Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / python / cpython2 / 1b407fe658fb298986d1ed3f8a71937ac1b3460c / . / Lib / http / server.py
blob: fccdc4c92eeb32ffe88a4c6b396a135e2e733d06 [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 Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]87import cgi
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 Kumaran5e8826c2010-10-03 18:04:52 +0000[diff] [blame]102import copy
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]103
104# Default error message template
105DEFAULT_ERROR_MESSAGE = """\
Senthil Kumaran1b407fe2011-03-20 10:44:30 +0800[diff] [blame^]106<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
107 "http://www.w3.org/TR/html4/strict.dtd">
Senthil Kumaran9f9193e2011-03-17 17:01:45 +0800[diff] [blame]108 <head>
Senthil Kumaran1b407fe2011-03-20 10:44:30 +0800[diff] [blame^]109 <meta http-equiv="Content-Type" content="text/html;charset=utf-8">
Senthil Kumaran9f9193e2011-03-17 17:01:45 +0800[diff] [blame]110 <title>Error response</title>
111 </head>
112 <body>
113 <h1>Error response</h1>
114 <p>Error code: %(code)d</p>
115 <p>Message: %(message)s.</p>
116 <p>Error code explanation: %(code)s - %(explain)s.</p>
117 </body>
118</html>
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]119"""
120
121DEFAULT_ERROR_CONTENT_TYPE = "text/html;charset=utf-8"
122
123def _quote_html(html):
124 return html.replace("&", "&amp;").replace("<", "&lt;").replace(">", "&gt;")
125
126class HTTPServer(socketserver.TCPServer):
127
128 allow_reuse_address = 1 # Seems to make sense in testing environment
129
130 def server_bind(self):
131 """Override server_bind to store the server name."""
132 socketserver.TCPServer.server_bind(self)
133 host, port = self.socket.getsockname()[:2]
134 self.server_name = socket.getfqdn(host)
135 self.server_port = port
136
137
138class BaseHTTPRequestHandler(socketserver.StreamRequestHandler):
139
140 """HTTP request handler base class.
141
142 The following explanation of HTTP serves to guide you through the
143 code as well as to expose any misunderstandings I may have about
144 HTTP (so you don't need to read the code to figure out I'm wrong
145 :-).
146
147 HTTP (HyperText Transfer Protocol) is an extensible protocol on
148 top of a reliable stream transport (e.g. TCP/IP). The protocol
149 recognizes three parts to a request:
150
151 1. One line identifying the request type and path
152 2. An optional set of RFC-822-style headers
153 3. An optional data part
154
155 The headers and data are separated by a blank line.
156
157 The first line of the request has the form
158
159 <command> <path> <version>
160
161 where <command> is a (case-sensitive) keyword such as GET or POST,
162 <path> is a string containing path information for the request,
163 and <version> should be the string "HTTP/1.0" or "HTTP/1.1".
164 <path> is encoded using the URL encoding scheme (using %xx to signify
165 the ASCII character with hex code xx).
166
167 The specification specifies that lines are separated by CRLF but
168 for compatibility with the widest range of clients recommends
169 servers also handle LF. Similarly, whitespace in the request line
170 is treated sensibly (allowing multiple spaces between components
171 and allowing trailing whitespace).
172
173 Similarly, for output, lines ought to be separated by CRLF pairs
174 but most clients grok LF characters just fine.
175
176 If the first line of the request has the form
177
178 <command> <path>
179
180 (i.e. <version> is left out) then this is assumed to be an HTTP
181 0.9 request; this form has no optional headers and data part and
182 the reply consists of just the data.
183
184 The reply form of the HTTP 1.x protocol again has three parts:
185
186 1. One line giving the response code
187 2. An optional set of RFC-822-style headers
188 3. The data
189
190 Again, the headers and data are separated by a blank line.
191
192 The response code line has the form
193
194 <version> <responsecode> <responsestring>
195
196 where <version> is the protocol version ("HTTP/1.0" or "HTTP/1.1"),
197 <responsecode> is a 3-digit response code indicating success or
198 failure of the request, and <responsestring> is an optional
199 human-readable string explaining what the response code means.
200
201 This server parses the request and the headers, and then calls a
202 function specific to the request type (<command>). Specifically,
203 a request SPAM will be handled by a method do_SPAM(). If no
204 such method exists the server sends an error response to the
205 client. If it exists, it is called with no arguments:
206
207 do_SPAM()
208
209 Note that the request name is case sensitive (i.e. SPAM and spam
210 are different requests).
211
212 The various request details are stored in instance variables:
213
214 - client_address is the client IP address in the form (host,
215 port);
216
217 - command, path and version are the broken-down request line;
218
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]219 - headers is an instance of email.message.Message (or a derived
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]220 class) containing the header information;
221
222 - rfile is a file object open for reading positioned at the
223 start of the optional input data part;
224
225 - wfile is a file object open for writing.
226
227 IT IS IMPORTANT TO ADHERE TO THE PROTOCOL FOR WRITING!
228
229 The first thing to be written must be the response line. Then
230 follow 0 or more header lines, then a blank line, and then the
231 actual data (if any). The meaning of the header lines depends on
232 the command executed by the server; in most cases, when data is
233 returned, there should be at least one header line of the form
234
235 Content-type: <type>/<subtype>
236
237 where <type> and <subtype> should be registered MIME types,
238 e.g. "text/html" or "text/plain".
239
240 """
241
242 # The Python system version, truncated to its first component.
243 sys_version = "Python/" + sys.version.split()[0]
244
245 # The server software version. You may want to override this.
246 # The format is multiple whitespace-separated strings,
247 # where each string is of the form name[/version].
248 server_version = "BaseHTTP/" + __version__
249
250 error_message_format = DEFAULT_ERROR_MESSAGE
251 error_content_type = DEFAULT_ERROR_CONTENT_TYPE
252
253 # The default request version. This only affects responses up until
254 # the point where the request line is parsed, so it mainly decides what
255 # the client gets back when sending a malformed request line.
256 # Most web servers default to HTTP 0.9, i.e. don't send a status line.
257 default_request_version = "HTTP/0.9"
258
259 def parse_request(self):
260 """Parse a request (internal).
261
262 The request should be stored in self.raw_requestline; the results
263 are in self.command, self.path, self.request_version and
264 self.headers.
265
266 Return True for success, False for failure; on failure, an
267 error is sent back.
268
269 """
270 self.command = None # set in case of error on the first line
271 self.request_version = version = self.default_request_version
272 self.close_connection = 1
273 requestline = str(self.raw_requestline, 'iso-8859-1')
274 if requestline[-2:] == '\r\n':
275 requestline = requestline[:-2]
276 elif requestline[-1:] == '\n':
277 requestline = requestline[:-1]
278 self.requestline = requestline
279 words = requestline.split()
280 if len(words) == 3:
281 [command, path, version] = words
282 if version[:5] != 'HTTP/':
283 self.send_error(400, "Bad request version (%r)" % version)
284 return False
285 try:
286 base_version_number = version.split('/', 1)[1]
287 version_number = base_version_number.split(".")
288 # RFC 2145 section 3.1 says there can be only one "." and
289 # - major and minor numbers MUST be treated as
290 # separate integers;
291 # - HTTP/2.4 is a lower version than HTTP/2.13, which in
292 # turn is lower than HTTP/12.3;
293 # - Leading zeros MUST be ignored by recipients.
294 if len(version_number) != 2:
295 raise ValueError
296 version_number = int(version_number[0]), int(version_number[1])
297 except (ValueError, IndexError):
298 self.send_error(400, "Bad request version (%r)" % version)
299 return False
300 if version_number >= (1, 1) and self.protocol_version >= "HTTP/1.1":
301 self.close_connection = 0
302 if version_number >= (2, 0):
303 self.send_error(505,
304 "Invalid HTTP Version (%s)" % base_version_number)
305 return False
306 elif len(words) == 2:
307 [command, path] = words
308 self.close_connection = 1
309 if command != 'GET':
310 self.send_error(400,
311 "Bad HTTP/0.9 request type (%r)" % command)
312 return False
313 elif not words:
314 return False
315 else:
316 self.send_error(400, "Bad request syntax (%r)" % requestline)
317 return False
318 self.command, self.path, self.request_version = command, path, version
319
320 # Examine the headers and look for a Connection directive.
Antoine Pitrouff1bbba2010-12-18 18:04:38 +0000[diff] [blame]321 try:
322 self.headers = http.client.parse_headers(self.rfile,
323 _class=self.MessageClass)
324 except http.client.LineTooLong:
325 self.send_error(400, "Line too long")
326 return False
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]327
328 conntype = self.headers.get('Connection', "")
329 if conntype.lower() == 'close':
330 self.close_connection = 1
331 elif (conntype.lower() == 'keep-alive' and
332 self.protocol_version >= "HTTP/1.1"):
333 self.close_connection = 0
334 return True
335
336 def handle_one_request(self):
337 """Handle a single HTTP request.
338
339 You normally don't need to override this method; see the class
340 __doc__ string for information on how to handle specific HTTP
341 commands such as GET and POST.
342
343 """
Antoine Pitrou3022ce12010-12-16 17:03:16 +0000[diff] [blame]344 self.raw_requestline = self.rfile.readline(65537)
345 if len(self.raw_requestline) > 65536:
346 self.requestline = ''
347 self.request_version = ''
348 self.command = ''
349 self.send_error(414)
350 return
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]351 if not self.raw_requestline:
352 self.close_connection = 1
353 return
354 if not self.parse_request(): # An error code has been sent, just exit
355 return
356 mname = 'do_' + self.command
357 if not hasattr(self, mname):
358 self.send_error(501, "Unsupported method (%r)" % self.command)
359 return
360 method = getattr(self, mname)
361 method()
362
363 def handle(self):
364 """Handle multiple requests if necessary."""
365 self.close_connection = 1
366
367 self.handle_one_request()
368 while not self.close_connection:
369 self.handle_one_request()
370
371 def send_error(self, code, message=None):
372 """Send and log an error reply.
373
374 Arguments are the error code, and a detailed message.
375 The detailed message defaults to the short entry matching the
376 response code.
377
378 This sends an error response (so it must be called before any
379 output has been generated), logs the error, and finally sends
380 a piece of HTML explaining the error to the user.
381
382 """
383
384 try:
385 shortmsg, longmsg = self.responses[code]
386 except KeyError:
387 shortmsg, longmsg = '???', '???'
388 if message is None:
389 message = shortmsg
390 explain = longmsg
391 self.log_error("code %d, message %s", code, message)
392 # using _quote_html to prevent Cross Site Scripting attacks (see bug #1100201)
393 content = (self.error_message_format %
394 {'code': code, 'message': _quote_html(message), 'explain': explain})
395 self.send_response(code, message)
396 self.send_header("Content-Type", self.error_content_type)
397 self.send_header('Connection', 'close')
398 self.end_headers()
399 if self.command != 'HEAD' and code >= 200 and code not in (204, 304):
400 self.wfile.write(content.encode('UTF-8', 'replace'))
401
402 def send_response(self, code, message=None):
403 """Send the response header and log the response code.
404
405 Also send two standard headers with the server software
406 version and the current date.
407
408 """
409 self.log_request(code)
410 if message is None:
411 if code in self.responses:
412 message = self.responses[code][0]
413 else:
414 message = ''
415 if self.request_version != 'HTTP/0.9':
416 self.wfile.write(("%s %d %s\r\n" %
417 (self.protocol_version, code, message)).encode('ASCII', 'strict'))
418 # print (self.protocol_version, code, message)
419 self.send_header('Server', self.version_string())
420 self.send_header('Date', self.date_time_string())
421
422 def send_header(self, keyword, value):
423 """Send a MIME header."""
424 if self.request_version != 'HTTP/0.9':
425 self.wfile.write(("%s: %s\r\n" % (keyword, value)).encode('ASCII', 'strict'))
426
427 if keyword.lower() == 'connection':
428 if value.lower() == 'close':
429 self.close_connection = 1
430 elif value.lower() == 'keep-alive':
431 self.close_connection = 0
432
433 def end_headers(self):
434 """Send the blank line ending the MIME headers."""
435 if self.request_version != 'HTTP/0.9':
436 self.wfile.write(b"\r\n")
437
438 def log_request(self, code='-', size='-'):
439 """Log an accepted request.
440
441 This is called by send_response().
442
443 """
444
445 self.log_message('"%s" %s %s',
446 self.requestline, str(code), str(size))
447
448 def log_error(self, format, *args):
449 """Log an error.
450
451 This is called when a request cannot be fulfilled. By
452 default it passes the message on to log_message().
453
454 Arguments are the same as for log_message().
455
456 XXX This should go to the separate error log.
457
458 """
459
460 self.log_message(format, *args)
461
462 def log_message(self, format, *args):
463 """Log an arbitrary message.
464
465 This is used by all other logging functions. Override
466 it if you have specific logging wishes.
467
468 The first argument, FORMAT, is a format string for the
469 message to be logged. If the format string contains
470 any % escapes requiring parameters, they should be
471 specified as subsequent arguments (it's just like
472 printf!).
473
474 The client host and current date/time are prefixed to
475 every message.
476
477 """
478
479 sys.stderr.write("%s - - [%s] %s\n" %
480 (self.address_string(),
481 self.log_date_time_string(),
482 format%args))
483
484 def version_string(self):
485 """Return the server software version string."""
486 return self.server_version + ' ' + self.sys_version
487
488 def date_time_string(self, timestamp=None):
489 """Return the current date and time formatted for a message header."""
490 if timestamp is None:
491 timestamp = time.time()
492 year, month, day, hh, mm, ss, wd, y, z = time.gmtime(timestamp)
493 s = "%s, %02d %3s %4d %02d:%02d:%02d GMT" % (
494 self.weekdayname[wd],
495 day, self.monthname[month], year,
496 hh, mm, ss)
497 return s
498
499 def log_date_time_string(self):
500 """Return the current time formatted for logging."""
501 now = time.time()
502 year, month, day, hh, mm, ss, x, y, z = time.localtime(now)
503 s = "%02d/%3s/%04d %02d:%02d:%02d" % (
504 day, self.monthname[month], year, hh, mm, ss)
505 return s
506
507 weekdayname = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
508
509 monthname = [None,
510 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun',
511 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
512
513 def address_string(self):
514 """Return the client address formatted for logging.
515
516 This version looks up the full hostname using gethostbyaddr(),
517 and tries to find a name that contains at least one dot.
518
519 """
520
521 host, port = self.client_address[:2]
522 return socket.getfqdn(host)
523
524 # Essentially static class variables
525
526 # The version of the HTTP protocol we support.
527 # Set this to HTTP/1.1 to enable automatic keepalive
528 protocol_version = "HTTP/1.0"
529
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]530 # MessageClass used to parse headers
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]531 MessageClass = http.client.HTTPMessage
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]532
533 # Table mapping response codes to messages; entries have the
534 # form {code: (shortmessage, longmessage)}.
535 # See RFC 2616.
536 responses = {
537 100: ('Continue', 'Request received, please continue'),
538 101: ('Switching Protocols',
539 'Switching to new protocol; obey Upgrade header'),
540
541 200: ('OK', 'Request fulfilled, document follows'),
542 201: ('Created', 'Document created, URL follows'),
543 202: ('Accepted',
544 'Request accepted, processing continues off-line'),
545 203: ('Non-Authoritative Information', 'Request fulfilled from cache'),
546 204: ('No Content', 'Request fulfilled, nothing follows'),
547 205: ('Reset Content', 'Clear input form for further input.'),
548 206: ('Partial Content', 'Partial content follows.'),
549
550 300: ('Multiple Choices',
551 'Object has several resources -- see URI list'),
552 301: ('Moved Permanently', 'Object moved permanently -- see URI list'),
553 302: ('Found', 'Object moved temporarily -- see URI list'),
554 303: ('See Other', 'Object moved -- see Method and URL list'),
555 304: ('Not Modified',
556 'Document has not changed since given time'),
557 305: ('Use Proxy',
558 'You must use proxy specified in Location to access this '
559 'resource.'),
560 307: ('Temporary Redirect',
561 'Object moved temporarily -- see URI list'),
562
563 400: ('Bad Request',
564 'Bad request syntax or unsupported method'),
565 401: ('Unauthorized',
566 'No permission -- see authorization schemes'),
567 402: ('Payment Required',
568 'No payment -- see charging schemes'),
569 403: ('Forbidden',
570 'Request forbidden -- authorization will not help'),
571 404: ('Not Found', 'Nothing matches the given URI'),
572 405: ('Method Not Allowed',
Senthil Kumaran613c61c2010-02-22 11:02:53 +0000[diff] [blame]573 'Specified method is invalid for this resource.'),
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]574 406: ('Not Acceptable', 'URI not available in preferred format.'),
575 407: ('Proxy Authentication Required', 'You must authenticate with '
576 'this proxy before proceeding.'),
577 408: ('Request Timeout', 'Request timed out; try again later.'),
578 409: ('Conflict', 'Request conflict.'),
579 410: ('Gone',
580 'URI no longer exists and has been permanently removed.'),
581 411: ('Length Required', 'Client must specify Content-Length.'),
582 412: ('Precondition Failed', 'Precondition in headers is false.'),
583 413: ('Request Entity Too Large', 'Entity is too large.'),
584 414: ('Request-URI Too Long', 'URI is too long.'),
585 415: ('Unsupported Media Type', 'Entity body in unsupported format.'),
586 416: ('Requested Range Not Satisfiable',
587 'Cannot satisfy request range.'),
588 417: ('Expectation Failed',
589 'Expect condition could not be satisfied.'),
590
591 500: ('Internal Server Error', 'Server got itself in trouble'),
592 501: ('Not Implemented',
593 'Server does not support this operation'),
594 502: ('Bad Gateway', 'Invalid responses from another server/proxy.'),
595 503: ('Service Unavailable',
596 'The server cannot process the request due to a high load'),
597 504: ('Gateway Timeout',
598 'The gateway server did not receive a timely response'),
599 505: ('HTTP Version Not Supported', 'Cannot fulfill request.'),
600 }
601
602
603class SimpleHTTPRequestHandler(BaseHTTPRequestHandler):
604
605 """Simple HTTP request handler with GET and HEAD commands.
606
607 This serves files from the current directory and any of its
608 subdirectories. The MIME type for files is determined by
609 calling the .guess_type() method.
610
611 The GET and HEAD requests are identical except that the HEAD
612 request omits the actual contents of the file.
613
614 """
615
616 server_version = "SimpleHTTP/" + __version__
617
618 def do_GET(self):
619 """Serve a GET request."""
620 f = self.send_head()
621 if f:
622 self.copyfile(f, self.wfile)
623 f.close()
624
625 def do_HEAD(self):
626 """Serve a HEAD request."""
627 f = self.send_head()
628 if f:
629 f.close()
630
631 def send_head(self):
632 """Common code for GET and HEAD commands.
633
634 This sends the response code and MIME headers.
635
636 Return value is either a file object (which has to be copied
637 to the outputfile by the caller unless the command was HEAD,
638 and must be closed by the caller under all circumstances), or
639 None, in which case the caller has nothing further to do.
640
641 """
642 path = self.translate_path(self.path)
643 f = None
644 if os.path.isdir(path):
645 if not self.path.endswith('/'):
646 # redirect browser - doing basically what apache does
647 self.send_response(301)
648 self.send_header("Location", self.path + "/")
649 self.end_headers()
650 return None
651 for index in "index.html", "index.htm":
652 index = os.path.join(path, index)
653 if os.path.exists(index):
654 path = index
655 break
656 else:
657 return self.list_directory(path)
658 ctype = self.guess_type(path)
659 try:
660 f = open(path, 'rb')
661 except IOError:
662 self.send_error(404, "File not found")
663 return None
664 self.send_response(200)
665 self.send_header("Content-type", ctype)
666 fs = os.fstat(f.fileno())
667 self.send_header("Content-Length", str(fs[6]))
668 self.send_header("Last-Modified", self.date_time_string(fs.st_mtime))
669 self.end_headers()
670 return f
671
672 def list_directory(self, path):
673 """Helper to produce a directory listing (absent index.html).
674
675 Return value is either a file object, or None (indicating an
676 error). In either case, the headers are sent, making the
677 interface the same as for send_head().
678
679 """
680 try:
681 list = os.listdir(path)
682 except os.error:
683 self.send_error(404, "No permission to list directory")
684 return None
685 list.sort(key=lambda a: a.lower())
686 r = []
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]687 displaypath = cgi.escape(urllib.parse.unquote(self.path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]688 r.append('<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">')
689 r.append("<html>\n<title>Directory listing for %s</title>\n" % displaypath)
690 r.append("<body>\n<h2>Directory listing for %s</h2>\n" % displaypath)
691 r.append("<hr>\n<ul>\n")
692 for name in list:
693 fullname = os.path.join(path, name)
694 displayname = linkname = name
695 # Append / for directories or @ for symbolic links
696 if os.path.isdir(fullname):
697 displayname = name + "/"
698 linkname = name + "/"
699 if os.path.islink(fullname):
700 displayname = name + "@"
701 # Note: a link to a directory displays with @ and links with /
702 r.append('<li><a href="%s">%s</a>\n'
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]703 % (urllib.parse.quote(linkname), cgi.escape(displayname)))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]704 r.append("</ul>\n<hr>\n</body>\n</html>\n")
705 enc = sys.getfilesystemencoding()
706 encoded = ''.join(r).encode(enc)
707 f = io.BytesIO()
708 f.write(encoded)
709 f.seek(0)
710 self.send_response(200)
711 self.send_header("Content-type", "text/html; charset=%s" % enc)
712 self.send_header("Content-Length", str(len(encoded)))
713 self.end_headers()
714 return f
715
716 def translate_path(self, path):
717 """Translate a /-separated PATH to the local filename syntax.
718
719 Components that mean special things to the local file system
720 (e.g. drive or directory names) are ignored. (XXX They should
721 probably be diagnosed.)
722
723 """
724 # abandon query parameters
725 path = path.split('?',1)[0]
726 path = path.split('#',1)[0]
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]727 path = posixpath.normpath(urllib.parse.unquote(path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]728 words = path.split('/')
729 words = filter(None, words)
730 path = os.getcwd()
731 for word in words:
732 drive, word = os.path.splitdrive(word)
733 head, word = os.path.split(word)
734 if word in (os.curdir, os.pardir): continue
735 path = os.path.join(path, word)
736 return path
737
738 def copyfile(self, source, outputfile):
739 """Copy all data between two file objects.
740
741 The SOURCE argument is a file object open for reading
742 (or anything with a read() method) and the DESTINATION
743 argument is a file object open for writing (or
744 anything with a write() method).
745
746 The only reason for overriding this would be to change
747 the block size or perhaps to replace newlines by CRLF
748 -- note however that this the default server uses this
749 to copy binary data as well.
750
751 """
752 shutil.copyfileobj(source, outputfile)
753
754 def guess_type(self, path):
755 """Guess the type of a file.
756
757 Argument is a PATH (a filename).
758
759 Return value is a string of the form type/subtype,
760 usable for a MIME Content-type header.
761
762 The default implementation looks the file's extension
763 up in the table self.extensions_map, using application/octet-stream
764 as a default; however it would be permissible (if
765 slow) to look inside the data to make a better guess.
766
767 """
768
769 base, ext = posixpath.splitext(path)
770 if ext in self.extensions_map:
771 return self.extensions_map[ext]
772 ext = ext.lower()
773 if ext in self.extensions_map:
774 return self.extensions_map[ext]
775 else:
776 return self.extensions_map['']
777
778 if not mimetypes.inited:
779 mimetypes.init() # try to read system mime.types
780 extensions_map = mimetypes.types_map.copy()
781 extensions_map.update({
782 '': 'application/octet-stream', # Default
783 '.py': 'text/plain',
784 '.c': 'text/plain',
785 '.h': 'text/plain',
786 })
787
788
789# Utilities for CGIHTTPRequestHandler
790
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]791# TODO(gregory.p.smith): Move this into an appropriate library.
792def _url_collapse_path_split(path):
793 """
794 Given a URL path, remove extra '/'s and '.' path elements and collapse
795 any '..' references.
796
797 Implements something akin to RFC-2396 5.2 step 6 to parse relative paths.
798
799 Returns: A tuple of (head, tail) where tail is everything after the final /
800 and head is everything before it. Head will always start with a '/' and,
801 if it contains anything else, never have a trailing '/'.
802
803 Raises: IndexError if too many '..' occur within the path.
804 """
805 # Similar to os.path.split(os.path.normpath(path)) but specific to URL
806 # path semantics rather than local operating system semantics.
807 path_parts = []
808 for part in path.split('/'):
809 if part == '.':
810 path_parts.append('')
811 else:
812 path_parts.append(part)
813 # Filter out blank non trailing parts before consuming the '..'.
814 path_parts = [part for part in path_parts[:-1] if part] + path_parts[-1:]
815 if path_parts:
816 tail_part = path_parts.pop()
817 else:
818 tail_part = ''
819 head_parts = []
820 for part in path_parts:
821 if part == '..':
822 head_parts.pop()
823 else:
824 head_parts.append(part)
825 if tail_part and tail_part == '..':
826 head_parts.pop()
827 tail_part = ''
828 return ('/' + '/'.join(head_parts), tail_part)
829
830
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]831nobody = None
832
833def nobody_uid():
834 """Internal routine to get nobody's uid"""
835 global nobody
836 if nobody:
837 return nobody
838 try:
839 import pwd
840 except ImportError:
841 return -1
842 try:
843 nobody = pwd.getpwnam('nobody')[2]
844 except KeyError:
845 nobody = 1 + max(map(lambda x: x[2], pwd.getpwall()))
846 return nobody
847
848
849def executable(path):
850 """Test for executable file."""
851 try:
852 st = os.stat(path)
853 except os.error:
854 return False
855 return st.st_mode & 0o111 != 0
856
857
858class CGIHTTPRequestHandler(SimpleHTTPRequestHandler):
859
860 """Complete HTTP server with GET, HEAD and POST commands.
861
862 GET and HEAD also support running CGI scripts.
863
864 The POST command is *only* implemented for CGI scripts.
865
866 """
867
868 # Determine platform specifics
869 have_fork = hasattr(os, 'fork')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]870
871 # Make rfile unbuffered -- we need to read one line and then pass
872 # the rest to a subprocess, so we can't use buffered input.
873 rbufsize = 0
874
875 def do_POST(self):
876 """Serve a POST request.
877
878 This is only implemented for CGI scripts.
879
880 """
881
882 if self.is_cgi():
883 self.run_cgi()
884 else:
885 self.send_error(501, "Can only POST to CGI scripts")
886
887 def send_head(self):
888 """Version of send_head that support CGI scripts"""
889 if self.is_cgi():
890 return self.run_cgi()
891 else:
892 return SimpleHTTPRequestHandler.send_head(self)
893
894 def is_cgi(self):
895 """Test whether self.path corresponds to a CGI script.
896
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]897 Returns True and updates the cgi_info attribute to the tuple
898 (dir, rest) if self.path requires running a CGI script.
899 Returns False otherwise.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]900
Benjamin Petersona7deeee2009-05-08 20:54:42 +0000[diff] [blame]901 If any exception is raised, the caller should assume that
902 self.path was rejected as invalid and act accordingly.
903
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]904 The default implementation tests whether the normalized url
905 path begins with one of the strings in self.cgi_directories
906 (and the next character is a '/' or the end of the string).
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]907
908 """
909
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]910 splitpath = _url_collapse_path_split(self.path)
911 if splitpath[0] in self.cgi_directories:
912 self.cgi_info = splitpath
913 return True
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]914 return False
915
916 cgi_directories = ['/cgi-bin', '/htbin']
917
918 def is_executable(self, path):
919 """Test whether argument path is an executable file."""
920 return executable(path)
921
922 def is_python(self, path):
923 """Test whether argument path is a Python script."""
924 head, tail = os.path.splitext(path)
925 return tail.lower() in (".py", ".pyw")
926
927 def run_cgi(self):
928 """Execute a CGI script."""
929 path = self.path
930 dir, rest = self.cgi_info
931
932 i = path.find('/', len(dir) + 1)
933 while i >= 0:
934 nextdir = path[:i]
935 nextrest = path[i+1:]
936
937 scriptdir = self.translate_path(nextdir)
938 if os.path.isdir(scriptdir):
939 dir, rest = nextdir, nextrest
940 i = path.find('/', len(dir) + 1)
941 else:
942 break
943
944 # find an explicit query string, if present.
945 i = rest.rfind('?')
946 if i >= 0:
947 rest, query = rest[:i], rest[i+1:]
948 else:
949 query = ''
950
951 # dissect the part after the directory name into a script name &
952 # a possible additional path, to be stored in PATH_INFO.
953 i = rest.find('/')
954 if i >= 0:
955 script, rest = rest[:i], rest[i:]
956 else:
957 script, rest = rest, ''
958
959 scriptname = dir + '/' + script
960 scriptfile = self.translate_path(scriptname)
961 if not os.path.exists(scriptfile):
962 self.send_error(404, "No such CGI script (%r)" % scriptname)
963 return
964 if not os.path.isfile(scriptfile):
965 self.send_error(403, "CGI script is not a plain file (%r)" %
966 scriptname)
967 return
968 ispy = self.is_python(scriptname)
969 if not ispy:
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]970 if not self.is_executable(scriptfile):
971 self.send_error(403, "CGI script is not executable (%r)" %
972 scriptname)
973 return
974
975 # Reference: http://hoohoo.ncsa.uiuc.edu/cgi/env.html
976 # XXX Much of the following could be prepared ahead of time!
Senthil Kumaran5e8826c2010-10-03 18:04:52 +0000[diff] [blame]977 env = copy.deepcopy(os.environ)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]978 env['SERVER_SOFTWARE'] = self.version_string()
979 env['SERVER_NAME'] = self.server.server_name
980 env['GATEWAY_INTERFACE'] = 'CGI/1.1'
981 env['SERVER_PROTOCOL'] = self.protocol_version
982 env['SERVER_PORT'] = str(self.server.server_port)
983 env['REQUEST_METHOD'] = self.command
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]984 uqrest = urllib.parse.unquote(rest)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]985 env['PATH_INFO'] = uqrest
986 env['PATH_TRANSLATED'] = self.translate_path(uqrest)
987 env['SCRIPT_NAME'] = scriptname
988 if query:
989 env['QUERY_STRING'] = query
990 host = self.address_string()
991 if host != self.client_address[0]:
992 env['REMOTE_HOST'] = host
993 env['REMOTE_ADDR'] = self.client_address[0]
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]994 authorization = self.headers.get("authorization")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]995 if authorization:
996 authorization = authorization.split()
997 if len(authorization) == 2:
998 import base64, binascii
999 env['AUTH_TYPE'] = authorization[0]
1000 if authorization[0].lower() == "basic":
1001 try:
1002 authorization = authorization[1].encode('ascii')
Georg Brandl706824f2009-06-04 09:42:55 +0000[diff] [blame]1003 authorization = base64.decodebytes(authorization).\
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1004 decode('ascii')
1005 except (binascii.Error, UnicodeError):
1006 pass
1007 else:
1008 authorization = authorization.split(':')
1009 if len(authorization) == 2:
1010 env['REMOTE_USER'] = authorization[0]
1011 # XXX REMOTE_IDENT
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1012 if self.headers.get('content-type') is None:
1013 env['CONTENT_TYPE'] = self.headers.get_content_type()
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1014 else:
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1015 env['CONTENT_TYPE'] = self.headers['content-type']
1016 length = self.headers.get('content-length')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1017 if length:
1018 env['CONTENT_LENGTH'] = length
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1019 referer = self.headers.get('referer')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1020 if referer:
1021 env['HTTP_REFERER'] = referer
1022 accept = []
1023 for line in self.headers.getallmatchingheaders('accept'):
1024 if line[:1] in "\t\n\r ":
1025 accept.append(line.strip())
1026 else:
1027 accept = accept + line[7:].split(',')
1028 env['HTTP_ACCEPT'] = ','.join(accept)
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1029 ua = self.headers.get('user-agent')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1030 if ua:
1031 env['HTTP_USER_AGENT'] = ua
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1032 co = filter(None, self.headers.get_all('cookie', []))
Georg Brandlcaa78fe2010-08-01 19:07:28 +0000[diff] [blame]1033 cookie_str = ', '.join(co)
1034 if cookie_str:
1035 env['HTTP_COOKIE'] = cookie_str
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1036 # XXX Other HTTP_* headers
1037 # Since we're setting the env in the parent, provide empty
1038 # values to override previously set values
1039 for k in ('QUERY_STRING', 'REMOTE_HOST', 'CONTENT_LENGTH',
1040 'HTTP_USER_AGENT', 'HTTP_COOKIE', 'HTTP_REFERER'):
1041 env.setdefault(k, "")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1042
1043 self.send_response(200, "Script output follows")
1044
1045 decoded_query = query.replace('+', ' ')
1046
1047 if self.have_fork:
1048 # Unix -- fork as we should
1049 args = [script]
1050 if '=' not in decoded_query:
1051 args.append(decoded_query)
1052 nobody = nobody_uid()
1053 self.wfile.flush() # Always flush before forking
1054 pid = os.fork()
1055 if pid != 0:
1056 # Parent
1057 pid, sts = os.waitpid(pid, 0)
1058 # throw away additional data [see bug #427345]
1059 while select.select([self.rfile], [], [], 0)[0]:
1060 if not self.rfile.read(1):
1061 break
1062 if sts:
1063 self.log_error("CGI script exit status %#x", sts)
1064 return
1065 # Child
1066 try:
1067 try:
1068 os.setuid(nobody)
1069 except os.error:
1070 pass
1071 os.dup2(self.rfile.fileno(), 0)
1072 os.dup2(self.wfile.fileno(), 1)
Senthil Kumaran5e8826c2010-10-03 18:04:52 +0000[diff] [blame]1073 os.execve(scriptfile, args, env)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1074 except:
1075 self.server.handle_error(self.request, self.client_address)
1076 os._exit(127)
1077
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1078 else:
1079 # Non-Unix -- use subprocess
1080 import subprocess
Senthil Kumaranca5130c2009-11-11 04:21:22 +0000[diff] [blame]1081 cmdline = [scriptfile]
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1082 if self.is_python(scriptfile):
1083 interp = sys.executable
1084 if interp.lower().endswith("w.exe"):
1085 # On Windows, use python.exe, not pythonw.exe
1086 interp = interp[:-5] + interp[-4:]
Senthil Kumaranca5130c2009-11-11 04:21:22 +0000[diff] [blame]1087 cmdline = [interp, '-u'] + cmdline
1088 if '=' not in query:
1089 cmdline.append(query)
1090 self.log_message("command: %s", subprocess.list2cmdline(cmdline))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1091 try:
1092 nbytes = int(length)
1093 except (TypeError, ValueError):
1094 nbytes = 0
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1095 p = subprocess.Popen(cmdline,
1096 stdin=subprocess.PIPE,
1097 stdout=subprocess.PIPE,
Senthil Kumaran5e8826c2010-10-03 18:04:52 +0000[diff] [blame]1098 stderr=subprocess.PIPE,
1099 env = env
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1100 )
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1101 if self.command.lower() == "post" and nbytes > 0:
1102 data = self.rfile.read(nbytes)
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1103 else:
1104 data = None
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1105 # throw away additional data [see bug #427345]
1106 while select.select([self.rfile._sock], [], [], 0)[0]:
1107 if not self.rfile._sock.recv(1):
1108 break
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1109 stdout, stderr = p.communicate(data)
1110 self.wfile.write(stdout)
1111 if stderr:
1112 self.log_error('%s', stderr)
Brian Curtin938ece72010-11-05 15:08:19 +0000[diff] [blame]1113 p.stderr.close()
1114 p.stdout.close()
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1115 status = p.returncode
1116 if status:
1117 self.log_error("CGI script exit status %#x", status)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1118 else:
1119 self.log_message("CGI script exited OK")
1120
1121
1122def test(HandlerClass = BaseHTTPRequestHandler,
1123 ServerClass = HTTPServer, protocol="HTTP/1.0"):
1124 """Test the HTTP request handler class.
1125
1126 This runs an HTTP server on port 8000 (or the first command line
1127 argument).
1128
1129 """
1130
1131 if sys.argv[1:]:
1132 port = int(sys.argv[1])
1133 else:
1134 port = 8000
1135 server_address = ('', port)
1136
1137 HandlerClass.protocol_version = protocol
1138 httpd = ServerClass(server_address, HandlerClass)
1139
1140 sa = httpd.socket.getsockname()
1141 print("Serving HTTP on", sa[0], "port", sa[1], "...")
Alexandre Vassalottib5292a22009-04-03 07:16:55 +0000[diff] [blame]1142 try:
1143 httpd.serve_forever()
1144 except KeyboardInterrupt:
1145 print("\nKeyboard interrupt received, exiting.")
1146 httpd.server_close()
1147 sys.exit(0)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1148
1149if __name__ == '__main__':
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1150 test(HandlerClass=SimpleHTTPRequestHandler)