Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / python / cpython3 / c5895dcc15d442ac45567b67a68ef76db8c20cd9 / . / Lib / http / server.py
blob: f1538f40f9a43f20b99ad0c921bc3f49d9cc6f9e [file] [log] [blame]
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1"""HTTP server classes.
2
3Note: BaseHTTPRequestHandler doesn't implement any HTTP request; see
4SimpleHTTPRequestHandler for simple implementations of GET, HEAD and POST,
5and CGIHTTPRequestHandler for CGI scripts.
6
7It does, however, optionally implement HTTP/1.1 persistent connections,
8as of version 0.3.
9
10Notes on CGIHTTPRequestHandler
11------------------------------
12
13This class implements GET and POST requests to cgi-bin scripts.
14
15If the os.fork() function is not present (e.g. on Windows),
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]16subprocess.Popen() is used as a fallback, with slightly altered semantics.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]17
18In all cases, the implementation is intentionally naive -- all
19requests are executed synchronously.
20
21SECURITY WARNING: DON'T USE THIS CODE UNLESS YOU ARE INSIDE A FIREWALL
22-- it may execute arbitrary Python code or external programs.
23
24Note that status code 200 is sent prior to execution of a CGI script, so
25scripts cannot send other status codes such as 302 (redirect).
26
27XXX To do:
28
29- log requests even later (to capture byte count)
30- log user-agent header and other interesting goodies
31- send error log to separate file
32"""
33
34
35# See also:
36#
37# HTTP Working Group T. Berners-Lee
38# INTERNET-DRAFT R. T. Fielding
39# <draft-ietf-http-v10-spec-00.txt> H. Frystyk Nielsen
40# Expires September 8, 1995 March 8, 1995
41#
42# URL: http://www.ics.uci.edu/pub/ietf/http/draft-ietf-http-v10-spec-00.txt
43#
44# and
45#
46# Network Working Group R. Fielding
47# Request for Comments: 2616 et al
48# Obsoletes: 2068 June 1999
49# Category: Standards Track
50#
51# URL: http://www.faqs.org/rfcs/rfc2616.html
52
53# Log files
54# ---------
55#
56# Here's a quote from the NCSA httpd docs about log file format.
57#
58# | The logfile format is as follows. Each line consists of:
59# |
60# | host rfc931 authuser [DD/Mon/YYYY:hh:mm:ss] "request" ddd bbbb
61# |
62# | host: Either the DNS name or the IP number of the remote client
63# | rfc931: Any information returned by identd for this person,
64# | - otherwise.
65# | authuser: If user sent a userid for authentication, the user name,
66# | - otherwise.
67# | DD: Day
68# | Mon: Month (calendar name)
69# | YYYY: Year
70# | hh: hour (24-hour format, the machine's timezone)
71# | mm: minutes
72# | ss: seconds
73# | request: The first line of the HTTP request as sent by the client.
74# | ddd: the status code returned by the server, - if not available.
75# | bbbb: the total number of bytes sent,
76# | *not including the HTTP/1.0 header*, - if not available
77# |
78# | You can determine the name of the file accessed through request.
79#
80# (Actually, the latter is only true if you know the server configuration
81# at the time the request was made!)
82
83__version__ = "0.6"
84
85__all__ = ["HTTPServer", "BaseHTTPRequestHandler"]
86
Georg Brandl1f7fffb2010-10-15 15:57:45 +0000[diff] [blame]87import html
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]88import email.message
89import email.parser
Jeremy Hylton914ab452009-03-27 17:16:06 +0000[diff] [blame]90import http.client
91import io
92import mimetypes
93import os
94import posixpath
95import select
96import shutil
97import socket # For gethostbyaddr()
98import socketserver
99import sys
100import time
101import urllib.parse
Senthil Kumaran42713722010-10-03 17:55:45 +0000[diff] [blame]102import copy
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]103
104# Default error message template
105DEFAULT_ERROR_MESSAGE = """\
106<head>
107<title>Error response</title>
108</head>
109<body>
110<h1>Error response</h1>
111<p>Error code %(code)d.
112<p>Message: %(message)s.
113<p>Error code explanation: %(code)s = %(explain)s.
114</body>
115"""
116
117DEFAULT_ERROR_CONTENT_TYPE = "text/html;charset=utf-8"
118
119def _quote_html(html):
120 return html.replace("&", "&amp;").replace("<", "&lt;").replace(">", "&gt;")
121
122class HTTPServer(socketserver.TCPServer):
123
124 allow_reuse_address = 1 # Seems to make sense in testing environment
125
126 def server_bind(self):
127 """Override server_bind to store the server name."""
128 socketserver.TCPServer.server_bind(self)
129 host, port = self.socket.getsockname()[:2]
130 self.server_name = socket.getfqdn(host)
131 self.server_port = port
132
133
134class BaseHTTPRequestHandler(socketserver.StreamRequestHandler):
135
136 """HTTP request handler base class.
137
138 The following explanation of HTTP serves to guide you through the
139 code as well as to expose any misunderstandings I may have about
140 HTTP (so you don't need to read the code to figure out I'm wrong
141 :-).
142
143 HTTP (HyperText Transfer Protocol) is an extensible protocol on
144 top of a reliable stream transport (e.g. TCP/IP). The protocol
145 recognizes three parts to a request:
146
147 1. One line identifying the request type and path
148 2. An optional set of RFC-822-style headers
149 3. An optional data part
150
151 The headers and data are separated by a blank line.
152
153 The first line of the request has the form
154
155 <command> <path> <version>
156
157 where <command> is a (case-sensitive) keyword such as GET or POST,
158 <path> is a string containing path information for the request,
159 and <version> should be the string "HTTP/1.0" or "HTTP/1.1".
160 <path> is encoded using the URL encoding scheme (using %xx to signify
161 the ASCII character with hex code xx).
162
163 The specification specifies that lines are separated by CRLF but
164 for compatibility with the widest range of clients recommends
165 servers also handle LF. Similarly, whitespace in the request line
166 is treated sensibly (allowing multiple spaces between components
167 and allowing trailing whitespace).
168
169 Similarly, for output, lines ought to be separated by CRLF pairs
170 but most clients grok LF characters just fine.
171
172 If the first line of the request has the form
173
174 <command> <path>
175
176 (i.e. <version> is left out) then this is assumed to be an HTTP
177 0.9 request; this form has no optional headers and data part and
178 the reply consists of just the data.
179
180 The reply form of the HTTP 1.x protocol again has three parts:
181
182 1. One line giving the response code
183 2. An optional set of RFC-822-style headers
184 3. The data
185
186 Again, the headers and data are separated by a blank line.
187
188 The response code line has the form
189
190 <version> <responsecode> <responsestring>
191
192 where <version> is the protocol version ("HTTP/1.0" or "HTTP/1.1"),
193 <responsecode> is a 3-digit response code indicating success or
194 failure of the request, and <responsestring> is an optional
195 human-readable string explaining what the response code means.
196
197 This server parses the request and the headers, and then calls a
198 function specific to the request type (<command>). Specifically,
199 a request SPAM will be handled by a method do_SPAM(). If no
200 such method exists the server sends an error response to the
201 client. If it exists, it is called with no arguments:
202
203 do_SPAM()
204
205 Note that the request name is case sensitive (i.e. SPAM and spam
206 are different requests).
207
208 The various request details are stored in instance variables:
209
210 - client_address is the client IP address in the form (host,
211 port);
212
213 - command, path and version are the broken-down request line;
214
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]215 - headers is an instance of email.message.Message (or a derived
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]216 class) containing the header information;
217
218 - rfile is a file object open for reading positioned at the
219 start of the optional input data part;
220
221 - wfile is a file object open for writing.
222
223 IT IS IMPORTANT TO ADHERE TO THE PROTOCOL FOR WRITING!
224
225 The first thing to be written must be the response line. Then
226 follow 0 or more header lines, then a blank line, and then the
227 actual data (if any). The meaning of the header lines depends on
228 the command executed by the server; in most cases, when data is
229 returned, there should be at least one header line of the form
230
231 Content-type: <type>/<subtype>
232
233 where <type> and <subtype> should be registered MIME types,
234 e.g. "text/html" or "text/plain".
235
236 """
237
238 # The Python system version, truncated to its first component.
239 sys_version = "Python/" + sys.version.split()[0]
240
241 # The server software version. You may want to override this.
242 # The format is multiple whitespace-separated strings,
243 # where each string is of the form name[/version].
244 server_version = "BaseHTTP/" + __version__
245
246 error_message_format = DEFAULT_ERROR_MESSAGE
247 error_content_type = DEFAULT_ERROR_CONTENT_TYPE
248
249 # The default request version. This only affects responses up until
250 # the point where the request line is parsed, so it mainly decides what
251 # the client gets back when sending a malformed request line.
252 # Most web servers default to HTTP 0.9, i.e. don't send a status line.
253 default_request_version = "HTTP/0.9"
254
255 def parse_request(self):
256 """Parse a request (internal).
257
258 The request should be stored in self.raw_requestline; the results
259 are in self.command, self.path, self.request_version and
260 self.headers.
261
262 Return True for success, False for failure; on failure, an
263 error is sent back.
264
265 """
266 self.command = None # set in case of error on the first line
267 self.request_version = version = self.default_request_version
268 self.close_connection = 1
269 requestline = str(self.raw_requestline, 'iso-8859-1')
270 if requestline[-2:] == '\r\n':
271 requestline = requestline[:-2]
272 elif requestline[-1:] == '\n':
273 requestline = requestline[:-1]
274 self.requestline = requestline
275 words = requestline.split()
276 if len(words) == 3:
277 [command, path, version] = words
278 if version[:5] != 'HTTP/':
279 self.send_error(400, "Bad request version (%r)" % version)
280 return False
281 try:
282 base_version_number = version.split('/', 1)[1]
283 version_number = base_version_number.split(".")
284 # RFC 2145 section 3.1 says there can be only one "." and
285 # - major and minor numbers MUST be treated as
286 # separate integers;
287 # - HTTP/2.4 is a lower version than HTTP/2.13, which in
288 # turn is lower than HTTP/12.3;
289 # - Leading zeros MUST be ignored by recipients.
290 if len(version_number) != 2:
291 raise ValueError
292 version_number = int(version_number[0]), int(version_number[1])
293 except (ValueError, IndexError):
294 self.send_error(400, "Bad request version (%r)" % version)
295 return False
296 if version_number >= (1, 1) and self.protocol_version >= "HTTP/1.1":
297 self.close_connection = 0
298 if version_number >= (2, 0):
299 self.send_error(505,
300 "Invalid HTTP Version (%s)" % base_version_number)
301 return False
302 elif len(words) == 2:
303 [command, path] = words
304 self.close_connection = 1
305 if command != 'GET':
306 self.send_error(400,
307 "Bad HTTP/0.9 request type (%r)" % command)
308 return False
309 elif not words:
310 return False
311 else:
312 self.send_error(400, "Bad request syntax (%r)" % requestline)
313 return False
314 self.command, self.path, self.request_version = command, path, version
315
316 # Examine the headers and look for a Connection directive.
Jeremy Hylton98eb6c22009-03-27 18:31:36 +0000[diff] [blame]317 self.headers = http.client.parse_headers(self.rfile,
318 _class=self.MessageClass)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]319
320 conntype = self.headers.get('Connection', "")
321 if conntype.lower() == 'close':
322 self.close_connection = 1
323 elif (conntype.lower() == 'keep-alive' and
324 self.protocol_version >= "HTTP/1.1"):
325 self.close_connection = 0
Senthil Kumaran0f476d42010-09-30 06:09:18 +0000[diff] [blame]326 # Examine the headers and look for an Expect directive
327 expect = self.headers.get('Expect', "")
328 if (expect.lower() == "100-continue" and
329 self.protocol_version >= "HTTP/1.1" and
330 self.request_version >= "HTTP/1.1"):
331 if not self.handle_expect_100():
332 return False
333 return True
334
335 def handle_expect_100(self):
336 """Decide what to do with an "Expect: 100-continue" header.
337
338 If the client is expecting a 100 Continue response, we must
339 respond with either a 100 Continue or a final response before
340 waiting for the request body. The default is to always respond
341 with a 100 Continue. You can behave differently (for example,
342 reject unauthorized requests) by overriding this method.
343
344 This method should either return True (possibly after sending
345 a 100 Continue response) or send an error response and return
346 False.
347
348 """
349 self.send_response_only(100)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]350 return True
351
352 def handle_one_request(self):
353 """Handle a single HTTP request.
354
355 You normally don't need to override this method; see the class
356 __doc__ string for information on how to handle specific HTTP
357 commands such as GET and POST.
358
359 """
Kristján Valur Jónsson985fc6a2009-07-01 10:01:31 +0000[diff] [blame]360 try:
Antoine Pitrouc4924372010-12-16 16:48:36 +0000[diff] [blame]361 self.raw_requestline = self.rfile.readline(65537)
362 if len(self.raw_requestline) > 65536:
363 self.requestline = ''
364 self.request_version = ''
365 self.command = ''
366 self.send_error(414)
367 return
Kristján Valur Jónsson985fc6a2009-07-01 10:01:31 +0000[diff] [blame]368 if not self.raw_requestline:
369 self.close_connection = 1
370 return
371 if not self.parse_request():
372 # An error code has been sent, just exit
373 return
374 mname = 'do_' + self.command
375 if not hasattr(self, mname):
376 self.send_error(501, "Unsupported method (%r)" % self.command)
377 return
378 method = getattr(self, mname)
379 method()
380 self.wfile.flush() #actually send the response if not already done.
381 except socket.timeout as e:
382 #a read or a write timed out. Discard this connection
383 self.log_error("Request timed out: %r", e)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]384 self.close_connection = 1
385 return
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]386
387 def handle(self):
388 """Handle multiple requests if necessary."""
389 self.close_connection = 1
390
391 self.handle_one_request()
392 while not self.close_connection:
393 self.handle_one_request()
394
395 def send_error(self, code, message=None):
396 """Send and log an error reply.
397
398 Arguments are the error code, and a detailed message.
399 The detailed message defaults to the short entry matching the
400 response code.
401
402 This sends an error response (so it must be called before any
403 output has been generated), logs the error, and finally sends
404 a piece of HTML explaining the error to the user.
405
406 """
407
408 try:
409 shortmsg, longmsg = self.responses[code]
410 except KeyError:
411 shortmsg, longmsg = '???', '???'
412 if message is None:
413 message = shortmsg
414 explain = longmsg
415 self.log_error("code %d, message %s", code, message)
416 # using _quote_html to prevent Cross Site Scripting attacks (see bug #1100201)
417 content = (self.error_message_format %
418 {'code': code, 'message': _quote_html(message), 'explain': explain})
419 self.send_response(code, message)
420 self.send_header("Content-Type", self.error_content_type)
421 self.send_header('Connection', 'close')
422 self.end_headers()
423 if self.command != 'HEAD' and code >= 200 and code not in (204, 304):
424 self.wfile.write(content.encode('UTF-8', 'replace'))
425
426 def send_response(self, code, message=None):
427 """Send the response header and log the response code.
428
429 Also send two standard headers with the server software
430 version and the current date.
431
432 """
433 self.log_request(code)
Senthil Kumaran0f476d42010-09-30 06:09:18 +0000[diff] [blame]434 self.send_response_only(code, message)
435 self.send_header('Server', self.version_string())
436 self.send_header('Date', self.date_time_string())
437
438 def send_response_only(self, code, message=None):
439 """Send the response header only."""
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]440 if message is None:
441 if code in self.responses:
442 message = self.responses[code][0]
443 else:
444 message = ''
445 if self.request_version != 'HTTP/0.9':
446 self.wfile.write(("%s %d %s\r\n" %
447 (self.protocol_version, code, message)).encode('ASCII', 'strict'))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]448
449 def send_header(self, keyword, value):
450 """Send a MIME header."""
451 if self.request_version != 'HTTP/0.9':
Senthil Kumarane4dad4f2010-11-21 14:36:14 +0000[diff] [blame]452 if not hasattr(self, '_headers_buffer'):
453 self._headers_buffer = []
454 self._headers_buffer.append(
455 ("%s: %s\r\n" % (keyword, value)).encode('ASCII', 'strict'))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]456
457 if keyword.lower() == 'connection':
458 if value.lower() == 'close':
459 self.close_connection = 1
460 elif value.lower() == 'keep-alive':
461 self.close_connection = 0
462
463 def end_headers(self):
464 """Send the blank line ending the MIME headers."""
465 if self.request_version != 'HTTP/0.9':
Senthil Kumarane4dad4f2010-11-21 14:36:14 +0000[diff] [blame]466 self._headers_buffer.append(b"\r\n")
467 self.wfile.write(b"".join(self._headers_buffer))
468 self._headers_buffer = []
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]469
470 def log_request(self, code='-', size='-'):
471 """Log an accepted request.
472
473 This is called by send_response().
474
475 """
476
477 self.log_message('"%s" %s %s',
478 self.requestline, str(code), str(size))
479
480 def log_error(self, format, *args):
481 """Log an error.
482
483 This is called when a request cannot be fulfilled. By
484 default it passes the message on to log_message().
485
486 Arguments are the same as for log_message().
487
488 XXX This should go to the separate error log.
489
490 """
491
492 self.log_message(format, *args)
493
494 def log_message(self, format, *args):
495 """Log an arbitrary message.
496
497 This is used by all other logging functions. Override
498 it if you have specific logging wishes.
499
500 The first argument, FORMAT, is a format string for the
501 message to be logged. If the format string contains
502 any % escapes requiring parameters, they should be
503 specified as subsequent arguments (it's just like
504 printf!).
505
506 The client host and current date/time are prefixed to
507 every message.
508
509 """
510
511 sys.stderr.write("%s - - [%s] %s\n" %
512 (self.address_string(),
513 self.log_date_time_string(),
514 format%args))
515
516 def version_string(self):
517 """Return the server software version string."""
518 return self.server_version + ' ' + self.sys_version
519
520 def date_time_string(self, timestamp=None):
521 """Return the current date and time formatted for a message header."""
522 if timestamp is None:
523 timestamp = time.time()
524 year, month, day, hh, mm, ss, wd, y, z = time.gmtime(timestamp)
525 s = "%s, %02d %3s %4d %02d:%02d:%02d GMT" % (
526 self.weekdayname[wd],
527 day, self.monthname[month], year,
528 hh, mm, ss)
529 return s
530
531 def log_date_time_string(self):
532 """Return the current time formatted for logging."""
533 now = time.time()
534 year, month, day, hh, mm, ss, x, y, z = time.localtime(now)
535 s = "%02d/%3s/%04d %02d:%02d:%02d" % (
536 day, self.monthname[month], year, hh, mm, ss)
537 return s
538
539 weekdayname = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
540
541 monthname = [None,
542 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun',
543 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
544
545 def address_string(self):
546 """Return the client address formatted for logging.
547
548 This version looks up the full hostname using gethostbyaddr(),
549 and tries to find a name that contains at least one dot.
550
551 """
552
553 host, port = self.client_address[:2]
554 return socket.getfqdn(host)
555
556 # Essentially static class variables
557
558 # The version of the HTTP protocol we support.
559 # Set this to HTTP/1.1 to enable automatic keepalive
560 protocol_version = "HTTP/1.0"
561
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]562 # MessageClass used to parse headers
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]563 MessageClass = http.client.HTTPMessage
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]564
565 # Table mapping response codes to messages; entries have the
566 # form {code: (shortmessage, longmessage)}.
567 # See RFC 2616.
568 responses = {
569 100: ('Continue', 'Request received, please continue'),
570 101: ('Switching Protocols',
571 'Switching to new protocol; obey Upgrade header'),
572
573 200: ('OK', 'Request fulfilled, document follows'),
574 201: ('Created', 'Document created, URL follows'),
575 202: ('Accepted',
576 'Request accepted, processing continues off-line'),
577 203: ('Non-Authoritative Information', 'Request fulfilled from cache'),
578 204: ('No Content', 'Request fulfilled, nothing follows'),
579 205: ('Reset Content', 'Clear input form for further input.'),
580 206: ('Partial Content', 'Partial content follows.'),
581
582 300: ('Multiple Choices',
583 'Object has several resources -- see URI list'),
584 301: ('Moved Permanently', 'Object moved permanently -- see URI list'),
585 302: ('Found', 'Object moved temporarily -- see URI list'),
586 303: ('See Other', 'Object moved -- see Method and URL list'),
587 304: ('Not Modified',
588 'Document has not changed since given time'),
589 305: ('Use Proxy',
590 'You must use proxy specified in Location to access this '
591 'resource.'),
592 307: ('Temporary Redirect',
593 'Object moved temporarily -- see URI list'),
594
595 400: ('Bad Request',
596 'Bad request syntax or unsupported method'),
597 401: ('Unauthorized',
598 'No permission -- see authorization schemes'),
599 402: ('Payment Required',
600 'No payment -- see charging schemes'),
601 403: ('Forbidden',
602 'Request forbidden -- authorization will not help'),
603 404: ('Not Found', 'Nothing matches the given URI'),
604 405: ('Method Not Allowed',
Senthil Kumaran7aa26212010-02-22 11:00:50 +0000[diff] [blame]605 'Specified method is invalid for this resource.'),
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]606 406: ('Not Acceptable', 'URI not available in preferred format.'),
607 407: ('Proxy Authentication Required', 'You must authenticate with '
608 'this proxy before proceeding.'),
609 408: ('Request Timeout', 'Request timed out; try again later.'),
610 409: ('Conflict', 'Request conflict.'),
611 410: ('Gone',
612 'URI no longer exists and has been permanently removed.'),
613 411: ('Length Required', 'Client must specify Content-Length.'),
614 412: ('Precondition Failed', 'Precondition in headers is false.'),
615 413: ('Request Entity Too Large', 'Entity is too large.'),
616 414: ('Request-URI Too Long', 'URI is too long.'),
617 415: ('Unsupported Media Type', 'Entity body in unsupported format.'),
618 416: ('Requested Range Not Satisfiable',
619 'Cannot satisfy request range.'),
620 417: ('Expectation Failed',
621 'Expect condition could not be satisfied.'),
622
623 500: ('Internal Server Error', 'Server got itself in trouble'),
624 501: ('Not Implemented',
625 'Server does not support this operation'),
626 502: ('Bad Gateway', 'Invalid responses from another server/proxy.'),
627 503: ('Service Unavailable',
628 'The server cannot process the request due to a high load'),
629 504: ('Gateway Timeout',
630 'The gateway server did not receive a timely response'),
631 505: ('HTTP Version Not Supported', 'Cannot fulfill request.'),
632 }
633
634
635class SimpleHTTPRequestHandler(BaseHTTPRequestHandler):
636
637 """Simple HTTP request handler with GET and HEAD commands.
638
639 This serves files from the current directory and any of its
640 subdirectories. The MIME type for files is determined by
641 calling the .guess_type() method.
642
643 The GET and HEAD requests are identical except that the HEAD
644 request omits the actual contents of the file.
645
646 """
647
648 server_version = "SimpleHTTP/" + __version__
649
650 def do_GET(self):
651 """Serve a GET request."""
652 f = self.send_head()
653 if f:
654 self.copyfile(f, self.wfile)
655 f.close()
656
657 def do_HEAD(self):
658 """Serve a HEAD request."""
659 f = self.send_head()
660 if f:
661 f.close()
662
663 def send_head(self):
664 """Common code for GET and HEAD commands.
665
666 This sends the response code and MIME headers.
667
668 Return value is either a file object (which has to be copied
669 to the outputfile by the caller unless the command was HEAD,
670 and must be closed by the caller under all circumstances), or
671 None, in which case the caller has nothing further to do.
672
673 """
674 path = self.translate_path(self.path)
675 f = None
676 if os.path.isdir(path):
677 if not self.path.endswith('/'):
678 # redirect browser - doing basically what apache does
679 self.send_response(301)
680 self.send_header("Location", self.path + "/")
681 self.end_headers()
682 return None
683 for index in "index.html", "index.htm":
684 index = os.path.join(path, index)
685 if os.path.exists(index):
686 path = index
687 break
688 else:
689 return self.list_directory(path)
690 ctype = self.guess_type(path)
691 try:
692 f = open(path, 'rb')
693 except IOError:
694 self.send_error(404, "File not found")
695 return None
696 self.send_response(200)
697 self.send_header("Content-type", ctype)
698 fs = os.fstat(f.fileno())
699 self.send_header("Content-Length", str(fs[6]))
700 self.send_header("Last-Modified", self.date_time_string(fs.st_mtime))
701 self.end_headers()
702 return f
703
704 def list_directory(self, path):
705 """Helper to produce a directory listing (absent index.html).
706
707 Return value is either a file object, or None (indicating an
708 error). In either case, the headers are sent, making the
709 interface the same as for send_head().
710
711 """
712 try:
713 list = os.listdir(path)
714 except os.error:
715 self.send_error(404, "No permission to list directory")
716 return None
717 list.sort(key=lambda a: a.lower())
718 r = []
Georg Brandl1f7fffb2010-10-15 15:57:45 +0000[diff] [blame]719 displaypath = html.escape(urllib.parse.unquote(self.path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]720 r.append('<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">')
721 r.append("<html>\n<title>Directory listing for %s</title>\n" % displaypath)
722 r.append("<body>\n<h2>Directory listing for %s</h2>\n" % displaypath)
723 r.append("<hr>\n<ul>\n")
724 for name in list:
725 fullname = os.path.join(path, name)
726 displayname = linkname = name
727 # Append / for directories or @ for symbolic links
728 if os.path.isdir(fullname):
729 displayname = name + "/"
730 linkname = name + "/"
731 if os.path.islink(fullname):
732 displayname = name + "@"
733 # Note: a link to a directory displays with @ and links with /
734 r.append('<li><a href="%s">%s</a>\n'
Georg Brandl1f7fffb2010-10-15 15:57:45 +0000[diff] [blame]735 % (urllib.parse.quote(linkname), html.escape(displayname)))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]736 r.append("</ul>\n<hr>\n</body>\n</html>\n")
737 enc = sys.getfilesystemencoding()
738 encoded = ''.join(r).encode(enc)
739 f = io.BytesIO()
740 f.write(encoded)
741 f.seek(0)
742 self.send_response(200)
743 self.send_header("Content-type", "text/html; charset=%s" % enc)
744 self.send_header("Content-Length", str(len(encoded)))
745 self.end_headers()
746 return f
747
748 def translate_path(self, path):
749 """Translate a /-separated PATH to the local filename syntax.
750
751 Components that mean special things to the local file system
752 (e.g. drive or directory names) are ignored. (XXX They should
753 probably be diagnosed.)
754
755 """
756 # abandon query parameters
757 path = path.split('?',1)[0]
758 path = path.split('#',1)[0]
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]759 path = posixpath.normpath(urllib.parse.unquote(path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]760 words = path.split('/')
761 words = filter(None, words)
762 path = os.getcwd()
763 for word in words:
764 drive, word = os.path.splitdrive(word)
765 head, word = os.path.split(word)
766 if word in (os.curdir, os.pardir): continue
767 path = os.path.join(path, word)
768 return path
769
770 def copyfile(self, source, outputfile):
771 """Copy all data between two file objects.
772
773 The SOURCE argument is a file object open for reading
774 (or anything with a read() method) and the DESTINATION
775 argument is a file object open for writing (or
776 anything with a write() method).
777
778 The only reason for overriding this would be to change
779 the block size or perhaps to replace newlines by CRLF
780 -- note however that this the default server uses this
781 to copy binary data as well.
782
783 """
784 shutil.copyfileobj(source, outputfile)
785
786 def guess_type(self, path):
787 """Guess the type of a file.
788
789 Argument is a PATH (a filename).
790
791 Return value is a string of the form type/subtype,
792 usable for a MIME Content-type header.
793
794 The default implementation looks the file's extension
795 up in the table self.extensions_map, using application/octet-stream
796 as a default; however it would be permissible (if
797 slow) to look inside the data to make a better guess.
798
799 """
800
801 base, ext = posixpath.splitext(path)
802 if ext in self.extensions_map:
803 return self.extensions_map[ext]
804 ext = ext.lower()
805 if ext in self.extensions_map:
806 return self.extensions_map[ext]
807 else:
808 return self.extensions_map['']
809
810 if not mimetypes.inited:
811 mimetypes.init() # try to read system mime.types
812 extensions_map = mimetypes.types_map.copy()
813 extensions_map.update({
814 '': 'application/octet-stream', # Default
815 '.py': 'text/plain',
816 '.c': 'text/plain',
817 '.h': 'text/plain',
818 })
819
820
821# Utilities for CGIHTTPRequestHandler
822
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]823# TODO(gregory.p.smith): Move this into an appropriate library.
824def _url_collapse_path_split(path):
825 """
826 Given a URL path, remove extra '/'s and '.' path elements and collapse
827 any '..' references.
828
829 Implements something akin to RFC-2396 5.2 step 6 to parse relative paths.
830
831 Returns: A tuple of (head, tail) where tail is everything after the final /
832 and head is everything before it. Head will always start with a '/' and,
833 if it contains anything else, never have a trailing '/'.
834
835 Raises: IndexError if too many '..' occur within the path.
836 """
837 # Similar to os.path.split(os.path.normpath(path)) but specific to URL
838 # path semantics rather than local operating system semantics.
839 path_parts = []
840 for part in path.split('/'):
841 if part == '.':
842 path_parts.append('')
843 else:
844 path_parts.append(part)
845 # Filter out blank non trailing parts before consuming the '..'.
846 path_parts = [part for part in path_parts[:-1] if part] + path_parts[-1:]
847 if path_parts:
848 tail_part = path_parts.pop()
849 else:
850 tail_part = ''
851 head_parts = []
852 for part in path_parts:
853 if part == '..':
854 head_parts.pop()
855 else:
856 head_parts.append(part)
857 if tail_part and tail_part == '..':
858 head_parts.pop()
859 tail_part = ''
860 return ('/' + '/'.join(head_parts), tail_part)
861
862
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]863nobody = None
864
865def nobody_uid():
866 """Internal routine to get nobody's uid"""
867 global nobody
868 if nobody:
869 return nobody
870 try:
871 import pwd
872 except ImportError:
873 return -1
874 try:
875 nobody = pwd.getpwnam('nobody')[2]
876 except KeyError:
Georg Brandlcbd2ab12010-12-04 10:39:14 +0000[diff] [blame]877 nobody = 1 + max(x[2] for x in pwd.getpwall())
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]878 return nobody
879
880
881def executable(path):
882 """Test for executable file."""
883 try:
884 st = os.stat(path)
885 except os.error:
886 return False
887 return st.st_mode & 0o111 != 0
888
889
890class CGIHTTPRequestHandler(SimpleHTTPRequestHandler):
891
892 """Complete HTTP server with GET, HEAD and POST commands.
893
894 GET and HEAD also support running CGI scripts.
895
896 The POST command is *only* implemented for CGI scripts.
897
898 """
899
900 # Determine platform specifics
901 have_fork = hasattr(os, 'fork')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]902
903 # Make rfile unbuffered -- we need to read one line and then pass
904 # the rest to a subprocess, so we can't use buffered input.
905 rbufsize = 0
906
907 def do_POST(self):
908 """Serve a POST request.
909
910 This is only implemented for CGI scripts.
911
912 """
913
914 if self.is_cgi():
915 self.run_cgi()
916 else:
917 self.send_error(501, "Can only POST to CGI scripts")
918
919 def send_head(self):
920 """Version of send_head that support CGI scripts"""
921 if self.is_cgi():
922 return self.run_cgi()
923 else:
924 return SimpleHTTPRequestHandler.send_head(self)
925
926 def is_cgi(self):
927 """Test whether self.path corresponds to a CGI script.
928
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]929 Returns True and updates the cgi_info attribute to the tuple
930 (dir, rest) if self.path requires running a CGI script.
931 Returns False otherwise.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]932
Benjamin Petersona7deeee2009-05-08 20:54:42 +0000[diff] [blame]933 If any exception is raised, the caller should assume that
934 self.path was rejected as invalid and act accordingly.
935
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]936 The default implementation tests whether the normalized url
937 path begins with one of the strings in self.cgi_directories
938 (and the next character is a '/' or the end of the string).
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]939
940 """
941
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]942 splitpath = _url_collapse_path_split(self.path)
943 if splitpath[0] in self.cgi_directories:
944 self.cgi_info = splitpath
945 return True
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]946 return False
947
948 cgi_directories = ['/cgi-bin', '/htbin']
949
950 def is_executable(self, path):
951 """Test whether argument path is an executable file."""
952 return executable(path)
953
954 def is_python(self, path):
955 """Test whether argument path is a Python script."""
956 head, tail = os.path.splitext(path)
957 return tail.lower() in (".py", ".pyw")
958
959 def run_cgi(self):
960 """Execute a CGI script."""
961 path = self.path
962 dir, rest = self.cgi_info
963
964 i = path.find('/', len(dir) + 1)
965 while i >= 0:
966 nextdir = path[:i]
967 nextrest = path[i+1:]
968
969 scriptdir = self.translate_path(nextdir)
970 if os.path.isdir(scriptdir):
971 dir, rest = nextdir, nextrest
972 i = path.find('/', len(dir) + 1)
973 else:
974 break
975
976 # find an explicit query string, if present.
977 i = rest.rfind('?')
978 if i >= 0:
979 rest, query = rest[:i], rest[i+1:]
980 else:
981 query = ''
982
983 # dissect the part after the directory name into a script name &
984 # a possible additional path, to be stored in PATH_INFO.
985 i = rest.find('/')
986 if i >= 0:
987 script, rest = rest[:i], rest[i:]
988 else:
989 script, rest = rest, ''
990
991 scriptname = dir + '/' + script
992 scriptfile = self.translate_path(scriptname)
993 if not os.path.exists(scriptfile):
994 self.send_error(404, "No such CGI script (%r)" % scriptname)
995 return
996 if not os.path.isfile(scriptfile):
997 self.send_error(403, "CGI script is not a plain file (%r)" %
998 scriptname)
999 return
1000 ispy = self.is_python(scriptname)
1001 if not ispy:
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1002 if not self.is_executable(scriptfile):
1003 self.send_error(403, "CGI script is not executable (%r)" %
1004 scriptname)
1005 return
1006
1007 # Reference: http://hoohoo.ncsa.uiuc.edu/cgi/env.html
1008 # XXX Much of the following could be prepared ahead of time!
Senthil Kumaran42713722010-10-03 17:55:45 +0000[diff] [blame]1009 env = copy.deepcopy(os.environ)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1010 env['SERVER_SOFTWARE'] = self.version_string()
1011 env['SERVER_NAME'] = self.server.server_name
1012 env['GATEWAY_INTERFACE'] = 'CGI/1.1'
1013 env['SERVER_PROTOCOL'] = self.protocol_version
1014 env['SERVER_PORT'] = str(self.server.server_port)
1015 env['REQUEST_METHOD'] = self.command
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]1016 uqrest = urllib.parse.unquote(rest)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1017 env['PATH_INFO'] = uqrest
1018 env['PATH_TRANSLATED'] = self.translate_path(uqrest)
1019 env['SCRIPT_NAME'] = scriptname
1020 if query:
1021 env['QUERY_STRING'] = query
1022 host = self.address_string()
1023 if host != self.client_address[0]:
1024 env['REMOTE_HOST'] = host
1025 env['REMOTE_ADDR'] = self.client_address[0]
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1026 authorization = self.headers.get("authorization")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1027 if authorization:
1028 authorization = authorization.split()
1029 if len(authorization) == 2:
1030 import base64, binascii
1031 env['AUTH_TYPE'] = authorization[0]
1032 if authorization[0].lower() == "basic":
1033 try:
1034 authorization = authorization[1].encode('ascii')
Georg Brandl706824f2009-06-04 09:42:55 +0000[diff] [blame]1035 authorization = base64.decodebytes(authorization).\
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1036 decode('ascii')
1037 except (binascii.Error, UnicodeError):
1038 pass
1039 else:
1040 authorization = authorization.split(':')
1041 if len(authorization) == 2:
1042 env['REMOTE_USER'] = authorization[0]
1043 # XXX REMOTE_IDENT
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1044 if self.headers.get('content-type') is None:
1045 env['CONTENT_TYPE'] = self.headers.get_content_type()
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1046 else:
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1047 env['CONTENT_TYPE'] = self.headers['content-type']
1048 length = self.headers.get('content-length')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1049 if length:
1050 env['CONTENT_LENGTH'] = length
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1051 referer = self.headers.get('referer')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1052 if referer:
1053 env['HTTP_REFERER'] = referer
1054 accept = []
1055 for line in self.headers.getallmatchingheaders('accept'):
1056 if line[:1] in "\t\n\r ":
1057 accept.append(line.strip())
1058 else:
1059 accept = accept + line[7:].split(',')
1060 env['HTTP_ACCEPT'] = ','.join(accept)
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1061 ua = self.headers.get('user-agent')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1062 if ua:
1063 env['HTTP_USER_AGENT'] = ua
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1064 co = filter(None, self.headers.get_all('cookie', []))
Georg Brandl62e2ca22010-07-31 21:54:24 +0000[diff] [blame]1065 cookie_str = ', '.join(co)
1066 if cookie_str:
1067 env['HTTP_COOKIE'] = cookie_str
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1068 # XXX Other HTTP_* headers
1069 # Since we're setting the env in the parent, provide empty
1070 # values to override previously set values
1071 for k in ('QUERY_STRING', 'REMOTE_HOST', 'CONTENT_LENGTH',
1072 'HTTP_USER_AGENT', 'HTTP_COOKIE', 'HTTP_REFERER'):
1073 env.setdefault(k, "")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1074
1075 self.send_response(200, "Script output follows")
1076
1077 decoded_query = query.replace('+', ' ')
1078
1079 if self.have_fork:
1080 # Unix -- fork as we should
1081 args = [script]
1082 if '=' not in decoded_query:
1083 args.append(decoded_query)
1084 nobody = nobody_uid()
1085 self.wfile.flush() # Always flush before forking
1086 pid = os.fork()
1087 if pid != 0:
1088 # Parent
1089 pid, sts = os.waitpid(pid, 0)
1090 # throw away additional data [see bug #427345]
1091 while select.select([self.rfile], [], [], 0)[0]:
1092 if not self.rfile.read(1):
1093 break
1094 if sts:
1095 self.log_error("CGI script exit status %#x", sts)
1096 return
1097 # Child
1098 try:
1099 try:
1100 os.setuid(nobody)
1101 except os.error:
1102 pass
1103 os.dup2(self.rfile.fileno(), 0)
1104 os.dup2(self.wfile.fileno(), 1)
Senthil Kumaran42713722010-10-03 17:55:45 +0000[diff] [blame]1105 os.execve(scriptfile, args, env)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1106 except:
1107 self.server.handle_error(self.request, self.client_address)
1108 os._exit(127)
1109
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1110 else:
1111 # Non-Unix -- use subprocess
1112 import subprocess
Senthil Kumarane29cd162009-11-11 04:17:53 +0000[diff] [blame]1113 cmdline = [scriptfile]
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1114 if self.is_python(scriptfile):
1115 interp = sys.executable
1116 if interp.lower().endswith("w.exe"):
1117 # On Windows, use python.exe, not pythonw.exe
1118 interp = interp[:-5] + interp[-4:]
Senthil Kumarane29cd162009-11-11 04:17:53 +0000[diff] [blame]1119 cmdline = [interp, '-u'] + cmdline
1120 if '=' not in query:
1121 cmdline.append(query)
1122 self.log_message("command: %s", subprocess.list2cmdline(cmdline))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1123 try:
1124 nbytes = int(length)
1125 except (TypeError, ValueError):
1126 nbytes = 0
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1127 p = subprocess.Popen(cmdline,
1128 stdin=subprocess.PIPE,
1129 stdout=subprocess.PIPE,
Senthil Kumaran42713722010-10-03 17:55:45 +0000[diff] [blame]1130 stderr=subprocess.PIPE,
1131 env = env
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1132 )
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1133 if self.command.lower() == "post" and nbytes > 0:
1134 data = self.rfile.read(nbytes)
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1135 else:
1136 data = None
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1137 # throw away additional data [see bug #427345]
1138 while select.select([self.rfile._sock], [], [], 0)[0]:
1139 if not self.rfile._sock.recv(1):
1140 break
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1141 stdout, stderr = p.communicate(data)
1142 self.wfile.write(stdout)
1143 if stderr:
1144 self.log_error('%s', stderr)
Brian Curtincbad4df2010-11-05 15:04:48 +0000[diff] [blame]1145 p.stderr.close()
1146 p.stdout.close()
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1147 status = p.returncode
1148 if status:
1149 self.log_error("CGI script exit status %#x", status)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1150 else:
1151 self.log_message("CGI script exited OK")
1152
1153
1154def test(HandlerClass = BaseHTTPRequestHandler,
1155 ServerClass = HTTPServer, protocol="HTTP/1.0"):
1156 """Test the HTTP request handler class.
1157
1158 This runs an HTTP server on port 8000 (or the first command line
1159 argument).
1160
1161 """
1162
1163 if sys.argv[1:]:
1164 port = int(sys.argv[1])
1165 else:
1166 port = 8000
1167 server_address = ('', port)
1168
1169 HandlerClass.protocol_version = protocol
1170 httpd = ServerClass(server_address, HandlerClass)
1171
1172 sa = httpd.socket.getsockname()
1173 print("Serving HTTP on", sa[0], "port", sa[1], "...")
Alexandre Vassalottib5292a22009-04-03 07:16:55 +0000[diff] [blame]1174 try:
1175 httpd.serve_forever()
1176 except KeyboardInterrupt:
1177 print("\nKeyboard interrupt received, exiting.")
1178 httpd.server_close()
1179 sys.exit(0)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1180
1181if __name__ == '__main__':
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1182 test(HandlerClass=SimpleHTTPRequestHandler)