1 """HTTP server base class. 2 3 Note: the class in this module doesn't implement any HTTP request; see 4 SimpleHTTPServer for simple implementations of GET, HEAD and POST 5 (including CGI scripts). It does, however, optionally implement HTTP/1.1 6 persistent connections, as of version 0.3. 7 8 Contents: 9 10 - BaseHTTPRequestHandler: HTTP request handler base class 11 - test: test function 12 13 XXX To do: 14 15 - log requests even later (to capture byte count) 16 - log user-agent header and other interesting goodies 17 - send error log to separate file 18 """ 19 20 21 # See also: 22 # 23 # HTTP Working Group T. Berners-Lee 24 # INTERNET-DRAFT R. T. Fielding 25 # <draft-ietf-http-v10-spec-00.txt> H. Frystyk Nielsen 26 # Expires September 8, 1995 March 8, 1995 27 # 28 # URL: http://www.ics.uci.edu/pub/ietf/http/draft-ietf-http-v10-spec-00.txt 29 # 30 # and 31 # 32 # Network Working Group R. Fielding 33 # Request for Comments: 2616 et al 34 # Obsoletes: 2068 June 1999 35 # Category: Standards Track 36 # 37 # URL: http://www.faqs.org/rfcs/rfc2616.html 38 39 # Log files 40 # --------- 41 # 42 # Here's a quote from the NCSA httpd docs about log file format. 43 # 44 # | The logfile format is as follows. Each line consists of: 45 # | 46 # | host rfc931 authuser [DD/Mon/YYYY:hh:mm:ss] "request" ddd bbbb 47 # | 48 # | host: Either the DNS name or the IP number of the remote client 49 # | rfc931: Any information returned by identd for this person, 50 # | - otherwise. 51 # | authuser: If user sent a userid for authentication, the user name, 52 # | - otherwise. 53 # | DD: Day 54 # | Mon: Month (calendar name) 55 # | YYYY: Year 56 # | hh: hour (24-hour format, the machine's timezone) 57 # | mm: minutes 58 # | ss: seconds 59 # | request: The first line of the HTTP request as sent by the client. 60 # | ddd: the status code returned by the server, - if not available. 61 # | bbbb: the total number of bytes sent, 62 # | *not including the HTTP/1.0 header*, - if not available 63 # | 64 # | You can determine the name of the file accessed through request. 65 # 66 # (Actually, the latter is only true if you know the server configuration 67 # at the time the request was made!) 68 69 __version__ = "0.3" 70 71 __all__ = ["HTTPServer", "BaseHTTPRequestHandler"] 72 73 import sys 74 import time 75 import socket # For gethostbyaddr() 76 from warnings import filterwarnings, catch_warnings 77 with catch_warnings(): 78 if sys.py3kwarning: 79 filterwarnings("ignore", ".*mimetools has been removed", 80 DeprecationWarning) 81 import mimetools 82 import SocketServer 83 84 # Default error message template 85 DEFAULT_ERROR_MESSAGE = """\ 86 <head> 87 <title>Error response</title> 88 </head> 89 <body> 90 <h1>Error response</h1> 91 <p>Error code %(code)d. 92 <p>Message: %(message)s. 93 <p>Error code explanation: %(code)s = %(explain)s. 94 </body> 95 """ 96 97 DEFAULT_ERROR_CONTENT_TYPE = "text/html" 98 99 def _quote_html(html): 100 return html.replace("&", "&").replace("<", "<").replace(">", ">") 101 102 class HTTPServer(SocketServer.TCPServer): 103 104 allow_reuse_address = 1 # Seems to make sense in testing environment 105 106 def server_bind(self): 107 """Override server_bind to store the server name.""" 108 SocketServer.TCPServer.server_bind(self) 109 host, port = self.socket.getsockname()[:2] 110 self.server_name = socket.getfqdn(host) 111 self.server_port = port 112 113 114 class BaseHTTPRequestHandler(SocketServer.StreamRequestHandler): 115 116 """HTTP request handler base class. 117 118 The following explanation of HTTP serves to guide you through the 119 code as well as to expose any misunderstandings I may have about 120 HTTP (so you don't need to read the code to figure out I'm wrong 121 :-). 122 123 HTTP (HyperText Transfer Protocol) is an extensible protocol on 124 top of a reliable stream transport (e.g. TCP/IP). The protocol 125 recognizes three parts to a request: 126 127 1. One line identifying the request type and path 128 2. An optional set of RFC-822-style headers 129 3. An optional data part 130 131 The headers and data are separated by a blank line. 132 133 The first line of the request has the form 134 135 <command> <path> <version> 136 137 where <command> is a (case-sensitive) keyword such as GET or POST, 138 <path> is a string containing path information for the request, 139 and <version> should be the string "HTTP/1.0" or "HTTP/1.1". 140 <path> is encoded using the URL encoding scheme (using %xx to signify 141 the ASCII character with hex code xx). 142 143 The specification specifies that lines are separated by CRLF but 144 for compatibility with the widest range of clients recommends 145 servers also handle LF. Similarly, whitespace in the request line 146 is treated sensibly (allowing multiple spaces between components 147 and allowing trailing whitespace). 148 149 Similarly, for output, lines ought to be separated by CRLF pairs 150 but most clients grok LF characters just fine. 151 152 If the first line of the request has the form 153 154 <command> <path> 155 156 (i.e. <version> is left out) then this is assumed to be an HTTP 157 0.9 request; this form has no optional headers and data part and 158 the reply consists of just the data. 159 160 The reply form of the HTTP 1.x protocol again has three parts: 161 162 1. One line giving the response code 163 2. An optional set of RFC-822-style headers 164 3. The data 165 166 Again, the headers and data are separated by a blank line. 167 168 The response code line has the form 169 170 <version> <responsecode> <responsestring> 171 172 where <version> is the protocol version ("HTTP/1.0" or "HTTP/1.1"), 173 <responsecode> is a 3-digit response code indicating success or 174 failure of the request, and <responsestring> is an optional 175 human-readable string explaining what the response code means. 176 177 This server parses the request and the headers, and then calls a 178 function specific to the request type (<command>). Specifically, 179 a request SPAM will be handled by a method do_SPAM(). If no 180 such method exists the server sends an error response to the 181 client. If it exists, it is called with no arguments: 182 183 do_SPAM() 184 185 Note that the request name is case sensitive (i.e. SPAM and spam 186 are different requests). 187 188 The various request details are stored in instance variables: 189 190 - client_address is the client IP address in the form (host, 191 port); 192 193 - command, path and version are the broken-down request line; 194 195 - headers is an instance of mimetools.Message (or a derived 196 class) containing the header information; 197 198 - rfile is a file object open for reading positioned at the 199 start of the optional input data part; 200 201 - wfile is a file object open for writing. 202 203 IT IS IMPORTANT TO ADHERE TO THE PROTOCOL FOR WRITING! 204 205 The first thing to be written must be the response line. Then 206 follow 0 or more header lines, then a blank line, and then the 207 actual data (if any). The meaning of the header lines depends on 208 the command executed by the server; in most cases, when data is 209 returned, there should be at least one header line of the form 210 211 Content-type: <type>/<subtype> 212 213 where <type> and <subtype> should be registered MIME types, 214 e.g. "text/html" or "text/plain". 215 216 """ 217 218 # The Python system version, truncated to its first component. 219 sys_version = "Python/" + sys.version.split()[0] 220 221 # The server software version. You may want to override this. 222 # The format is multiple whitespace-separated strings, 223 # where each string is of the form name[/version]. 224 server_version = "BaseHTTP/" + __version__ 225 226 # The default request version. This only affects responses up until 227 # the point where the request line is parsed, so it mainly decides what 228 # the client gets back when sending a malformed request line. 229 # Most web servers default to HTTP 0.9, i.e. don't send a status line. 230 default_request_version = "HTTP/0.9" 231 232 def parse_request(self): 233 """Parse a request (internal). 234 235 The request should be stored in self.raw_requestline; the results 236 are in self.command, self.path, self.request_version and 237 self.headers. 238 239 Return True for success, False for failure; on failure, an 240 error is sent back. 241 242 """ 243 self.command = None # set in case of error on the first line 244 self.request_version = version = self.default_request_version 245 self.close_connection = 1 246 requestline = self.raw_requestline 247 requestline = requestline.rstrip('\r\n') 248 self.requestline = requestline 249 words = requestline.split() 250 if len(words) == 3: 251 command, path, version = words 252 if version[:5] != 'HTTP/': 253 self.send_error(400, "Bad request version (%r)" % version) 254 return False 255 try: 256 base_version_number = version.split('/', 1)[1] 257 version_number = base_version_number.split(".") 258 # RFC 2145 section 3.1 says there can be only one "." and 259 # - major and minor numbers MUST be treated as 260 # separate integers; 261 # - HTTP/2.4 is a lower version than HTTP/2.13, which in 262 # turn is lower than HTTP/12.3; 263 # - Leading zeros MUST be ignored by recipients. 264 if len(version_number) != 2: 265 raise ValueError 266 version_number = int(version_number[0]), int(version_number[1]) 267 except (ValueError, IndexError): 268 self.send_error(400, "Bad request version (%r)" % version) 269 return False 270 if version_number >= (1, 1) and self.protocol_version >= "HTTP/1.1": 271 self.close_connection = 0 272 if version_number >= (2, 0): 273 self.send_error(505, 274 "Invalid HTTP Version (%s)" % base_version_number) 275 return False 276 elif len(words) == 2: 277 command, path = words 278 self.close_connection = 1 279 if command != 'GET': 280 self.send_error(400, 281 "Bad HTTP/0.9 request type (%r)" % command) 282 return False 283 elif not words: 284 return False 285 else: 286 self.send_error(400, "Bad request syntax (%r)" % requestline) 287 return False 288 self.command, self.path, self.request_version = command, path, version 289 290 # Examine the headers and look for a Connection directive 291 self.headers = self.MessageClass(self.rfile, 0) 292 293 conntype = self.headers.get('Connection', "") 294 if conntype.lower() == 'close': 295 self.close_connection = 1 296 elif (conntype.lower() == 'keep-alive' and 297 self.protocol_version >= "HTTP/1.1"): 298 self.close_connection = 0 299 return True 300 301 def handle_one_request(self): 302 """Handle a single HTTP request. 303 304 You normally don't need to override this method; see the class 305 __doc__ string for information on how to handle specific HTTP 306 commands such as GET and POST. 307 308 """ 309 try: 310 self.raw_requestline = self.rfile.readline(65537) 311 if len(self.raw_requestline) > 65536: 312 self.requestline = '' 313 self.request_version = '' 314 self.command = '' 315 self.send_error(414) 316 return 317 if not self.raw_requestline: 318 self.close_connection = 1 319 return 320 if not self.parse_request(): 321 # An error code has been sent, just exit 322 return 323 mname = 'do_' + self.command 324 if not hasattr(self, mname): 325 self.send_error(501, "Unsupported method (%r)" % self.command) 326 return 327 method = getattr(self, mname) 328 method() 329 self.wfile.flush() #actually send the response if not already done. 330 except socket.timeout, e: 331 #a read or a write timed out. Discard this connection 332 self.log_error("Request timed out: %r", e) 333 self.close_connection = 1 334 return 335 336 def handle(self): 337 """Handle multiple requests if necessary.""" 338 self.close_connection = 1 339 340 self.handle_one_request() 341 while not self.close_connection: 342 self.handle_one_request() 343 344 def send_error(self, code, message=None): 345 """Send and log an error reply. 346 347 Arguments are the error code, and a detailed message. 348 The detailed message defaults to the short entry matching the 349 response code. 350 351 This sends an error response (so it must be called before any 352 output has been generated), logs the error, and finally sends 353 a piece of HTML explaining the error to the user. 354 355 """ 356 357 try: 358 short, long = self.responses[code] 359 except KeyError: 360 short, long = '???', '???' 361 if message is None: 362 message = short 363 explain = long 364 self.log_error("code %d, message %s", code, message) 365 self.send_response(code, message) 366 self.send_header('Connection', 'close') 367 368 # Message body is omitted for cases described in: 369 # - RFC7230: 3.3. 1xx, 204(No Content), 304(Not Modified) 370 # - RFC7231: 6.3.6. 205(Reset Content) 371 content = None 372 if code >= 200 and code not in (204, 205, 304): 373 # HTML encode to prevent Cross Site Scripting attacks 374 # (see bug #1100201) 375 content = (self.error_message_format % { 376 'code': code, 377 'message': _quote_html(message), 378 'explain': explain 379 }) 380 self.send_header("Content-Type", self.error_content_type) 381 self.end_headers() 382 383 if self.command != 'HEAD' and content: 384 self.wfile.write(content) 385 386 error_message_format = DEFAULT_ERROR_MESSAGE 387 error_content_type = DEFAULT_ERROR_CONTENT_TYPE 388 389 def send_response(self, code, message=None): 390 """Send the response header and log the response code. 391 392 Also send two standard headers with the server software 393 version and the current date. 394 395 """ 396 self.log_request(code) 397 if message is None: 398 if code in self.responses: 399 message = self.responses[code][0] 400 else: 401 message = '' 402 if self.request_version != 'HTTP/0.9': 403 self.wfile.write("%s %d %s\r\n" % 404 (self.protocol_version, code, message)) 405 # print (self.protocol_version, code, message) 406 self.send_header('Server', self.version_string()) 407 self.send_header('Date', self.date_time_string()) 408 409 def send_header(self, keyword, value): 410 """Send a MIME header.""" 411 if self.request_version != 'HTTP/0.9': 412 self.wfile.write("%s: %s\r\n" % (keyword, value)) 413 414 if keyword.lower() == 'connection': 415 if value.lower() == 'close': 416 self.close_connection = 1 417 elif value.lower() == 'keep-alive': 418 self.close_connection = 0 419 420 def end_headers(self): 421 """Send the blank line ending the MIME headers.""" 422 if self.request_version != 'HTTP/0.9': 423 self.wfile.write("\r\n") 424 425 def log_request(self, code='-', size='-'): 426 """Log an accepted request. 427 428 This is called by send_response(). 429 430 """ 431 432 self.log_message('"%s" %s %s', 433 self.requestline, str(code), str(size)) 434 435 def log_error(self, format, *args): 436 """Log an error. 437 438 This is called when a request cannot be fulfilled. By 439 default it passes the message on to log_message(). 440 441 Arguments are the same as for log_message(). 442 443 XXX This should go to the separate error log. 444 445 """ 446 447 self.log_message(format, *args) 448 449 def log_message(self, format, *args): 450 """Log an arbitrary message. 451 452 This is used by all other logging functions. Override 453 it if you have specific logging wishes. 454 455 The first argument, FORMAT, is a format string for the 456 message to be logged. If the format string contains 457 any % escapes requiring parameters, they should be 458 specified as subsequent arguments (it's just like 459 printf!). 460 461 The client ip address and current date/time are prefixed to every 462 message. 463 464 """ 465 466 sys.stderr.write("%s - - [%s] %s\n" % 467 (self.client_address[0], 468 self.log_date_time_string(), 469 format%args)) 470 471 def version_string(self): 472 """Return the server software version string.""" 473 return self.server_version + ' ' + self.sys_version 474 475 def date_time_string(self, timestamp=None): 476 """Return the current date and time formatted for a message header.""" 477 if timestamp is None: 478 timestamp = time.time() 479 year, month, day, hh, mm, ss, wd, y, z = time.gmtime(timestamp) 480 s = "%s, %02d %3s %4d %02d:%02d:%02d GMT" % ( 481 self.weekdayname[wd], 482 day, self.monthname[month], year, 483 hh, mm, ss) 484 return s 485 486 def log_date_time_string(self): 487 """Return the current time formatted for logging.""" 488 now = time.time() 489 year, month, day, hh, mm, ss, x, y, z = time.localtime(now) 490 s = "%02d/%3s/%04d %02d:%02d:%02d" % ( 491 day, self.monthname[month], year, hh, mm, ss) 492 return s 493 494 weekdayname = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun'] 495 496 monthname = [None, 497 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 498 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec'] 499 500 def address_string(self): 501 """Return the client address formatted for logging. 502 503 This version looks up the full hostname using gethostbyaddr(), 504 and tries to find a name that contains at least one dot. 505 506 """ 507 508 host, port = self.client_address[:2] 509 return socket.getfqdn(host) 510 511 # Essentially static class variables 512 513 # The version of the HTTP protocol we support. 514 # Set this to HTTP/1.1 to enable automatic keepalive 515 protocol_version = "HTTP/1.0" 516 517 # The Message-like class used to parse headers 518 MessageClass = mimetools.Message 519 520 # Table mapping response codes to messages; entries have the 521 # form {code: (shortmessage, longmessage)}. 522 # See RFC 2616. 523 responses = { 524 100: ('Continue', 'Request received, please continue'), 525 101: ('Switching Protocols', 526 'Switching to new protocol; obey Upgrade header'), 527 528 200: ('OK', 'Request fulfilled, document follows'), 529 201: ('Created', 'Document created, URL follows'), 530 202: ('Accepted', 531 'Request accepted, processing continues off-line'), 532 203: ('Non-Authoritative Information', 'Request fulfilled from cache'), 533 204: ('No Content', 'Request fulfilled, nothing follows'), 534 205: ('Reset Content', 'Clear input form for further input.'), 535 206: ('Partial Content', 'Partial content follows.'), 536 537 300: ('Multiple Choices', 538 'Object has several resources -- see URI list'), 539 301: ('Moved Permanently', 'Object moved permanently -- see URI list'), 540 302: ('Found', 'Object moved temporarily -- see URI list'), 541 303: ('See Other', 'Object moved -- see Method and URL list'), 542 304: ('Not Modified', 543 'Document has not changed since given time'), 544 305: ('Use Proxy', 545 'You must use proxy specified in Location to access this ' 546 'resource.'), 547 307: ('Temporary Redirect', 548 'Object moved temporarily -- see URI list'), 549 550 400: ('Bad Request', 551 'Bad request syntax or unsupported method'), 552 401: ('Unauthorized', 553 'No permission -- see authorization schemes'), 554 402: ('Payment Required', 555 'No payment -- see charging schemes'), 556 403: ('Forbidden', 557 'Request forbidden -- authorization will not help'), 558 404: ('Not Found', 'Nothing matches the given URI'), 559 405: ('Method Not Allowed', 560 'Specified method is invalid for this resource.'), 561 406: ('Not Acceptable', 'URI not available in preferred format.'), 562 407: ('Proxy Authentication Required', 'You must authenticate with ' 563 'this proxy before proceeding.'), 564 408: ('Request Timeout', 'Request timed out; try again later.'), 565 409: ('Conflict', 'Request conflict.'), 566 410: ('Gone', 567 'URI no longer exists and has been permanently removed.'), 568 411: ('Length Required', 'Client must specify Content-Length.'), 569 412: ('Precondition Failed', 'Precondition in headers is false.'), 570 413: ('Request Entity Too Large', 'Entity is too large.'), 571 414: ('Request-URI Too Long', 'URI is too long.'), 572 415: ('Unsupported Media Type', 'Entity body in unsupported format.'), 573 416: ('Requested Range Not Satisfiable', 574 'Cannot satisfy request range.'), 575 417: ('Expectation Failed', 576 'Expect condition could not be satisfied.'), 577 578 500: ('Internal Server Error', 'Server got itself in trouble'), 579 501: ('Not Implemented', 580 'Server does not support this operation'), 581 502: ('Bad Gateway', 'Invalid responses from another server/proxy.'), 582 503: ('Service Unavailable', 583 'The server cannot process the request due to a high load'), 584 504: ('Gateway Timeout', 585 'The gateway server did not receive a timely response'), 586 505: ('HTTP Version Not Supported', 'Cannot fulfill request.'), 587 } 588 589 590 def test(HandlerClass = BaseHTTPRequestHandler, 591 ServerClass = HTTPServer, protocol="HTTP/1.0"): 592 """Test the HTTP request handler class. 593 594 This runs an HTTP server on port 8000 (or the first command line 595 argument). 596 597 """ 598 599 if sys.argv[1:]: 600 port = int(sys.argv[1]) 601 else: 602 port = 8000 603 server_address = ('', port) 604 605 HandlerClass.protocol_version = protocol 606 httpd = ServerClass(server_address, HandlerClass) 607 608 sa = httpd.socket.getsockname() 609 print "Serving HTTP on", sa[0], "port", sa[1], "..." 610 httpd.serve_forever() 611 612 613 if __name__ == '__main__': 614 test() 615
