008-12-15 03:00:50 +0000 | [diff] [blame] | 379 | self.chunked = False |
| 380 | self.will_close = True |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 381 | self.headers = self.msg = email.message_from_string('') |
Jeremy Hylton | 110941a | 2000-10-12 19:58:36 +0000 | [diff] [blame] | 382 | return |
| 383 | |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 384 | self.headers = self.msg = parse_headers(self.fp) |
Barry Warsaw | 820c120 | 2008-06-12 04:06:45 +0000 | [diff] [blame] | 385 | |
Jeremy Hylton | 30f8674 | 2000-09-18 22:50:38 +0000 | [diff] [blame] | 386 | if self.debuglevel > 0: |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 387 | for hdr in self.headers: |
Jeremy Hylton | e5d0e84 | 2007-08-03 13:45:24 +0000 | [diff] [blame] | 388 | print("header:", hdr, end=" ") |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 389 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 390 | # are we using the chunked-style of transfer encoding? |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 391 | tr_enc = self.headers.get("transfer-encoding") |
Jeremy Hylton | d229b3a | 2002-09-03 19:24:24 +0000 | [diff] [blame] | 392 | if tr_enc and tr_enc.lower() == "chunked": |
Jeremy Hylton | 236156f | 2008-12-15 03:00:50 +0000 | [diff] [blame] | 393 | self.chunked = True |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 394 | self.chunk_left = None |
| 395 | else: |
Jeremy Hylton | 236156f | 2008-12-15 03:00:50 +0000 | [diff] [blame] | 396 | self.chunked = False |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 397 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 398 | # will the connection close at the end of the response? |
Jeremy Hylton | 22b3a49 | 2002-11-13 17:27:43 +0000 | [diff] [blame] | 399 | self.will_close = self._check_close() |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 400 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 401 | # do we have a Content-Length? |
| 402 | # NOTE: RFC 2616, S4.4, #3 says we ignore this if tr_enc is "chunked" |
Jeremy Hylton | 0ee5eeb | 2007-08-04 03:25:17 +0000 | [diff] [blame] | 403 | self.length = None |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 404 | length = self.headers.get("content-length") |
Barry Warsaw | 820c120 | 2008-06-12 04:06:45 +0000 | [diff] [blame] | 405 | |
| 406 | # are we using the chunked-style of transfer encoding? |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 407 | tr_enc = self.headers.get("transfer-encoding") |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 408 | if length and not self.chunked: |
Jeremy Hylton | 30a8181 | 2000-09-14 20:34:27 +0000 | [diff] [blame] | 409 | try: |
| 410 | self.length = int(length) |
| 411 | except ValueError: |
Christian Heimes | a612dc0 | 2008-02-24 13:08:18 +0000 | [diff] [blame] | 412 | self.length = None |
| 413 | else: |
| 414 | if self.length < 0: # ignore nonsensical negative lengths |
| 415 | self.length = None |
| 416 | else: |
| 417 | self.length = None |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 418 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 419 | # does the body have a fixed length? (of zero) |
Martin v. Löwis | 39a3178 | 2004-09-18 09:03:49 +0000 | [diff] [blame] | 420 | if (status == NO_CONTENT or status == NOT_MODIFIED or |
Jeremy Hylton | c1b2cb9 | 2003-05-05 16:13:58 +0000 | [diff] [blame] | 421 | 100 <= status < 200 or # 1xx codes |
Jeremy Hylton | e5d0e84 | 2007-08-03 13:45:24 +0000 | [diff] [blame] | 422 | self._method == "HEAD"): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 423 | self.length = 0 |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 424 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 425 | # if the connection remains open, and we aren't using chunked, and |
| 426 | # a content-length was not provided, then assume that the connection |
| 427 | # WILL close. |
Jeremy Hylton | 0ee5eeb | 2007-08-04 03:25:17 +0000 | [diff] [blame] | 428 | if (not self.will_close and |
| 429 | not self.chunked and |
| 430 | self.length is None): |
Jeremy Hylton | 236156f | 2008-12-15 03:00:50 +0000 | [diff] [blame] | 431 | self.will_close = True |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 432 | |
Jeremy Hylton | 22b3a49 | 2002-11-13 17:27:43 +0000 | [diff] [blame] | 433 | def _check_close(self): |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 434 | conn = self.headers.get("connection") |
Jeremy Hylton | 22b3a49 | 2002-11-13 17:27:43 +0000 | [diff] [blame] | 435 | if self.version == 11: |
| 436 | # An HTTP/1.1 proxy is assumed to stay open unless |
| 437 | # explicitly closed. |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 438 | conn = self.headers.get("connection") |
Raymond Hettinger | bac788a | 2004-05-04 09:21:43 +0000 | [diff] [blame] | 439 | if conn and "close" in conn.lower(): |
Jeremy Hylton | 22b3a49 | 2002-11-13 17:27:43 +0000 | [diff] [blame] | 440 | return True |
| 441 | return False |
| 442 | |
Jeremy Hylton | 2c17825 | 2004-08-07 16:28:14 +0000 | [diff] [blame] | 443 | # Some HTTP/1.0 implementations have support for persistent |
| 444 | # connections, using rules different than HTTP/1.1. |
Jeremy Hylton | 22b3a49 | 2002-11-13 17:27:43 +0000 | [diff] [blame] | 445 | |
Christian Heimes | 895627f | 2007-12-08 17:28:33 +0000 | [diff] [blame] | 446 | # For older HTTP, Keep-Alive indicates persistent connection. |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 447 | if self.headers.get("keep-alive"): |
Jeremy Hylton | 22b3a49 | 2002-11-13 17:27:43 +0000 | [diff] [blame] | 448 | return False |
Tim Peters | 77c06fb | 2002-11-24 02:35:35 +0000 | [diff] [blame] | 449 | |
Jeremy Hylton | 2c17825 | 2004-08-07 16:28:14 +0000 | [diff] [blame] | 450 | # At least Akamai returns a "Connection: Keep-Alive" header, |
| 451 | # which was supposed to be sent by the client. |
| 452 | if conn and "keep-alive" in conn.lower(): |
| 453 | return False |
| 454 | |
Jeremy Hylton | 22b3a49 | 2002-11-13 17:27:43 +0000 | [diff] [blame] | 455 | # Proxy-Connection is a netscape hack. |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 456 | pconn = self.headers.get("proxy-connection") |
Raymond Hettinger | bac788a | 2004-05-04 09:21:43 +0000 | [diff] [blame] | 457 | if pconn and "keep-alive" in pconn.lower(): |
Jeremy Hylton | 22b3a49 | 2002-11-13 17:27:43 +0000 | [diff] [blame] | 458 | return False |
| 459 | |
| 460 | # otherwise, assume it will close |
| 461 | return True |
| 462 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 463 | def close(self): |
| 464 | if self.fp: |
| 465 | self.fp.close() |
| 466 | self.fp = None |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 467 | |
Jeremy Hylton | df5f6b5 | 2007-08-08 17:36:33 +0000 | [diff] [blame] | 468 | # These implementations are for the benefit of io.BufferedReader. |
| 469 | |
| 470 | # XXX This class should probably be revised to act more like |
| 471 | # the "raw stream" that BufferedReader expects. |
| 472 | |
| 473 | @property |
| 474 | def closed(self): |
| 475 | return self.isclosed() |
| 476 | |
| 477 | def flush(self): |
| 478 | self.fp.flush() |
| 479 | |
Jeremy Hylton | a7cff02 | 2009-04-01 02:35:56 +0000 | [diff] [blame] | 480 | def readable(self): |
| 481 | return True |
| 482 | |
Jeremy Hylton | df5f6b5 | 2007-08-08 17:36:33 +0000 | [diff] [blame] | 483 | # End of "raw stream" methods |
| 484 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 485 | def isclosed(self): |
| 486 | # NOTE: it is possible that we will not ever call self.close(). This |
| 487 | # case occurs when will_close is TRUE, length is None, and we |
| 488 | # read up to the last byte, but NOT past it. |
| 489 | # |
| 490 | # IMPLIES: if will_close is FALSE, then self.close() will ALWAYS be |
| 491 | # called, meaning self.isclosed() is meaningful. |
| 492 | return self.fp is None |
| 493 | |
| 494 | def read(self, amt=None): |
| 495 | if self.fp is None: |
Guido van Rossum | 8ce8a78 | 2007-11-01 19:42:39 +0000 | [diff] [blame] | 496 | return b"" |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 497 | |
Senthil Kumaran | 2682638 | 2010-04-28 17:44:30 +0000 | [diff] [blame] | 498 | if self._method == "HEAD": |
Senthil Kumaran | 41142f5 | 2010-06-04 16:43:10 +0000 | [diff] [blame] | 499 | self.close() |
Senthil Kumaran | 2682638 | 2010-04-28 17:44:30 +0000 | [diff] [blame] | 500 | return b"" |
| 501 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 502 | if self.chunked: |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 503 | return self._read_chunked(amt) |
Tim Peters | 230a60c | 2002-11-09 05:08:07 +0000 | [diff] [blame] | 504 | |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 505 | if amt is None: |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 506 | # unbounded read |
Jeremy Hylton | def9d2a | 2004-11-07 16:13:49 +0000 | [diff] [blame] | 507 | if self.length is None: |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 508 | s = self.fp.read() |
| 509 | else: |
| 510 | s = self._safe_read(self.length) |
Jeremy Hylton | def9d2a | 2004-11-07 16:13:49 +0000 | [diff] [blame] | 511 | self.length = 0 |
Tim Peters | 07e99cb | 2001-01-14 23:47:14 +0000 | [diff] [blame] | 512 | self.close() # we read everything |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 513 | return s |
| 514 | |
| 515 | if self.length is not None: |
| 516 | if amt > self.length: |
| 517 | # clip the read to the "end of response" |
| 518 | amt = self.length |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 519 | |
| 520 | # we do not use _safe_read() here because this may be a .will_close |
| 521 | # connection, and the user is reading more bytes than will be provided |
| 522 | # (for example, reading in 1k chunks) |
| 523 | s = self.fp.read(amt) |
Jeremy Hylton | def9d2a | 2004-11-07 16:13:49 +0000 | [diff] [blame] | 524 | if self.length is not None: |
| 525 | self.length -= len(s) |
Guido van Rossum | 8ce8a78 | 2007-11-01 19:42:39 +0000 | [diff] [blame] | 526 | if not self.length: |
| 527 | self.close() |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 528 | return s |
| 529 | |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 530 | def _read_chunked(self, amt): |
| 531 | assert self.chunked != _UNKNOWN |
| 532 | chunk_left = self.chunk_left |
Chris Withers | 5a86acb | 2009-09-04 17:15:46 +0000 | [diff] [blame] | 533 | value = [] |
Raymond Hettinger | b2e0b92 | 2003-02-26 22:45:18 +0000 | [diff] [blame] | 534 | while True: |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 535 | if chunk_left is None: |
Antoine Pitrou | ff1bbba | 2010-12-18 18:04:38 +0000 | [diff] [blame] | 536 | line = self.fp.readline(_MAXLINE + 1) |
| 537 | if len(line) > _MAXLINE: |
| 538 | raise LineTooLong("chunk size") |
Georg Brandl | 95ba469 | 2008-01-26 09:45:58 +0000 | [diff] [blame] | 539 | i = line.find(b";") |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 540 | if i >= 0: |
| 541 | line = line[:i] # strip chunk-extensions |
Christian Heimes | a612dc0 | 2008-02-24 13:08:18 +0000 | [diff] [blame] | 542 | try: |
| 543 | chunk_left = int(line, 16) |
| 544 | except ValueError: |
| 545 | # close the connection as protocol synchronisation is |
| 546 | # probably lost |
| 547 | self.close() |
Chris Withers | 5a86acb | 2009-09-04 17:15:46 +0000 | [diff] [blame] | 548 | raise IncompleteRead(b''.join(value)) |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 549 | if chunk_left == 0: |
| 550 | break |
| 551 | if amt is None: |
Chris Withers | 5a86acb | 2009-09-04 17:15:46 +0000 | [diff] [blame] | 552 | value.append(self._safe_read(chunk_left)) |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 553 | elif amt < chunk_left: |
Chris Withers | 5a86acb | 2009-09-04 17:15:46 +0000 | [diff] [blame] | 554 | value.append(self._safe_read(amt)) |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 555 | self.chunk_left = chunk_left - amt |
Chris Withers | 5a86acb | 2009-09-04 17:15:46 +0000 | [diff] [blame] | 556 | return b''.join(value) |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 557 | elif amt == chunk_left: |
Chris Withers | 5a86acb | 2009-09-04 17:15:46 +0000 | [diff] [blame] | 558 | value.append(self._safe_read(amt)) |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 559 | self._safe_read(2) # toss the CRLF at the end of the chunk |
| 560 | self.chunk_left = None |
Chris Withers | 5a86acb | 2009-09-04 17:15:46 +0000 | [diff] [blame] | 561 | return b''.join(value) |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 562 | else: |
Chris Withers | 5a86acb | 2009-09-04 17:15:46 +0000 | [diff] [blame] | 563 | value.append(self._safe_read(chunk_left)) |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 564 | amt -= chunk_left |
| 565 | |
| 566 | # we read the whole chunk, get another |
| 567 | self._safe_read(2) # toss the CRLF at the end of the chunk |
| 568 | chunk_left = None |
| 569 | |
| 570 | # read and discard trailer up to the CRLF terminator |
| 571 | ### note: we shouldn't have any trailers! |
Raymond Hettinger | b2e0b92 | 2003-02-26 22:45:18 +0000 | [diff] [blame] | 572 | while True: |
Antoine Pitrou | ff1bbba | 2010-12-18 18:04:38 +0000 | [diff] [blame] | 573 | line = self.fp.readline(_MAXLINE + 1) |
| 574 | if len(line) > _MAXLINE: |
| 575 | raise LineTooLong("trailer line") |
Christian Heimes | 0bd4e11 | 2008-02-12 22:59:25 +0000 | [diff] [blame] | 576 | if not line: |
| 577 | # a vanishingly small number of sites EOF without |
| 578 | # sending the trailer |
| 579 | break |
Guido van Rossum | 8ce8a78 | 2007-11-01 19:42:39 +0000 | [diff] [blame] | 580 | if line == b"\r\n": |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 581 | break |
| 582 | |
| 583 | # we read everything; close the "file" |
Jeremy Hylton | d4c472c | 2002-09-03 20:49:06 +0000 | [diff] [blame] | 584 | self.close() |
| 585 | |
Chris Withers | 5a86acb | 2009-09-04 17:15:46 +0000 | [diff] [blame] | 586 | return b''.join(value) |
Tim Peters | 230a60c | 2002-11-09 05:08:07 +0000 | [diff] [blame] | 587 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 588 | def _safe_read(self, amt): |
| 589 | """Read the number of bytes requested, compensating for partial reads. |
| 590 | |
| 591 | Normally, we have a blocking socket, but a read() can be interrupted |
| 592 | by a signal (resulting in a partial read). |
| 593 | |
| 594 | Note that we cannot distinguish between EOF and an interrupt when zero |
| 595 | bytes have been read. IncompleteRead() will be raised in this |
| 596 | situation. |
| 597 | |
| 598 | This function should be used when <amt> bytes "should" be present for |
| 599 | reading. If the bytes are truly not available (due to EOF), then the |
| 600 | IncompleteRead exception can be used to detect the problem. |
| 601 | """ |
Georg Brandl | 80ba8e8 | 2005-09-29 20:16:07 +0000 | [diff] [blame] | 602 | s = [] |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 603 | while amt > 0: |
Georg Brandl | 80ba8e8 | 2005-09-29 20:16:07 +0000 | [diff] [blame] | 604 | chunk = self.fp.read(min(amt, MAXAMOUNT)) |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 605 | if not chunk: |
Benjamin Peterson | 6accb98 | 2009-03-02 22:50:25 +0000 | [diff] [blame] | 606 | raise IncompleteRead(b''.join(s), amt) |
Georg Brandl | 80ba8e8 | 2005-09-29 20:16:07 +0000 | [diff] [blame] | 607 | s.append(chunk) |
Raymond Hettinger | b2e0b92 | 2003-02-26 22:45:18 +0000 | [diff] [blame] | 608 | amt -= len(chunk) |
Guido van Rossum | a00f123 | 2007-09-12 19:43:09 +0000 | [diff] [blame] | 609 | return b"".join(s) |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 610 | |
Antoine Pitrou | b353c12 | 2009-02-11 00:39:14 +0000 | [diff] [blame] | 611 | def fileno(self): |
| 612 | return self.fp.fileno() |
| 613 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 614 | def getheader(self, name, default=None): |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 615 | if self.headers is None: |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 616 | raise ResponseNotReady() |
Senthil Kumaran | 3357840 | 2010-08-02 12:01:21 +0000 | [diff] [blame] | 617 | headers = self.headers.get_all(name) or default |
| 618 | if isinstance(headers, str) or not hasattr(headers, '__iter__'): |
| 619 | return headers |
| 620 | else: |
| 621 | return ', '.join(headers) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 622 | |
Martin v. Löwis | deacce2 | 2004-08-18 12:46:26 +0000 | [diff] [blame] | 623 | def getheaders(self): |
| 624 | """Return list of (header, value) tuples.""" |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 625 | if self.headers is None: |
Martin v. Löwis | deacce2 | 2004-08-18 12:46:26 +0000 | [diff] [blame] | 626 | raise ResponseNotReady() |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 627 | return list(self.headers.items()) |
Martin v. Löwis | deacce2 | 2004-08-18 12:46:26 +0000 | [diff] [blame] | 628 | |
Antoine Pitrou | b353c12 | 2009-02-11 00:39:14 +0000 | [diff] [blame] | 629 | # We override IOBase.__iter__ so that it doesn't check for closed-ness |
| 630 | |
| 631 | def __iter__(self): |
| 632 | return self |
| 633 | |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 634 | # For compatibility with old-style urllib responses. |
| 635 | |
| 636 | def info(self): |
| 637 | return self.headers |
| 638 | |
| 639 | def geturl(self): |
| 640 | return self.url |
| 641 | |
| 642 | def getcode(self): |
| 643 | return self.status |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 644 | |
| 645 | class HTTPConnection: |
| 646 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 647 | _http_vsn = 11 |
| 648 | _http_vsn_str = 'HTTP/1.1' |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 649 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 650 | response_class = HTTPResponse |
| 651 | default_port = HTTP_PORT |
| 652 | auto_open = 1 |
Jeremy Hylton | 30f8674 | 2000-09-18 22:50:38 +0000 | [diff] [blame] | 653 | debuglevel = 0 |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 654 | strict = 0 |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 655 | |
Georg Brandl | f78e02b | 2008-06-10 17:40:04 +0000 | [diff] [blame] | 656 | def __init__(self, host, port=None, strict=None, |
| 657 | timeout=socket._GLOBAL_DEFAULT_TIMEOUT): |
Guido van Rossum | d8faa36 | 2007-04-27 19:54:29 +0000 | [diff] [blame] | 658 | self.timeout = timeout |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 659 | self.sock = None |
Jeremy Hylton | 8531b1b | 2002-07-16 21:21:11 +0000 | [diff] [blame] | 660 | self._buffer = [] |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 661 | self.__response = None |
| 662 | self.__state = _CS_IDLE |
Jeremy Hylton | c1b2cb9 | 2003-05-05 16:13:58 +0000 | [diff] [blame] | 663 | self._method = None |
Senthil Kumaran | 0ac1f83 | 2009-07-26 12:39:47 +0000 | [diff] [blame] | 664 | self._tunnel_host = None |
| 665 | self._tunnel_port = None |
Georg Brandl | 2774310 | 2011-02-25 10:18:11 +0000 | [diff] [blame] | 666 | self._tunnel_headers = {} |
Tim Peters | c411dba | 2002-07-16 21:35:23 +0000 | [diff] [blame] | 667 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 668 | self._set_hostport(host, port) |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 669 | if strict is not None: |
| 670 | self.strict = strict |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 671 | |
Senthil Kumaran | 4b9fbeb | 2009-12-20 07:18:22 +0000 | [diff] [blame] | 672 | def _set_tunnel(self, host, port=None, headers=None): |
Senthil Kumaran | 0ac1f83 | 2009-07-26 12:39:47 +0000 | [diff] [blame] | 673 | self._tunnel_host = host |
| 674 | self._tunnel_port = port |
Senthil Kumaran | 4b9fbeb | 2009-12-20 07:18:22 +0000 | [diff] [blame] | 675 | if headers: |
| 676 | self._tunnel_headers = headers |
| 677 | else: |
| 678 | self._tunnel_headers.clear() |
Senthil Kumaran | 0ac1f83 | 2009-07-26 12:39:47 +0000 | [diff] [blame] | 679 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 680 | def _set_hostport(self, host, port): |
| 681 | if port is None: |
Skip Montanaro | 10e6e0e | 2004-09-14 16:32:02 +0000 | [diff] [blame] | 682 | i = host.rfind(':') |
Skip Montanaro | cae14d2 | 2004-09-14 17:55:21 +0000 | [diff] [blame] | 683 | j = host.rfind(']') # ipv6 addresses have [...] |
| 684 | if i > j: |
Skip Montanaro | 9d38997 | 2002-03-24 16:53:50 +0000 | [diff] [blame] | 685 | try: |
| 686 | port = int(host[i+1:]) |
| 687 | except ValueError: |
Jeremy Hylton | fbd7994 | 2002-07-02 20:19:08 +0000 | [diff] [blame] | 688 | raise InvalidURL("nonnumeric port: '%s'" % host[i+1:]) |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 689 | host = host[:i] |
| 690 | else: |
| 691 | port = self.default_port |
Raymond Hettinger | 4d03791 | 2004-10-14 15:23:38 +0000 | [diff] [blame] | 692 | if host and host[0] == '[' and host[-1] == ']': |
Brett Cannon | 0a1af4a | 2004-09-15 23:26:23 +0000 | [diff] [blame] | 693 | host = host[1:-1] |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 694 | self.host = host |
| 695 | self.port = port |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 696 | |
Jeremy Hylton | 30f8674 | 2000-09-18 22:50:38 +0000 | [diff] [blame] | 697 | def set_debuglevel(self, level): |
| 698 | self.debuglevel = level |
| 699 | |
Senthil Kumaran | 0ac1f83 | 2009-07-26 12:39:47 +0000 | [diff] [blame] | 700 | def _tunnel(self): |
| 701 | self._set_hostport(self._tunnel_host, self._tunnel_port) |
Georg Brandl | 2774310 | 2011-02-25 10:18:11 +0000 | [diff] [blame] | 702 | connect_str = "CONNECT %s:%d HTTP/1.0\r\n" % (self.host, self.port) |
Senthil Kumaran | 0ac1f83 | 2009-07-26 12:39:47 +0000 | [diff] [blame] | 703 | connect_bytes = connect_str.encode("ascii") |
| 704 | self.send(connect_bytes) |
Georg Brandl | 2774310 | 2011-02-25 10:18:11 +0000 | [diff] [blame] | 705 | for header, value in self._tunnel_headers.items(): |
Senthil Kumaran | 4b9fbeb | 2009-12-20 07:18:22 +0000 | [diff] [blame] | 706 | header_str = "%s: %s\r\n" % (header, value) |
| 707 | header_bytes = header_str.encode("ascii") |
| 708 | self.send(header_bytes) |
Georg Brandl | 2774310 | 2011-02-25 10:18:11 +0000 | [diff] [blame] | 709 | self.send(b'\r\n') |
Senthil Kumaran | 4b9fbeb | 2009-12-20 07:18:22 +0000 | [diff] [blame] | 710 | |
Senthil Kumaran | 0ac1f83 | 2009-07-26 12:39:47 +0000 | [diff] [blame] | 711 | response = self.response_class(self.sock, strict = self.strict, |
Senthil Kumaran | 4b9fbeb | 2009-12-20 07:18:22 +0000 | [diff] [blame] | 712 | method = self._method) |
Senthil Kumaran | 0ac1f83 | 2009-07-26 12:39:47 +0000 | [diff] [blame] | 713 | (version, code, message) = response._read_status() |
Senthil Kumaran | 4b9fbeb | 2009-12-20 07:18:22 +0000 | [diff] [blame] | 714 | |
Senthil Kumaran | 0ac1f83 | 2009-07-26 12:39:47 +0000 | [diff] [blame] | 715 | if code != 200: |
| 716 | self.close() |
| 717 | raise socket.error("Tunnel connection failed: %d %s" % (code, |
| 718 | message.strip())) |
| 719 | while True: |
Antoine Pitrou | ff1bbba | 2010-12-18 18:04:38 +0000 | [diff] [blame] | 720 | line = response.fp.readline(_MAXLINE + 1) |
| 721 | if len(line) > _MAXLINE: |
| 722 | raise LineTooLong("header line") |
Senthil Kumaran | 0ac1f83 | 2009-07-26 12:39:47 +0000 | [diff] [blame] | 723 | if line == b'\r\n': |
| 724 | break |
| 725 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 726 | def connect(self): |
| 727 | """Connect to the host and port specified in __init__.""" |
Guido van Rossum | d8faa36 | 2007-04-27 19:54:29 +0000 | [diff] [blame] | 728 | self.sock = socket.create_connection((self.host,self.port), |
| 729 | self.timeout) |
Senthil Kumaran | 0ac1f83 | 2009-07-26 12:39:47 +0000 | [diff] [blame] | 730 | if self._tunnel_host: |
| 731 | self._tunnel() |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 732 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 733 | def close(self): |
| 734 | """Close the connection to the HTTP server.""" |
| 735 | if self.sock: |
Tim Peters | 07e99cb | 2001-01-14 23:47:14 +0000 | [diff] [blame] | 736 | self.sock.close() # close it manually... there may be other refs |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 737 | self.sock = None |
| 738 | if self.__response: |
| 739 | self.__response.close() |
| 740 | self.__response = None |
| 741 | self.__state = _CS_IDLE |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 742 | |
Senthil Kumaran | 996e13c | 2010-10-02 10:35:24 +0000 | [diff] [blame] | 743 | def send(self, data): |
| 744 | """Send `data' to the server.""" |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 745 | if self.sock is None: |
| 746 | if self.auto_open: |
| 747 | self.connect() |
| 748 | else: |
| 749 | raise NotConnected() |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 750 | |
Jeremy Hylton | e3252ec | 2002-07-16 21:41:43 +0000 | [diff] [blame] | 751 | # send the data to the server. if we get a broken pipe, then close |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 752 | # the socket. we want to reconnect when somebody tries to send again. |
| 753 | # |
| 754 | # NOTE: we DO propagate the error, though, because we cannot simply |
| 755 | # ignore the error... the caller will know if they can retry. |
Jeremy Hylton | 30f8674 | 2000-09-18 22:50:38 +0000 | [diff] [blame] | 756 | if self.debuglevel > 0: |
Senthil Kumaran | 996e13c | 2010-10-02 10:35:24 +0000 | [diff] [blame] | 757 | print("send:", repr(data)) |
Jeremy Hylton | 636950f | 2009-03-28 04:34:21 +0000 | [diff] [blame] | 758 | blocksize = 8192 |
Senthil Kumaran | 996e13c | 2010-10-02 10:35:24 +0000 | [diff] [blame] | 759 | if hasattr(data, "read") : |
Jeremy Hylton | 636950f | 2009-03-28 04:34:21 +0000 | [diff] [blame] | 760 | if self.debuglevel > 0: |
| 761 | print("sendIng a read()able") |
| 762 | encode = False |
Antoine Pitrou | 595ad32 | 2009-09-29 19:02:24 +0000 | [diff] [blame] | 763 | try: |
Senthil Kumaran | 996e13c | 2010-10-02 10:35:24 +0000 | [diff] [blame] | 764 | mode = data.mode |
Antoine Pitrou | 595ad32 | 2009-09-29 19:02:24 +0000 | [diff] [blame] | 765 | except AttributeError: |
| 766 | # io.BytesIO and other file-like objects don't have a `mode` |
| 767 | # attribute. |
| 768 | pass |
| 769 | else: |
| 770 | if "b" not in mode: |
| 771 | encode = True |
| 772 | if self.debuglevel > 0: |
| 773 | print("encoding file using iso-8859-1") |
Jeremy Hylton | 636950f | 2009-03-28 04:34:21 +0000 | [diff] [blame] | 774 | while 1: |
Senthil Kumaran | 996e13c | 2010-10-02 10:35:24 +0000 | [diff] [blame] | 775 | datablock = data.read(blocksize) |
| 776 | if not datablock: |
Jeremy Hylton | 636950f | 2009-03-28 04:34:21 +0000 | [diff] [blame] | 777 | break |
| 778 | if encode: |
Senthil Kumaran | 996e13c | 2010-10-02 10:35:24 +0000 | [diff] [blame] | 779 | datablock = datablock.encode("iso-8859-1") |
| 780 | self.sock.sendall(datablock) |
Jeremy Hylton | 636950f | 2009-03-28 04:34:21 +0000 | [diff] [blame] | 781 | else: |
Senthil Kumaran | 996e13c | 2010-10-02 10:35:24 +0000 | [diff] [blame] | 782 | self.sock.sendall(data) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 783 | |
Jeremy Hylton | 8531b1b | 2002-07-16 21:21:11 +0000 | [diff] [blame] | 784 | def _output(self, s): |
| 785 | """Add a line of output to the current request buffer. |
Tim Peters | 469cdad | 2002-08-08 20:19:19 +0000 | [diff] [blame] | 786 | |
Jeremy Hylton | e3252ec | 2002-07-16 21:41:43 +0000 | [diff] [blame] | 787 | Assumes that the line does *not* end with \\r\\n. |
Jeremy Hylton | 8531b1b | 2002-07-16 21:21:11 +0000 | [diff] [blame] | 788 | """ |
| 789 | self._buffer.append(s) |
| 790 | |
Benjamin Peterson | 1742e40 | 2008-11-30 22:15:29 +0000 | [diff] [blame] | 791 | def _send_output(self, message_body=None): |
Jeremy Hylton | 8531b1b | 2002-07-16 21:21:11 +0000 | [diff] [blame] | 792 | """Send the currently buffered request and clear the buffer. |
| 793 | |
Jeremy Hylton | e3252ec | 2002-07-16 21:41:43 +0000 | [diff] [blame] | 794 | Appends an extra \\r\\n to the buffer. |
Benjamin Peterson | 822b21c | 2009-01-18 00:04:57 +0000 | [diff] [blame] | 795 | A message_body may be specified, to be appended to the request. |
Jeremy Hylton | 8531b1b | 2002-07-16 21:21:11 +0000 | [diff] [blame] | 796 | """ |
Martin v. Löwis | dd5a860 | 2007-06-30 09:22:09 +0000 | [diff] [blame] | 797 | self._buffer.extend((b"", b"")) |
| 798 | msg = b"\r\n".join(self._buffer) |
Jeremy Hylton | 8531b1b | 2002-07-16 21:21:11 +0000 | [diff] [blame] | 799 | del self._buffer[:] |
Benjamin Peterson | 1742e40 | 2008-11-30 22:15:29 +0000 | [diff] [blame] | 800 | # If msg and message_body are sent in a single send() call, |
| 801 | # it will avoid performance problems caused by the interaction |
| 802 | # between delayed ack and the Nagle algorithim. |
Benjamin Peterson | 822b21c | 2009-01-18 00:04:57 +0000 | [diff] [blame] | 803 | if isinstance(message_body, bytes): |
Benjamin Peterson | 1742e40 | 2008-11-30 22:15:29 +0000 | [diff] [blame] | 804 | msg += message_body |
Benjamin Peterson | 822b21c | 2009-01-18 00:04:57 +0000 | [diff] [blame] | 805 | message_body = None |
Jeremy Hylton | 8531b1b | 2002-07-16 21:21:11 +0000 | [diff] [blame] | 806 | self.send(msg) |
Benjamin Peterson | 822b21c | 2009-01-18 00:04:57 +0000 | [diff] [blame] | 807 | if message_body is not None: |
Jeremy Hylton | 236654b | 2009-03-27 20:24:34 +0000 | [diff] [blame] | 808 | # message_body was not a string (i.e. it is a file), and |
| 809 | # we must run the risk of Nagle. |
Benjamin Peterson | 822b21c | 2009-01-18 00:04:57 +0000 | [diff] [blame] | 810 | self.send(message_body) |
Jeremy Hylton | 8531b1b | 2002-07-16 21:21:11 +0000 | [diff] [blame] | 811 | |
Martin v. Löwis | af7dc8d | 2003-11-19 19:51:55 +0000 | [diff] [blame] | 812 | def putrequest(self, method, url, skip_host=0, skip_accept_encoding=0): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 813 | """Send a request to the server. |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 814 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 815 | `method' specifies an HTTP request method, e.g. 'GET'. |
| 816 | `url' specifies the object being requested, e.g. '/index.html'. |
Martin v. Löwis | af7dc8d | 2003-11-19 19:51:55 +0000 | [diff] [blame] | 817 | `skip_host' if True does not add automatically a 'Host:' header |
| 818 | `skip_accept_encoding' if True does not add automatically an |
| 819 | 'Accept-Encoding:' header |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 820 | """ |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 821 | |
Greg Stein | 616a58d | 2003-06-24 06:35:19 +0000 | [diff] [blame] | 822 | # if a prior response has been completed, then forget about it. |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 823 | if self.__response and self.__response.isclosed(): |
| 824 | self.__response = None |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 825 | |
Tim Peters | 58eb11c | 2004-01-18 20:29:55 +0000 | [diff] [blame] | 826 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 827 | # in certain cases, we cannot issue another request on this connection. |
| 828 | # this occurs when: |
| 829 | # 1) we are in the process of sending a request. (_CS_REQ_STARTED) |
| 830 | # 2) a response to a previous request has signalled that it is going |
| 831 | # to close the connection upon completion. |
| 832 | # 3) the headers for the previous response have not been read, thus |
| 833 | # we cannot determine whether point (2) is true. (_CS_REQ_SENT) |
| 834 | # |
| 835 | # if there is no prior response, then we can request at will. |
| 836 | # |
| 837 | # if point (2) is true, then we will have passed the socket to the |
| 838 | # response (effectively meaning, "there is no prior response"), and |
| 839 | # will open a new one when a new request is made. |
| 840 | # |
| 841 | # Note: if a prior response exists, then we *can* start a new request. |
| 842 | # We are not allowed to begin fetching the response to this new |
| 843 | # request, however, until that prior response is complete. |
| 844 | # |
| 845 | if self.__state == _CS_IDLE: |
| 846 | self.__state = _CS_REQ_STARTED |
| 847 | else: |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 848 | raise CannotSendRequest(self.__state) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 849 | |
Jeremy Hylton | c1b2cb9 | 2003-05-05 16:13:58 +0000 | [diff] [blame] | 850 | # Save the method we use, we need it later in the response phase |
| 851 | self._method = method |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 852 | if not url: |
| 853 | url = '/' |
Martin v. Löwis | dd5a860 | 2007-06-30 09:22:09 +0000 | [diff] [blame] | 854 | request = '%s %s %s' % (method, url, self._http_vsn_str) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 855 | |
Martin v. Löwis | dd5a860 | 2007-06-30 09:22:09 +0000 | [diff] [blame] | 856 | # Non-ASCII characters should have been eliminated earlier |
| 857 | self._output(request.encode('ascii')) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 858 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 859 | if self._http_vsn == 11: |
| 860 | # Issue some standard headers for better HTTP/1.1 compliance |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 861 | |
Jeremy Hylton | 3921ff6 | 2002-03-09 06:07:23 +0000 | [diff] [blame] | 862 | if not skip_host: |
| 863 | # this header is issued *only* for HTTP/1.1 |
| 864 | # connections. more specifically, this means it is |
| 865 | # only issued when the client uses the new |
| 866 | # HTTPConnection() class. backwards-compat clients |
| 867 | # will be using HTTP/1.0 and those clients may be |
| 868 | # issuing this header themselves. we should NOT issue |
| 869 | # it twice; some web servers (such as Apache) barf |
| 870 | # when they see two Host: headers |
Guido van Rossum | f6922aa | 2001-01-14 21:03:01 +0000 | [diff] [blame] | 871 | |
Jeremy Hylton | 3921ff6 | 2002-03-09 06:07:23 +0000 | [diff] [blame] | 872 | # If we need a non-standard port,include it in the |
| 873 | # header. If the request is going through a proxy, |
| 874 | # but the host of the actual URL, not the host of the |
| 875 | # proxy. |
Jeremy Hylton | 8acf1e0 | 2002-03-08 19:35:51 +0000 | [diff] [blame] | 876 | |
Jeremy Hylton | 3921ff6 | 2002-03-09 06:07:23 +0000 | [diff] [blame] | 877 | netloc = '' |
| 878 | if url.startswith('http'): |
| 879 | nil, netloc, nil, nil, nil = urlsplit(url) |
| 880 | |
| 881 | if netloc: |
Thomas Wouters | 477c8d5 | 2006-05-27 19:21:47 +0000 | [diff] [blame] | 882 | try: |
| 883 | netloc_enc = netloc.encode("ascii") |
| 884 | except UnicodeEncodeError: |
| 885 | netloc_enc = netloc.encode("idna") |
| 886 | self.putheader('Host', netloc_enc) |
Jeremy Hylton | 3921ff6 | 2002-03-09 06:07:23 +0000 | [diff] [blame] | 887 | else: |
Thomas Wouters | 477c8d5 | 2006-05-27 19:21:47 +0000 | [diff] [blame] | 888 | try: |
| 889 | host_enc = self.host.encode("ascii") |
| 890 | except UnicodeEncodeError: |
| 891 | host_enc = self.host.encode("idna") |
Senthil Kumaran | 2e89cf3 | 2010-11-14 03:14:52 +0000 | [diff] [blame] | 892 | |
| 893 | # As per RFC 273, IPv6 address should be wrapped with [] |
| 894 | # when used as Host header |
| 895 | |
| 896 | if self.host.find(':') >= 0: |
| 897 | host_enc = b'[' + host_enc + b']' |
| 898 | |
Georg Brandl | 86b2fb9 | 2008-07-16 03:43:04 +0000 | [diff] [blame] | 899 | if self.port == self.default_port: |
Thomas Wouters | 477c8d5 | 2006-05-27 19:21:47 +0000 | [diff] [blame] | 900 | self.putheader('Host', host_enc) |
| 901 | else: |
Guido van Rossum | 98297ee | 2007-11-06 21:34:58 +0000 | [diff] [blame] | 902 | host_enc = host_enc.decode("ascii") |
Thomas Wouters | 477c8d5 | 2006-05-27 19:21:47 +0000 | [diff] [blame] | 903 | self.putheader('Host', "%s:%s" % (host_enc, self.port)) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 904 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 905 | # note: we are assuming that clients will not attempt to set these |
| 906 | # headers since *this* library must deal with the |
| 907 | # consequences. this also means that when the supporting |
| 908 | # libraries are updated to recognize other forms, then this |
| 909 | # code should be changed (removed or updated). |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 910 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 911 | # we only want a Content-Encoding of "identity" since we don't |
| 912 | # support encodings such as x-gzip or x-deflate. |
Martin v. Löwis | af7dc8d | 2003-11-19 19:51:55 +0000 | [diff] [blame] | 913 | if not skip_accept_encoding: |
| 914 | self.putheader('Accept-Encoding', 'identity') |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 915 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 916 | # we can accept "chunked" Transfer-Encodings, but no others |
| 917 | # NOTE: no TE header implies *only* "chunked" |
| 918 | #self.putheader('TE', 'chunked') |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 919 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 920 | # if TE is supplied in the header, then it must appear in a |
| 921 | # Connection header. |
| 922 | #self.putheader('Connection', 'TE') |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 923 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 924 | else: |
| 925 | # For HTTP/1.0, the server will assume "not chunked" |
| 926 | pass |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 927 | |
Benjamin Peterson | f608c61 | 2008-11-16 18:33:53 +0000 | [diff] [blame] | 928 | def putheader(self, header, *values): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 929 | """Send a request header line to the server. |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 930 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 931 | For example: h.putheader('Accept', 'text/html') |
| 932 | """ |
| 933 | if self.__state != _CS_REQ_STARTED: |
| 934 | raise CannotSendHeader() |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 935 | |
Guido van Rossum | 98297ee | 2007-11-06 21:34:58 +0000 | [diff] [blame] | 936 | if hasattr(header, 'encode'): |
| 937 | header = header.encode('ascii') |
Benjamin Peterson | f608c61 | 2008-11-16 18:33:53 +0000 | [diff] [blame] | 938 | values = list(values) |
| 939 | for i, one_value in enumerate(values): |
| 940 | if hasattr(one_value, 'encode'): |
| 941 | values[i] = one_value.encode('ascii') |
Senthil Kumaran | 9730456 | 2010-10-03 18:25:01 +0000 | [diff] [blame] | 942 | elif isinstance(one_value, int): |
| 943 | values[i] = str(one_value).encode('ascii') |
Benjamin Peterson | f608c61 | 2008-11-16 18:33:53 +0000 | [diff] [blame] | 944 | value = b'\r\n\t'.join(values) |
Guido van Rossum | 98297ee | 2007-11-06 21:34:58 +0000 | [diff] [blame] | 945 | header = header + b': ' + value |
| 946 | self._output(header) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 947 | |
Benjamin Peterson | 1742e40 | 2008-11-30 22:15:29 +0000 | [diff] [blame] | 948 | def endheaders(self, message_body=None): |
| 949 | """Indicate that the last header line has been sent to the server. |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 950 | |
Benjamin Peterson | 1742e40 | 2008-11-30 22:15:29 +0000 | [diff] [blame] | 951 | This method sends the request to the server. The optional |
| 952 | message_body argument can be used to pass message body |
| 953 | associated with the request. The message body will be sent in |
| 954 | the same packet as the message headers if possible. The |
| 955 | message_body should be a string. |
| 956 | """ |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 957 | if self.__state == _CS_REQ_STARTED: |
| 958 | self.__state = _CS_REQ_SENT |
| 959 | else: |
| 960 | raise CannotSendHeader() |
Benjamin Peterson | 1742e40 | 2008-11-30 22:15:29 +0000 | [diff] [blame] | 961 | self._send_output(message_body) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 962 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 963 | def request(self, method, url, body=None, headers={}): |
| 964 | """Send a complete request to the server.""" |
Jeremy Hylton | 636950f | 2009-03-28 04:34:21 +0000 | [diff] [blame] | 965 | self._send_request(method, url, body, headers) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 966 | |
Benjamin Peterson | 1742e40 | 2008-11-30 22:15:29 +0000 | [diff] [blame] | 967 | def _set_content_length(self, body): |
| 968 | # Set the content-length based on the body. |
| 969 | thelen = None |
| 970 | try: |
| 971 | thelen = str(len(body)) |
| 972 | except TypeError as te: |
| 973 | # If this is a file-like object, try to |
| 974 | # fstat its file descriptor |
Benjamin Peterson | 1742e40 | 2008-11-30 22:15:29 +0000 | [diff] [blame] | 975 | try: |
| 976 | thelen = str(os.fstat(body.fileno()).st_size) |
| 977 | except (AttributeError, OSError): |
| 978 | # Don't send a length if this failed |
| 979 | if self.debuglevel > 0: print("Cannot stat!!") |
| 980 | |
| 981 | if thelen is not None: |
| 982 | self.putheader('Content-Length', thelen) |
| 983 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 984 | def _send_request(self, method, url, body, headers): |
Jeremy Hylton | 636950f | 2009-03-28 04:34:21 +0000 | [diff] [blame] | 985 | # Honor explicitly requested Host: and Accept-Encoding: headers. |
Jeremy Hylton | 2c17825 | 2004-08-07 16:28:14 +0000 | [diff] [blame] | 986 | header_names = dict.fromkeys([k.lower() for k in headers]) |
| 987 | skips = {} |
| 988 | if 'host' in header_names: |
| 989 | skips['skip_host'] = 1 |
| 990 | if 'accept-encoding' in header_names: |
| 991 | skips['skip_accept_encoding'] = 1 |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 992 | |
Jeremy Hylton | 2c17825 | 2004-08-07 16:28:14 +0000 | [diff] [blame] | 993 | self.putrequest(method, url, **skips) |
| 994 | |
| 995 | if body and ('content-length' not in header_names): |
Benjamin Peterson | 1742e40 | 2008-11-30 22:15:29 +0000 | [diff] [blame] | 996 | self._set_content_length(body) |
Guido van Rossum | cc2b016 | 2007-02-11 06:12:03 +0000 | [diff] [blame] | 997 | for hdr, value in headers.items(): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 998 | self.putheader(hdr, value) |
Jeremy Hylton | ef9f48e | 2009-03-26 22:04:05 +0000 | [diff] [blame] | 999 | if isinstance(body, str): |
Jeremy Hylton | 236654b | 2009-03-27 20:24:34 +0000 | [diff] [blame] | 1000 | # RFC 2616 Section 3.7.1 says that text default has a |
| 1001 | # default charset of iso-8859-1. |
| 1002 | body = body.encode('iso-8859-1') |
Jeremy Hylton | ef9f48e | 2009-03-26 22:04:05 +0000 | [diff] [blame] | 1003 | self.endheaders(body) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1004 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1005 | def getresponse(self): |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 1006 | """Get the response from the server. |
| 1007 | |
| 1008 | If the HTTPConnection is in the correct state, returns an |
| 1009 | instance of HTTPResponse or of whatever object is returned by |
| 1010 | class the response_class variable. |
| 1011 | |
| 1012 | If a request has not been sent or if a previous response has |
| 1013 | not be handled, ResponseNotReady is raised. If the HTTP |
| 1014 | response indicates that the connection should be closed, then |
| 1015 | it will be closed before the response is returned. When the |
| 1016 | connection is closed, the underlying socket is closed. |
| 1017 | """ |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1018 | |
Greg Stein | 616a58d | 2003-06-24 06:35:19 +0000 | [diff] [blame] | 1019 | # if a prior response has been completed, then forget about it. |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1020 | if self.__response and self.__response.isclosed(): |
| 1021 | self.__response = None |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1022 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1023 | # if a prior response exists, then it must be completed (otherwise, we |
| 1024 | # cannot read this response's header to determine the connection-close |
| 1025 | # behavior) |
| 1026 | # |
| 1027 | # note: if a prior response existed, but was connection-close, then the |
| 1028 | # socket and response were made independent of this HTTPConnection |
| 1029 | # object since a new request requires that we open a whole new |
| 1030 | # connection |
| 1031 | # |
| 1032 | # this means the prior response had one of two states: |
| 1033 | # 1) will_close: this connection was reset and the prior socket and |
| 1034 | # response operate independently |
| 1035 | # 2) persistent: the response was retained and we await its |
| 1036 | # isclosed() status to become true. |
| 1037 | # |
| 1038 | if self.__state != _CS_REQ_SENT or self.__response: |
Jeremy Hylton | 6c5e28c | 2009-03-31 14:35:53 +0000 | [diff] [blame] | 1039 | raise ResponseNotReady(self.__state) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1040 | |
Jeremy Hylton | 30f8674 | 2000-09-18 22:50:38 +0000 | [diff] [blame] | 1041 | if self.debuglevel > 0: |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 1042 | response = self.response_class(self.sock, self.debuglevel, |
Tim Peters | c2659cf | 2003-05-12 20:19:37 +0000 | [diff] [blame] | 1043 | strict=self.strict, |
Jeremy Hylton | c1b2cb9 | 2003-05-05 16:13:58 +0000 | [diff] [blame] | 1044 | method=self._method) |
Jeremy Hylton | 30f8674 | 2000-09-18 22:50:38 +0000 | [diff] [blame] | 1045 | else: |
Jeremy Hylton | c1b2cb9 | 2003-05-05 16:13:58 +0000 | [diff] [blame] | 1046 | response = self.response_class(self.sock, strict=self.strict, |
| 1047 | method=self._method) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1048 | |
Jeremy Hylton | 39c0380 | 2002-07-12 14:04:09 +0000 | [diff] [blame] | 1049 | response.begin() |
Jeremy Hylton | be4fcf1 | 2002-06-28 22:38:01 +0000 | [diff] [blame] | 1050 | assert response.will_close != _UNKNOWN |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1051 | self.__state = _CS_IDLE |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1052 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1053 | if response.will_close: |
Guido van Rossum | d8faa36 | 2007-04-27 19:54:29 +0000 | [diff] [blame] | 1054 | # this effectively passes the connection to the response |
| 1055 | self.close() |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1056 | else: |
| 1057 | # remember this, so we can tell when it is complete |
| 1058 | self.__response = response |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1059 | |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1060 | return response |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1061 | |
Thomas Wouters | 47b49bf | 2007-08-30 22:15:33 +0000 | [diff] [blame] | 1062 | try: |
| 1063 | import ssl |
| 1064 | except ImportError: |
| 1065 | pass |
| 1066 | else: |
| 1067 | class HTTPSConnection(HTTPConnection): |
| 1068 | "This class allows communication via SSL." |
Jeremy Hylton | 29d27ac | 2002-07-09 21:22:36 +0000 | [diff] [blame] | 1069 | |
Thomas Wouters | 47b49bf | 2007-08-30 22:15:33 +0000 | [diff] [blame] | 1070 | default_port = HTTPS_PORT |
Jeremy Hylton | 29d27ac | 2002-07-09 21:22:36 +0000 | [diff] [blame] | 1071 | |
Thomas Wouters | 47b49bf | 2007-08-30 22:15:33 +0000 | [diff] [blame] | 1072 | def __init__(self, host, port=None, key_file=None, cert_file=None, |
Georg Brandl | f78e02b | 2008-06-10 17:40:04 +0000 | [diff] [blame] | 1073 | strict=None, timeout=socket._GLOBAL_DEFAULT_TIMEOUT): |
Thomas Wouters | 47b49bf | 2007-08-30 22:15:33 +0000 | [diff] [blame] | 1074 | HTTPConnection.__init__(self, host, port, strict, timeout) |
| 1075 | self.key_file = key_file |
| 1076 | self.cert_file = cert_file |
Jeremy Hylton | 29d27ac | 2002-07-09 21:22:36 +0000 | [diff] [blame] | 1077 | |
Thomas Wouters | 47b49bf | 2007-08-30 22:15:33 +0000 | [diff] [blame] | 1078 | def connect(self): |
| 1079 | "Connect to a host on a given (SSL) port." |
Jeremy Hylton | 29d27ac | 2002-07-09 21:22:36 +0000 | [diff] [blame] | 1080 | |
Jeremy Hylton | 636950f | 2009-03-28 04:34:21 +0000 | [diff] [blame] | 1081 | sock = socket.create_connection((self.host, self.port), |
| 1082 | self.timeout) |
Senthil Kumaran | 0ac1f83 | 2009-07-26 12:39:47 +0000 | [diff] [blame] | 1083 | |
| 1084 | if self._tunnel_host: |
| 1085 | self.sock = sock |
| 1086 | self._tunnel() |
| 1087 | |
Thomas Wouters | 1b7f891 | 2007-09-19 03:06:30 +0000 | [diff] [blame] | 1088 | self.sock = ssl.wrap_socket(sock, self.key_file, self.cert_file) |
Jeremy Hylton | 29d27ac | 2002-07-09 21:22:36 +0000 | [diff] [blame] | 1089 | |
Jeremy Hylton | 29d27ac | 2002-07-09 21:22:36 +0000 | [diff] [blame] | 1090 | |
Thomas Wouters | 47b49bf | 2007-08-30 22:15:33 +0000 | [diff] [blame] | 1091 | def FakeSocket (sock, sslobj): |
Thomas Wouters | 89d996e | 2007-09-08 17:39:28 +0000 | [diff] [blame] | 1092 | warnings.warn("FakeSocket is deprecated, and won't be in 3.x. " + |
Thomas Wouters | 1b7f891 | 2007-09-19 03:06:30 +0000 | [diff] [blame] | 1093 | "Use the result of ssl.wrap_socket() directly instead.", |
Thomas Wouters | 89d996e | 2007-09-08 17:39:28 +0000 | [diff] [blame] | 1094 | DeprecationWarning, stacklevel=2) |
Thomas Wouters | 47b49bf | 2007-08-30 22:15:33 +0000 | [diff] [blame] | 1095 | return sslobj |
Jeremy Hylton | 29d27ac | 2002-07-09 21:22:36 +0000 | [diff] [blame] | 1096 | |
Thomas Wouters | 47b49bf | 2007-08-30 22:15:33 +0000 | [diff] [blame] | 1097 | __all__.append("HTTPSConnection") |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1098 | |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1099 | class HTTPException(Exception): |
Jeremy Hylton | 12f4f35 | 2002-07-06 18:55:01 +0000 | [diff] [blame] | 1100 | # Subclasses that define an __init__ must call Exception.__init__ |
| 1101 | # or define self.args. Otherwise, str() will fail. |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1102 | pass |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1103 | |
| 1104 | class NotConnected(HTTPException): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1105 | pass |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1106 | |
Skip Montanaro | 9d38997 | 2002-03-24 16:53:50 +0000 | [diff] [blame] | 1107 | class InvalidURL(HTTPException): |
| 1108 | pass |
| 1109 | |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1110 | class UnknownProtocol(HTTPException): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1111 | def __init__(self, version): |
Jeremy Hylton | 12f4f35 | 2002-07-06 18:55:01 +0000 | [diff] [blame] | 1112 | self.args = version, |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1113 | self.version = version |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1114 | |
| 1115 | class UnknownTransferEncoding(HTTPException): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1116 | pass |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1117 | |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1118 | class UnimplementedFileMode(HTTPException): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1119 | pass |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1120 | |
| 1121 | class IncompleteRead(HTTPException): |
Benjamin Peterson | 6accb98 | 2009-03-02 22:50:25 +0000 | [diff] [blame] | 1122 | def __init__(self, partial, expected=None): |
Jeremy Hylton | 12f4f35 | 2002-07-06 18:55:01 +0000 | [diff] [blame] | 1123 | self.args = partial, |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1124 | self.partial = partial |
Benjamin Peterson | 6accb98 | 2009-03-02 22:50:25 +0000 | [diff] [blame] | 1125 | self.expected = expected |
| 1126 | def __repr__(self): |
| 1127 | if self.expected is not None: |
| 1128 | e = ', %i more expected' % self.expected |
| 1129 | else: |
| 1130 | e = '' |
| 1131 | return 'IncompleteRead(%i bytes read%s)' % (len(self.partial), e) |
| 1132 | def __str__(self): |
| 1133 | return repr(self) |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1134 | |
| 1135 | class ImproperConnectionState(HTTPException): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1136 | pass |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1137 | |
| 1138 | class CannotSendRequest(ImproperConnectionState): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1139 | pass |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1140 | |
| 1141 | class CannotSendHeader(ImproperConnectionState): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1142 | pass |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1143 | |
| 1144 | class ResponseNotReady(ImproperConnectionState): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1145 | pass |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1146 | |
| 1147 | class BadStatusLine(HTTPException): |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1148 | def __init__(self, line): |
Jeremy Hylton | 12f4f35 | 2002-07-06 18:55:01 +0000 | [diff] [blame] | 1149 | self.args = line, |
Greg Stein | dd6eefb | 2000-07-18 09:09:48 +0000 | [diff] [blame] | 1150 | self.line = line |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1151 | |
Antoine Pitrou | ff1bbba | 2010-12-18 18:04:38 +0000 | [diff] [blame] | 1152 | class LineTooLong(HTTPException): |
| 1153 | def __init__(self, line_type): |
| 1154 | HTTPException.__init__(self, "got more than %d bytes when reading %s" |
| 1155 | % (_MAXLINE, line_type)) |
| 1156 | |
Greg Stein | 5e0fa40 | 2000-06-26 08:28:01 +0000 | [diff] [blame] | 1157 | # for backwards compatibility |
| 1158 | error = HTTPException |
| 1159 | |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 1160 | class LineAndFileWrapper: |
| 1161 | """A limited file-like object for HTTP/0.9 responses.""" |
| 1162 | |
| 1163 | # The status-line parsing code calls readline(), which normally |
| 1164 | # get the HTTP status line. For a 0.9 response, however, this is |
| 1165 | # actually the first line of the body! Clients need to get a |
| 1166 | # readable file object that contains that line. |
| 1167 | |
| 1168 | def __init__(self, line, file): |
| 1169 | self._line = line |
| 1170 | self._file = file |
| 1171 | self._line_consumed = 0 |
| 1172 | self._line_offset = 0 |
| 1173 | self._line_left = len(line) |
| 1174 | |
| 1175 | def __getattr__(self, attr): |
| 1176 | return getattr(self._file, attr) |
| 1177 | |
| 1178 | def _done(self): |
| 1179 | # called when the last byte is read from the line. After the |
| 1180 | # call, all read methods are delegated to the underlying file |
Skip Montanaro | 74b9a7a | 2003-02-25 17:48:15 +0000 | [diff] [blame] | 1181 | # object. |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 1182 | self._line_consumed = 1 |
| 1183 | self.read = self._file.read |
| 1184 | self.readline = self._file.readline |
| 1185 | self.readlines = self._file.readlines |
| 1186 | |
| 1187 | def read(self, amt=None): |
Hye-Shik Chang | 39aef79 | 2004-06-05 13:30:56 +0000 | [diff] [blame] | 1188 | if self._line_consumed: |
| 1189 | return self._file.read(amt) |
| 1190 | assert self._line_left |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 1191 | if amt is None or amt > self._line_left: |
| 1192 | s = self._line[self._line_offset:] |
| 1193 | self._done() |
| 1194 | if amt is None: |
| 1195 | return s + self._file.read() |
| 1196 | else: |
Tim Peters | c411dba | 2002-07-16 21:35:23 +0000 | [diff] [blame] | 1197 | return s + self._file.read(amt - len(s)) |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 1198 | else: |
| 1199 | assert amt <= self._line_left |
| 1200 | i = self._line_offset |
| 1201 | j = i + amt |
| 1202 | s = self._line[i:j] |
| 1203 | self._line_offset = j |
| 1204 | self._line_left -= amt |
| 1205 | if self._line_left == 0: |
| 1206 | self._done() |
| 1207 | return s |
Tim Peters | c411dba | 2002-07-16 21:35:23 +0000 | [diff] [blame] | 1208 | |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 1209 | def readline(self): |
Hye-Shik Chang | 39aef79 | 2004-06-05 13:30:56 +0000 | [diff] [blame] | 1210 | if self._line_consumed: |
| 1211 | return self._file.readline() |
| 1212 | assert self._line_left |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 1213 | s = self._line[self._line_offset:] |
| 1214 | self._done() |
| 1215 | return s |
| 1216 | |
| 1217 | def readlines(self, size=None): |
Hye-Shik Chang | 39aef79 | 2004-06-05 13:30:56 +0000 | [diff] [blame] | 1218 | if self._line_consumed: |
| 1219 | return self._file.readlines(size) |
| 1220 | assert self._line_left |
Jeremy Hylton | d46aa37 | 2002-07-06 18:48:07 +0000 | [diff] [blame] | 1221 | L = [self._line[self._line_offset:]] |
| 1222 | self._done() |
| 1223 | if size is None: |
| 1224 | return L + self._file.readlines() |
| 1225 | else: |
| 1226 | return L + self._file.readlines(size) |