Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / python / cpython2 / 388f956c35d335446c4715c8f2c28631f3c54ab1 / . / Lib / http / server.py
blob: 4fa58a259e391d4899fac524d3f144530644831d [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
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]102
103# Default error message template
104DEFAULT_ERROR_MESSAGE = """\
105<head>
106<title>Error response</title>
107</head>
108<body>
109<h1>Error response</h1>
110<p>Error code %(code)d.
111<p>Message: %(message)s.
112<p>Error code explanation: %(code)s = %(explain)s.
113</body>
114"""
115
116DEFAULT_ERROR_CONTENT_TYPE = "text/html;charset=utf-8"
117
118def _quote_html(html):
119 return html.replace("&", "&amp;").replace("<", "&lt;").replace(">", "&gt;")
120
121class HTTPServer(socketserver.TCPServer):
122
123 allow_reuse_address = 1 # Seems to make sense in testing environment
124
125 def server_bind(self):
126 """Override server_bind to store the server name."""
127 socketserver.TCPServer.server_bind(self)
128 host, port = self.socket.getsockname()[:2]
129 self.server_name = socket.getfqdn(host)
130 self.server_port = port
131
132
133class BaseHTTPRequestHandler(socketserver.StreamRequestHandler):
134
135 """HTTP request handler base class.
136
137 The following explanation of HTTP serves to guide you through the
138 code as well as to expose any misunderstandings I may have about
139 HTTP (so you don't need to read the code to figure out I'm wrong
140 :-).
141
142 HTTP (HyperText Transfer Protocol) is an extensible protocol on
143 top of a reliable stream transport (e.g. TCP/IP). The protocol
144 recognizes three parts to a request:
145
146 1. One line identifying the request type and path
147 2. An optional set of RFC-822-style headers
148 3. An optional data part
149
150 The headers and data are separated by a blank line.
151
152 The first line of the request has the form
153
154 <command> <path> <version>
155
156 where <command> is a (case-sensitive) keyword such as GET or POST,
157 <path> is a string containing path information for the request,
158 and <version> should be the string "HTTP/1.0" or "HTTP/1.1".
159 <path> is encoded using the URL encoding scheme (using %xx to signify
160 the ASCII character with hex code xx).
161
162 The specification specifies that lines are separated by CRLF but
163 for compatibility with the widest range of clients recommends
164 servers also handle LF. Similarly, whitespace in the request line
165 is treated sensibly (allowing multiple spaces between components
166 and allowing trailing whitespace).
167
168 Similarly, for output, lines ought to be separated by CRLF pairs
169 but most clients grok LF characters just fine.
170
171 If the first line of the request has the form
172
173 <command> <path>
174
175 (i.e. <version> is left out) then this is assumed to be an HTTP
176 0.9 request; this form has no optional headers and data part and
177 the reply consists of just the data.
178
179 The reply form of the HTTP 1.x protocol again has three parts:
180
181 1. One line giving the response code
182 2. An optional set of RFC-822-style headers
183 3. The data
184
185 Again, the headers and data are separated by a blank line.
186
187 The response code line has the form
188
189 <version> <responsecode> <responsestring>
190
191 where <version> is the protocol version ("HTTP/1.0" or "HTTP/1.1"),
192 <responsecode> is a 3-digit response code indicating success or
193 failure of the request, and <responsestring> is an optional
194 human-readable string explaining what the response code means.
195
196 This server parses the request and the headers, and then calls a
197 function specific to the request type (<command>). Specifically,
198 a request SPAM will be handled by a method do_SPAM(). If no
199 such method exists the server sends an error response to the
200 client. If it exists, it is called with no arguments:
201
202 do_SPAM()
203
204 Note that the request name is case sensitive (i.e. SPAM and spam
205 are different requests).
206
207 The various request details are stored in instance variables:
208
209 - client_address is the client IP address in the form (host,
210 port);
211
212 - command, path and version are the broken-down request line;
213
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]214 - headers is an instance of email.message.Message (or a derived
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]215 class) containing the header information;
216
217 - rfile is a file object open for reading positioned at the
218 start of the optional input data part;
219
220 - wfile is a file object open for writing.
221
222 IT IS IMPORTANT TO ADHERE TO THE PROTOCOL FOR WRITING!
223
224 The first thing to be written must be the response line. Then
225 follow 0 or more header lines, then a blank line, and then the
226 actual data (if any). The meaning of the header lines depends on
227 the command executed by the server; in most cases, when data is
228 returned, there should be at least one header line of the form
229
230 Content-type: <type>/<subtype>
231
232 where <type> and <subtype> should be registered MIME types,
233 e.g. "text/html" or "text/plain".
234
235 """
236
237 # The Python system version, truncated to its first component.
238 sys_version = "Python/" + sys.version.split()[0]
239
240 # The server software version. You may want to override this.
241 # The format is multiple whitespace-separated strings,
242 # where each string is of the form name[/version].
243 server_version = "BaseHTTP/" + __version__
244
245 error_message_format = DEFAULT_ERROR_MESSAGE
246 error_content_type = DEFAULT_ERROR_CONTENT_TYPE
247
248 # The default request version. This only affects responses up until
249 # the point where the request line is parsed, so it mainly decides what
250 # the client gets back when sending a malformed request line.
251 # Most web servers default to HTTP 0.9, i.e. don't send a status line.
252 default_request_version = "HTTP/0.9"
253
254 def parse_request(self):
255 """Parse a request (internal).
256
257 The request should be stored in self.raw_requestline; the results
258 are in self.command, self.path, self.request_version and
259 self.headers.
260
261 Return True for success, False for failure; on failure, an
262 error is sent back.
263
264 """
265 self.command = None # set in case of error on the first line
266 self.request_version = version = self.default_request_version
267 self.close_connection = 1
268 requestline = str(self.raw_requestline, 'iso-8859-1')
269 if requestline[-2:] == '\r\n':
270 requestline = requestline[:-2]
271 elif requestline[-1:] == '\n':
272 requestline = requestline[:-1]
273 self.requestline = requestline
274 words = requestline.split()
275 if len(words) == 3:
276 [command, path, version] = words
277 if version[:5] != 'HTTP/':
278 self.send_error(400, "Bad request version (%r)" % version)
279 return False
280 try:
281 base_version_number = version.split('/', 1)[1]
282 version_number = base_version_number.split(".")
283 # RFC 2145 section 3.1 says there can be only one "." and
284 # - major and minor numbers MUST be treated as
285 # separate integers;
286 # - HTTP/2.4 is a lower version than HTTP/2.13, which in
287 # turn is lower than HTTP/12.3;
288 # - Leading zeros MUST be ignored by recipients.
289 if len(version_number) != 2:
290 raise ValueError
291 version_number = int(version_number[0]), int(version_number[1])
292 except (ValueError, IndexError):
293 self.send_error(400, "Bad request version (%r)" % version)
294 return False
295 if version_number >= (1, 1) and self.protocol_version >= "HTTP/1.1":
296 self.close_connection = 0
297 if version_number >= (2, 0):
298 self.send_error(505,
299 "Invalid HTTP Version (%s)" % base_version_number)
300 return False
301 elif len(words) == 2:
302 [command, path] = words
303 self.close_connection = 1
304 if command != 'GET':
305 self.send_error(400,
306 "Bad HTTP/0.9 request type (%r)" % command)
307 return False
308 elif not words:
309 return False
310 else:
311 self.send_error(400, "Bad request syntax (%r)" % requestline)
312 return False
313 self.command, self.path, self.request_version = command, path, version
314
315 # Examine the headers and look for a Connection directive.
Jeremy Hylton98eb6c22009-03-27 18:31:36 +0000[diff] [blame]316 self.headers = http.client.parse_headers(self.rfile,
317 _class=self.MessageClass)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]318
319 conntype = self.headers.get('Connection', "")
320 if conntype.lower() == 'close':
321 self.close_connection = 1
322 elif (conntype.lower() == 'keep-alive' and
323 self.protocol_version >= "HTTP/1.1"):
324 self.close_connection = 0
Senthil Kumaran0f476d42010-09-30 06:09:18 +0000[diff] [blame]325 # Examine the headers and look for an Expect directive
326 expect = self.headers.get('Expect', "")
327 if (expect.lower() == "100-continue" and
328 self.protocol_version >= "HTTP/1.1" and
329 self.request_version >= "HTTP/1.1"):
330 if not self.handle_expect_100():
331 return False
332 return True
333
334 def handle_expect_100(self):
335 """Decide what to do with an "Expect: 100-continue" header.
336
337 If the client is expecting a 100 Continue response, we must
338 respond with either a 100 Continue or a final response before
339 waiting for the request body. The default is to always respond
340 with a 100 Continue. You can behave differently (for example,
341 reject unauthorized requests) by overriding this method.
342
343 This method should either return True (possibly after sending
344 a 100 Continue response) or send an error response and return
345 False.
346
347 """
348 self.send_response_only(100)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]349 return True
350
351 def handle_one_request(self):
352 """Handle a single HTTP request.
353
354 You normally don't need to override this method; see the class
355 __doc__ string for information on how to handle specific HTTP
356 commands such as GET and POST.
357
358 """
Kristján Valur Jónsson985fc6a2009-07-01 10:01:31 +0000[diff] [blame]359 try:
360 self.raw_requestline = self.rfile.readline()
361 if not self.raw_requestline:
362 self.close_connection = 1
363 return
364 if not self.parse_request():
365 # An error code has been sent, just exit
366 return
367 mname = 'do_' + self.command
368 if not hasattr(self, mname):
369 self.send_error(501, "Unsupported method (%r)" % self.command)
370 return
371 method = getattr(self, mname)
372 method()
373 self.wfile.flush() #actually send the response if not already done.
374 except socket.timeout as e:
375 #a read or a write timed out. Discard this connection
376 self.log_error("Request timed out: %r", e)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]377 self.close_connection = 1
378 return
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]379
380 def handle(self):
381 """Handle multiple requests if necessary."""
382 self.close_connection = 1
383
384 self.handle_one_request()
385 while not self.close_connection:
386 self.handle_one_request()
387
388 def send_error(self, code, message=None):
389 """Send and log an error reply.
390
391 Arguments are the error code, and a detailed message.
392 The detailed message defaults to the short entry matching the
393 response code.
394
395 This sends an error response (so it must be called before any
396 output has been generated), logs the error, and finally sends
397 a piece of HTML explaining the error to the user.
398
399 """
400
401 try:
402 shortmsg, longmsg = self.responses[code]
403 except KeyError:
404 shortmsg, longmsg = '???', '???'
405 if message is None:
406 message = shortmsg
407 explain = longmsg
408 self.log_error("code %d, message %s", code, message)
409 # using _quote_html to prevent Cross Site Scripting attacks (see bug #1100201)
410 content = (self.error_message_format %
411 {'code': code, 'message': _quote_html(message), 'explain': explain})
412 self.send_response(code, message)
413 self.send_header("Content-Type", self.error_content_type)
414 self.send_header('Connection', 'close')
415 self.end_headers()
416 if self.command != 'HEAD' and code >= 200 and code not in (204, 304):
417 self.wfile.write(content.encode('UTF-8', 'replace'))
418
419 def send_response(self, code, message=None):
420 """Send the response header and log the response code.
421
422 Also send two standard headers with the server software
423 version and the current date.
424
425 """
426 self.log_request(code)
Senthil Kumaran0f476d42010-09-30 06:09:18 +0000[diff] [blame]427 self.send_response_only(code, message)
428 self.send_header('Server', self.version_string())
429 self.send_header('Date', self.date_time_string())
430
431 def send_response_only(self, code, message=None):
432 """Send the response header only."""
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]433 if message is None:
434 if code in self.responses:
435 message = self.responses[code][0]
436 else:
437 message = ''
438 if self.request_version != 'HTTP/0.9':
439 self.wfile.write(("%s %d %s\r\n" %
440 (self.protocol_version, code, message)).encode('ASCII', 'strict'))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]441
442 def send_header(self, keyword, value):
443 """Send a MIME header."""
444 if self.request_version != 'HTTP/0.9':
445 self.wfile.write(("%s: %s\r\n" % (keyword, value)).encode('ASCII', 'strict'))
446
447 if keyword.lower() == 'connection':
448 if value.lower() == 'close':
449 self.close_connection = 1
450 elif value.lower() == 'keep-alive':
451 self.close_connection = 0
452
453 def end_headers(self):
454 """Send the blank line ending the MIME headers."""
455 if self.request_version != 'HTTP/0.9':
456 self.wfile.write(b"\r\n")
457
458 def log_request(self, code='-', size='-'):
459 """Log an accepted request.
460
461 This is called by send_response().
462
463 """
464
465 self.log_message('"%s" %s %s',
466 self.requestline, str(code), str(size))
467
468 def log_error(self, format, *args):
469 """Log an error.
470
471 This is called when a request cannot be fulfilled. By
472 default it passes the message on to log_message().
473
474 Arguments are the same as for log_message().
475
476 XXX This should go to the separate error log.
477
478 """
479
480 self.log_message(format, *args)
481
482 def log_message(self, format, *args):
483 """Log an arbitrary message.
484
485 This is used by all other logging functions. Override
486 it if you have specific logging wishes.
487
488 The first argument, FORMAT, is a format string for the
489 message to be logged. If the format string contains
490 any % escapes requiring parameters, they should be
491 specified as subsequent arguments (it's just like
492 printf!).
493
494 The client host and current date/time are prefixed to
495 every message.
496
497 """
498
499 sys.stderr.write("%s - - [%s] %s\n" %
500 (self.address_string(),
501 self.log_date_time_string(),
502 format%args))
503
504 def version_string(self):
505 """Return the server software version string."""
506 return self.server_version + ' ' + self.sys_version
507
508 def date_time_string(self, timestamp=None):
509 """Return the current date and time formatted for a message header."""
510 if timestamp is None:
511 timestamp = time.time()
512 year, month, day, hh, mm, ss, wd, y, z = time.gmtime(timestamp)
513 s = "%s, %02d %3s %4d %02d:%02d:%02d GMT" % (
514 self.weekdayname[wd],
515 day, self.monthname[month], year,
516 hh, mm, ss)
517 return s
518
519 def log_date_time_string(self):
520 """Return the current time formatted for logging."""
521 now = time.time()
522 year, month, day, hh, mm, ss, x, y, z = time.localtime(now)
523 s = "%02d/%3s/%04d %02d:%02d:%02d" % (
524 day, self.monthname[month], year, hh, mm, ss)
525 return s
526
527 weekdayname = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
528
529 monthname = [None,
530 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun',
531 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
532
533 def address_string(self):
534 """Return the client address formatted for logging.
535
536 This version looks up the full hostname using gethostbyaddr(),
537 and tries to find a name that contains at least one dot.
538
539 """
540
541 host, port = self.client_address[:2]
542 return socket.getfqdn(host)
543
544 # Essentially static class variables
545
546 # The version of the HTTP protocol we support.
547 # Set this to HTTP/1.1 to enable automatic keepalive
548 protocol_version = "HTTP/1.0"
549
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]550 # MessageClass used to parse headers
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]551 MessageClass = http.client.HTTPMessage
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]552
553 # Table mapping response codes to messages; entries have the
554 # form {code: (shortmessage, longmessage)}.
555 # See RFC 2616.
556 responses = {
557 100: ('Continue', 'Request received, please continue'),
558 101: ('Switching Protocols',
559 'Switching to new protocol; obey Upgrade header'),
560
561 200: ('OK', 'Request fulfilled, document follows'),
562 201: ('Created', 'Document created, URL follows'),
563 202: ('Accepted',
564 'Request accepted, processing continues off-line'),
565 203: ('Non-Authoritative Information', 'Request fulfilled from cache'),
566 204: ('No Content', 'Request fulfilled, nothing follows'),
567 205: ('Reset Content', 'Clear input form for further input.'),
568 206: ('Partial Content', 'Partial content follows.'),
569
570 300: ('Multiple Choices',
571 'Object has several resources -- see URI list'),
572 301: ('Moved Permanently', 'Object moved permanently -- see URI list'),
573 302: ('Found', 'Object moved temporarily -- see URI list'),
574 303: ('See Other', 'Object moved -- see Method and URL list'),
575 304: ('Not Modified',
576 'Document has not changed since given time'),
577 305: ('Use Proxy',
578 'You must use proxy specified in Location to access this '
579 'resource.'),
580 307: ('Temporary Redirect',
581 'Object moved temporarily -- see URI list'),
582
583 400: ('Bad Request',
584 'Bad request syntax or unsupported method'),
585 401: ('Unauthorized',
586 'No permission -- see authorization schemes'),
587 402: ('Payment Required',
588 'No payment -- see charging schemes'),
589 403: ('Forbidden',
590 'Request forbidden -- authorization will not help'),
591 404: ('Not Found', 'Nothing matches the given URI'),
592 405: ('Method Not Allowed',
Senthil Kumaran7aa26212010-02-22 11:00:50 +0000[diff] [blame]593 'Specified method is invalid for this resource.'),
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]594 406: ('Not Acceptable', 'URI not available in preferred format.'),
595 407: ('Proxy Authentication Required', 'You must authenticate with '
596 'this proxy before proceeding.'),
597 408: ('Request Timeout', 'Request timed out; try again later.'),
598 409: ('Conflict', 'Request conflict.'),
599 410: ('Gone',
600 'URI no longer exists and has been permanently removed.'),
601 411: ('Length Required', 'Client must specify Content-Length.'),
602 412: ('Precondition Failed', 'Precondition in headers is false.'),
603 413: ('Request Entity Too Large', 'Entity is too large.'),
604 414: ('Request-URI Too Long', 'URI is too long.'),
605 415: ('Unsupported Media Type', 'Entity body in unsupported format.'),
606 416: ('Requested Range Not Satisfiable',
607 'Cannot satisfy request range.'),
608 417: ('Expectation Failed',
609 'Expect condition could not be satisfied.'),
610
611 500: ('Internal Server Error', 'Server got itself in trouble'),
612 501: ('Not Implemented',
613 'Server does not support this operation'),
614 502: ('Bad Gateway', 'Invalid responses from another server/proxy.'),
615 503: ('Service Unavailable',
616 'The server cannot process the request due to a high load'),
617 504: ('Gateway Timeout',
618 'The gateway server did not receive a timely response'),
619 505: ('HTTP Version Not Supported', 'Cannot fulfill request.'),
620 }
621
622
623class SimpleHTTPRequestHandler(BaseHTTPRequestHandler):
624
625 """Simple HTTP request handler with GET and HEAD commands.
626
627 This serves files from the current directory and any of its
628 subdirectories. The MIME type for files is determined by
629 calling the .guess_type() method.
630
631 The GET and HEAD requests are identical except that the HEAD
632 request omits the actual contents of the file.
633
634 """
635
636 server_version = "SimpleHTTP/" + __version__
637
638 def do_GET(self):
639 """Serve a GET request."""
640 f = self.send_head()
641 if f:
642 self.copyfile(f, self.wfile)
643 f.close()
644
645 def do_HEAD(self):
646 """Serve a HEAD request."""
647 f = self.send_head()
648 if f:
649 f.close()
650
651 def send_head(self):
652 """Common code for GET and HEAD commands.
653
654 This sends the response code and MIME headers.
655
656 Return value is either a file object (which has to be copied
657 to the outputfile by the caller unless the command was HEAD,
658 and must be closed by the caller under all circumstances), or
659 None, in which case the caller has nothing further to do.
660
661 """
662 path = self.translate_path(self.path)
663 f = None
664 if os.path.isdir(path):
665 if not self.path.endswith('/'):
666 # redirect browser - doing basically what apache does
667 self.send_response(301)
668 self.send_header("Location", self.path + "/")
669 self.end_headers()
670 return None
671 for index in "index.html", "index.htm":
672 index = os.path.join(path, index)
673 if os.path.exists(index):
674 path = index
675 break
676 else:
677 return self.list_directory(path)
678 ctype = self.guess_type(path)
679 try:
680 f = open(path, 'rb')
681 except IOError:
682 self.send_error(404, "File not found")
683 return None
684 self.send_response(200)
685 self.send_header("Content-type", ctype)
686 fs = os.fstat(f.fileno())
687 self.send_header("Content-Length", str(fs[6]))
688 self.send_header("Last-Modified", self.date_time_string(fs.st_mtime))
689 self.end_headers()
690 return f
691
692 def list_directory(self, path):
693 """Helper to produce a directory listing (absent index.html).
694
695 Return value is either a file object, or None (indicating an
696 error). In either case, the headers are sent, making the
697 interface the same as for send_head().
698
699 """
700 try:
701 list = os.listdir(path)
702 except os.error:
703 self.send_error(404, "No permission to list directory")
704 return None
705 list.sort(key=lambda a: a.lower())
706 r = []
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]707 displaypath = cgi.escape(urllib.parse.unquote(self.path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]708 r.append('<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">')
709 r.append("<html>\n<title>Directory listing for %s</title>\n" % displaypath)
710 r.append("<body>\n<h2>Directory listing for %s</h2>\n" % displaypath)
711 r.append("<hr>\n<ul>\n")
712 for name in list:
713 fullname = os.path.join(path, name)
714 displayname = linkname = name
715 # Append / for directories or @ for symbolic links
716 if os.path.isdir(fullname):
717 displayname = name + "/"
718 linkname = name + "/"
719 if os.path.islink(fullname):
720 displayname = name + "@"
721 # Note: a link to a directory displays with @ and links with /
722 r.append('<li><a href="%s">%s</a>\n'
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]723 % (urllib.parse.quote(linkname), cgi.escape(displayname)))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]724 r.append("</ul>\n<hr>\n</body>\n</html>\n")
725 enc = sys.getfilesystemencoding()
726 encoded = ''.join(r).encode(enc)
727 f = io.BytesIO()
728 f.write(encoded)
729 f.seek(0)
730 self.send_response(200)
731 self.send_header("Content-type", "text/html; charset=%s" % enc)
732 self.send_header("Content-Length", str(len(encoded)))
733 self.end_headers()
734 return f
735
736 def translate_path(self, path):
737 """Translate a /-separated PATH to the local filename syntax.
738
739 Components that mean special things to the local file system
740 (e.g. drive or directory names) are ignored. (XXX They should
741 probably be diagnosed.)
742
743 """
744 # abandon query parameters
745 path = path.split('?',1)[0]
746 path = path.split('#',1)[0]
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]747 path = posixpath.normpath(urllib.parse.unquote(path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]748 words = path.split('/')
749 words = filter(None, words)
750 path = os.getcwd()
751 for word in words:
752 drive, word = os.path.splitdrive(word)
753 head, word = os.path.split(word)
754 if word in (os.curdir, os.pardir): continue
755 path = os.path.join(path, word)
756 return path
757
758 def copyfile(self, source, outputfile):
759 """Copy all data between two file objects.
760
761 The SOURCE argument is a file object open for reading
762 (or anything with a read() method) and the DESTINATION
763 argument is a file object open for writing (or
764 anything with a write() method).
765
766 The only reason for overriding this would be to change
767 the block size or perhaps to replace newlines by CRLF
768 -- note however that this the default server uses this
769 to copy binary data as well.
770
771 """
772 shutil.copyfileobj(source, outputfile)
773
774 def guess_type(self, path):
775 """Guess the type of a file.
776
777 Argument is a PATH (a filename).
778
779 Return value is a string of the form type/subtype,
780 usable for a MIME Content-type header.
781
782 The default implementation looks the file's extension
783 up in the table self.extensions_map, using application/octet-stream
784 as a default; however it would be permissible (if
785 slow) to look inside the data to make a better guess.
786
787 """
788
789 base, ext = posixpath.splitext(path)
790 if ext in self.extensions_map:
791 return self.extensions_map[ext]
792 ext = ext.lower()
793 if ext in self.extensions_map:
794 return self.extensions_map[ext]
795 else:
796 return self.extensions_map['']
797
798 if not mimetypes.inited:
799 mimetypes.init() # try to read system mime.types
800 extensions_map = mimetypes.types_map.copy()
801 extensions_map.update({
802 '': 'application/octet-stream', # Default
803 '.py': 'text/plain',
804 '.c': 'text/plain',
805 '.h': 'text/plain',
806 })
807
808
809# Utilities for CGIHTTPRequestHandler
810
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]811# TODO(gregory.p.smith): Move this into an appropriate library.
812def _url_collapse_path_split(path):
813 """
814 Given a URL path, remove extra '/'s and '.' path elements and collapse
815 any '..' references.
816
817 Implements something akin to RFC-2396 5.2 step 6 to parse relative paths.
818
819 Returns: A tuple of (head, tail) where tail is everything after the final /
820 and head is everything before it. Head will always start with a '/' and,
821 if it contains anything else, never have a trailing '/'.
822
823 Raises: IndexError if too many '..' occur within the path.
824 """
825 # Similar to os.path.split(os.path.normpath(path)) but specific to URL
826 # path semantics rather than local operating system semantics.
827 path_parts = []
828 for part in path.split('/'):
829 if part == '.':
830 path_parts.append('')
831 else:
832 path_parts.append(part)
833 # Filter out blank non trailing parts before consuming the '..'.
834 path_parts = [part for part in path_parts[:-1] if part] + path_parts[-1:]
835 if path_parts:
836 tail_part = path_parts.pop()
837 else:
838 tail_part = ''
839 head_parts = []
840 for part in path_parts:
841 if part == '..':
842 head_parts.pop()
843 else:
844 head_parts.append(part)
845 if tail_part and tail_part == '..':
846 head_parts.pop()
847 tail_part = ''
848 return ('/' + '/'.join(head_parts), tail_part)
849
850
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]851nobody = None
852
853def nobody_uid():
854 """Internal routine to get nobody's uid"""
855 global nobody
856 if nobody:
857 return nobody
858 try:
859 import pwd
860 except ImportError:
861 return -1
862 try:
863 nobody = pwd.getpwnam('nobody')[2]
864 except KeyError:
865 nobody = 1 + max(map(lambda x: x[2], pwd.getpwall()))
866 return nobody
867
868
869def executable(path):
870 """Test for executable file."""
871 try:
872 st = os.stat(path)
873 except os.error:
874 return False
875 return st.st_mode & 0o111 != 0
876
877
878class CGIHTTPRequestHandler(SimpleHTTPRequestHandler):
879
880 """Complete HTTP server with GET, HEAD and POST commands.
881
882 GET and HEAD also support running CGI scripts.
883
884 The POST command is *only* implemented for CGI scripts.
885
886 """
887
888 # Determine platform specifics
889 have_fork = hasattr(os, 'fork')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]890
891 # Make rfile unbuffered -- we need to read one line and then pass
892 # the rest to a subprocess, so we can't use buffered input.
893 rbufsize = 0
894
895 def do_POST(self):
896 """Serve a POST request.
897
898 This is only implemented for CGI scripts.
899
900 """
901
902 if self.is_cgi():
903 self.run_cgi()
904 else:
905 self.send_error(501, "Can only POST to CGI scripts")
906
907 def send_head(self):
908 """Version of send_head that support CGI scripts"""
909 if self.is_cgi():
910 return self.run_cgi()
911 else:
912 return SimpleHTTPRequestHandler.send_head(self)
913
914 def is_cgi(self):
915 """Test whether self.path corresponds to a CGI script.
916
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]917 Returns True and updates the cgi_info attribute to the tuple
918 (dir, rest) if self.path requires running a CGI script.
919 Returns False otherwise.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]920
Benjamin Petersona7deeee2009-05-08 20:54:42 +0000[diff] [blame]921 If any exception is raised, the caller should assume that
922 self.path was rejected as invalid and act accordingly.
923
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]924 The default implementation tests whether the normalized url
925 path begins with one of the strings in self.cgi_directories
926 (and the next character is a '/' or the end of the string).
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]927
928 """
929
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]930 splitpath = _url_collapse_path_split(self.path)
931 if splitpath[0] in self.cgi_directories:
932 self.cgi_info = splitpath
933 return True
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]934 return False
935
936 cgi_directories = ['/cgi-bin', '/htbin']
937
938 def is_executable(self, path):
939 """Test whether argument path is an executable file."""
940 return executable(path)
941
942 def is_python(self, path):
943 """Test whether argument path is a Python script."""
944 head, tail = os.path.splitext(path)
945 return tail.lower() in (".py", ".pyw")
946
947 def run_cgi(self):
948 """Execute a CGI script."""
949 path = self.path
950 dir, rest = self.cgi_info
951
952 i = path.find('/', len(dir) + 1)
953 while i >= 0:
954 nextdir = path[:i]
955 nextrest = path[i+1:]
956
957 scriptdir = self.translate_path(nextdir)
958 if os.path.isdir(scriptdir):
959 dir, rest = nextdir, nextrest
960 i = path.find('/', len(dir) + 1)
961 else:
962 break
963
964 # find an explicit query string, if present.
965 i = rest.rfind('?')
966 if i >= 0:
967 rest, query = rest[:i], rest[i+1:]
968 else:
969 query = ''
970
971 # dissect the part after the directory name into a script name &
972 # a possible additional path, to be stored in PATH_INFO.
973 i = rest.find('/')
974 if i >= 0:
975 script, rest = rest[:i], rest[i:]
976 else:
977 script, rest = rest, ''
978
979 scriptname = dir + '/' + script
980 scriptfile = self.translate_path(scriptname)
981 if not os.path.exists(scriptfile):
982 self.send_error(404, "No such CGI script (%r)" % scriptname)
983 return
984 if not os.path.isfile(scriptfile):
985 self.send_error(403, "CGI script is not a plain file (%r)" %
986 scriptname)
987 return
988 ispy = self.is_python(scriptname)
989 if not ispy:
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]990 if not self.is_executable(scriptfile):
991 self.send_error(403, "CGI script is not executable (%r)" %
992 scriptname)
993 return
994
995 # Reference: http://hoohoo.ncsa.uiuc.edu/cgi/env.html
996 # XXX Much of the following could be prepared ahead of time!
997 env = {}
998 env['SERVER_SOFTWARE'] = self.version_string()
999 env['SERVER_NAME'] = self.server.server_name
1000 env['GATEWAY_INTERFACE'] = 'CGI/1.1'
1001 env['SERVER_PROTOCOL'] = self.protocol_version
1002 env['SERVER_PORT'] = str(self.server.server_port)
1003 env['REQUEST_METHOD'] = self.command
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]1004 uqrest = urllib.parse.unquote(rest)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1005 env['PATH_INFO'] = uqrest
1006 env['PATH_TRANSLATED'] = self.translate_path(uqrest)
1007 env['SCRIPT_NAME'] = scriptname
1008 if query:
1009 env['QUERY_STRING'] = query
1010 host = self.address_string()
1011 if host != self.client_address[0]:
1012 env['REMOTE_HOST'] = host
1013 env['REMOTE_ADDR'] = self.client_address[0]
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1014 authorization = self.headers.get("authorization")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1015 if authorization:
1016 authorization = authorization.split()
1017 if len(authorization) == 2:
1018 import base64, binascii
1019 env['AUTH_TYPE'] = authorization[0]
1020 if authorization[0].lower() == "basic":
1021 try:
1022 authorization = authorization[1].encode('ascii')
Georg Brandl706824f2009-06-04 09:42:55 +0000[diff] [blame]1023 authorization = base64.decodebytes(authorization).\
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1024 decode('ascii')
1025 except (binascii.Error, UnicodeError):
1026 pass
1027 else:
1028 authorization = authorization.split(':')
1029 if len(authorization) == 2:
1030 env['REMOTE_USER'] = authorization[0]
1031 # XXX REMOTE_IDENT
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1032 if self.headers.get('content-type') is None:
1033 env['CONTENT_TYPE'] = self.headers.get_content_type()
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1034 else:
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1035 env['CONTENT_TYPE'] = self.headers['content-type']
1036 length = self.headers.get('content-length')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1037 if length:
1038 env['CONTENT_LENGTH'] = length
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1039 referer = self.headers.get('referer')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1040 if referer:
1041 env['HTTP_REFERER'] = referer
1042 accept = []
1043 for line in self.headers.getallmatchingheaders('accept'):
1044 if line[:1] in "\t\n\r ":
1045 accept.append(line.strip())
1046 else:
1047 accept = accept + line[7:].split(',')
1048 env['HTTP_ACCEPT'] = ','.join(accept)
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1049 ua = self.headers.get('user-agent')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1050 if ua:
1051 env['HTTP_USER_AGENT'] = ua
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1052 co = filter(None, self.headers.get_all('cookie', []))
Georg Brandl62e2ca22010-07-31 21:54:24 +0000[diff] [blame]1053 cookie_str = ', '.join(co)
1054 if cookie_str:
1055 env['HTTP_COOKIE'] = cookie_str
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1056 # XXX Other HTTP_* headers
1057 # Since we're setting the env in the parent, provide empty
1058 # values to override previously set values
1059 for k in ('QUERY_STRING', 'REMOTE_HOST', 'CONTENT_LENGTH',
1060 'HTTP_USER_AGENT', 'HTTP_COOKIE', 'HTTP_REFERER'):
1061 env.setdefault(k, "")
1062 os.environ.update(env)
1063
1064 self.send_response(200, "Script output follows")
1065
1066 decoded_query = query.replace('+', ' ')
1067
1068 if self.have_fork:
1069 # Unix -- fork as we should
1070 args = [script]
1071 if '=' not in decoded_query:
1072 args.append(decoded_query)
1073 nobody = nobody_uid()
1074 self.wfile.flush() # Always flush before forking
1075 pid = os.fork()
1076 if pid != 0:
1077 # Parent
1078 pid, sts = os.waitpid(pid, 0)
1079 # throw away additional data [see bug #427345]
1080 while select.select([self.rfile], [], [], 0)[0]:
1081 if not self.rfile.read(1):
1082 break
1083 if sts:
1084 self.log_error("CGI script exit status %#x", sts)
1085 return
1086 # Child
1087 try:
1088 try:
1089 os.setuid(nobody)
1090 except os.error:
1091 pass
1092 os.dup2(self.rfile.fileno(), 0)
1093 os.dup2(self.wfile.fileno(), 1)
1094 os.execve(scriptfile, args, os.environ)
1095 except:
1096 self.server.handle_error(self.request, self.client_address)
1097 os._exit(127)
1098
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1099 else:
1100 # Non-Unix -- use subprocess
1101 import subprocess
Senthil Kumarane29cd162009-11-11 04:17:53 +0000[diff] [blame]1102 cmdline = [scriptfile]
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1103 if self.is_python(scriptfile):
1104 interp = sys.executable
1105 if interp.lower().endswith("w.exe"):
1106 # On Windows, use python.exe, not pythonw.exe
1107 interp = interp[:-5] + interp[-4:]
Senthil Kumarane29cd162009-11-11 04:17:53 +0000[diff] [blame]1108 cmdline = [interp, '-u'] + cmdline
1109 if '=' not in query:
1110 cmdline.append(query)
1111 self.log_message("command: %s", subprocess.list2cmdline(cmdline))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1112 try:
1113 nbytes = int(length)
1114 except (TypeError, ValueError):
1115 nbytes = 0
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1116 p = subprocess.Popen(cmdline,
1117 stdin=subprocess.PIPE,
1118 stdout=subprocess.PIPE,
Senthil Kumarane29cd162009-11-11 04:17:53 +0000[diff] [blame]1119 stderr=subprocess.PIPE
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1120 )
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1121 if self.command.lower() == "post" and nbytes > 0:
1122 data = self.rfile.read(nbytes)
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1123 else:
1124 data = None
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1125 # throw away additional data [see bug #427345]
1126 while select.select([self.rfile._sock], [], [], 0)[0]:
1127 if not self.rfile._sock.recv(1):
1128 break
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1129 stdout, stderr = p.communicate(data)
1130 self.wfile.write(stdout)
1131 if stderr:
1132 self.log_error('%s', stderr)
1133 status = p.returncode
1134 if status:
1135 self.log_error("CGI script exit status %#x", status)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1136 else:
1137 self.log_message("CGI script exited OK")
1138
1139
1140def test(HandlerClass = BaseHTTPRequestHandler,
1141 ServerClass = HTTPServer, protocol="HTTP/1.0"):
1142 """Test the HTTP request handler class.
1143
1144 This runs an HTTP server on port 8000 (or the first command line
1145 argument).
1146
1147 """
1148
1149 if sys.argv[1:]:
1150 port = int(sys.argv[1])
1151 else:
1152 port = 8000
1153 server_address = ('', port)
1154
1155 HandlerClass.protocol_version = protocol
1156 httpd = ServerClass(server_address, HandlerClass)
1157
1158 sa = httpd.socket.getsockname()
1159 print("Serving HTTP on", sa[0], "port", sa[1], "...")
Alexandre Vassalottib5292a22009-04-03 07:16:55 +0000[diff] [blame]1160 try:
1161 httpd.serve_forever()
1162 except KeyboardInterrupt:
1163 print("\nKeyboard interrupt received, exiting.")
1164 httpd.server_close()
1165 sys.exit(0)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1166
1167if __name__ == '__main__':
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1168 test(HandlerClass=SimpleHTTPRequestHandler)