Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / python / cpython2 / a73dc9d5e8ba2320a5aebf8c24ab665ccdfed5f8 / . / Lib / http / server.py
blob: 8a6ba70a8ab362ef35935615c848a7d92cc38beb [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:
361 self.raw_requestline = self.rfile.readline()
362 if not self.raw_requestline:
363 self.close_connection = 1
364 return
365 if not self.parse_request():
366 # An error code has been sent, just exit
367 return
368 mname = 'do_' + self.command
369 if not hasattr(self, mname):
370 self.send_error(501, "Unsupported method (%r)" % self.command)
371 return
372 method = getattr(self, mname)
373 method()
374 self.wfile.flush() #actually send the response if not already done.
375 except socket.timeout as e:
376 #a read or a write timed out. Discard this connection
377 self.log_error("Request timed out: %r", e)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]378 self.close_connection = 1
379 return
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]380
381 def handle(self):
382 """Handle multiple requests if necessary."""
383 self.close_connection = 1
384
385 self.handle_one_request()
386 while not self.close_connection:
387 self.handle_one_request()
388
389 def send_error(self, code, message=None):
390 """Send and log an error reply.
391
392 Arguments are the error code, and a detailed message.
393 The detailed message defaults to the short entry matching the
394 response code.
395
396 This sends an error response (so it must be called before any
397 output has been generated), logs the error, and finally sends
398 a piece of HTML explaining the error to the user.
399
400 """
401
402 try:
403 shortmsg, longmsg = self.responses[code]
404 except KeyError:
405 shortmsg, longmsg = '???', '???'
406 if message is None:
407 message = shortmsg
408 explain = longmsg
409 self.log_error("code %d, message %s", code, message)
410 # using _quote_html to prevent Cross Site Scripting attacks (see bug #1100201)
411 content = (self.error_message_format %
412 {'code': code, 'message': _quote_html(message), 'explain': explain})
413 self.send_response(code, message)
414 self.send_header("Content-Type", self.error_content_type)
415 self.send_header('Connection', 'close')
416 self.end_headers()
417 if self.command != 'HEAD' and code >= 200 and code not in (204, 304):
418 self.wfile.write(content.encode('UTF-8', 'replace'))
419
420 def send_response(self, code, message=None):
421 """Send the response header and log the response code.
422
423 Also send two standard headers with the server software
424 version and the current date.
425
426 """
427 self.log_request(code)
Senthil Kumaran0f476d42010-09-30 06:09:18 +0000[diff] [blame]428 self.send_response_only(code, message)
429 self.send_header('Server', self.version_string())
430 self.send_header('Date', self.date_time_string())
431
432 def send_response_only(self, code, message=None):
433 """Send the response header only."""
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]434 if message is None:
435 if code in self.responses:
436 message = self.responses[code][0]
437 else:
438 message = ''
439 if self.request_version != 'HTTP/0.9':
440 self.wfile.write(("%s %d %s\r\n" %
441 (self.protocol_version, code, message)).encode('ASCII', 'strict'))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]442
443 def send_header(self, keyword, value):
444 """Send a MIME header."""
445 if self.request_version != 'HTTP/0.9':
446 self.wfile.write(("%s: %s\r\n" % (keyword, value)).encode('ASCII', 'strict'))
447
448 if keyword.lower() == 'connection':
449 if value.lower() == 'close':
450 self.close_connection = 1
451 elif value.lower() == 'keep-alive':
452 self.close_connection = 0
453
454 def end_headers(self):
455 """Send the blank line ending the MIME headers."""
456 if self.request_version != 'HTTP/0.9':
457 self.wfile.write(b"\r\n")
458
459 def log_request(self, code='-', size='-'):
460 """Log an accepted request.
461
462 This is called by send_response().
463
464 """
465
466 self.log_message('"%s" %s %s',
467 self.requestline, str(code), str(size))
468
469 def log_error(self, format, *args):
470 """Log an error.
471
472 This is called when a request cannot be fulfilled. By
473 default it passes the message on to log_message().
474
475 Arguments are the same as for log_message().
476
477 XXX This should go to the separate error log.
478
479 """
480
481 self.log_message(format, *args)
482
483 def log_message(self, format, *args):
484 """Log an arbitrary message.
485
486 This is used by all other logging functions. Override
487 it if you have specific logging wishes.
488
489 The first argument, FORMAT, is a format string for the
490 message to be logged. If the format string contains
491 any % escapes requiring parameters, they should be
492 specified as subsequent arguments (it's just like
493 printf!).
494
495 The client host and current date/time are prefixed to
496 every message.
497
498 """
499
500 sys.stderr.write("%s - - [%s] %s\n" %
501 (self.address_string(),
502 self.log_date_time_string(),
503 format%args))
504
505 def version_string(self):
506 """Return the server software version string."""
507 return self.server_version + ' ' + self.sys_version
508
509 def date_time_string(self, timestamp=None):
510 """Return the current date and time formatted for a message header."""
511 if timestamp is None:
512 timestamp = time.time()
513 year, month, day, hh, mm, ss, wd, y, z = time.gmtime(timestamp)
514 s = "%s, %02d %3s %4d %02d:%02d:%02d GMT" % (
515 self.weekdayname[wd],
516 day, self.monthname[month], year,
517 hh, mm, ss)
518 return s
519
520 def log_date_time_string(self):
521 """Return the current time formatted for logging."""
522 now = time.time()
523 year, month, day, hh, mm, ss, x, y, z = time.localtime(now)
524 s = "%02d/%3s/%04d %02d:%02d:%02d" % (
525 day, self.monthname[month], year, hh, mm, ss)
526 return s
527
528 weekdayname = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
529
530 monthname = [None,
531 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun',
532 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
533
534 def address_string(self):
535 """Return the client address formatted for logging.
536
537 This version looks up the full hostname using gethostbyaddr(),
538 and tries to find a name that contains at least one dot.
539
540 """
541
542 host, port = self.client_address[:2]
543 return socket.getfqdn(host)
544
545 # Essentially static class variables
546
547 # The version of the HTTP protocol we support.
548 # Set this to HTTP/1.1 to enable automatic keepalive
549 protocol_version = "HTTP/1.0"
550
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]551 # MessageClass used to parse headers
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]552 MessageClass = http.client.HTTPMessage
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]553
554 # Table mapping response codes to messages; entries have the
555 # form {code: (shortmessage, longmessage)}.
556 # See RFC 2616.
557 responses = {
558 100: ('Continue', 'Request received, please continue'),
559 101: ('Switching Protocols',
560 'Switching to new protocol; obey Upgrade header'),
561
562 200: ('OK', 'Request fulfilled, document follows'),
563 201: ('Created', 'Document created, URL follows'),
564 202: ('Accepted',
565 'Request accepted, processing continues off-line'),
566 203: ('Non-Authoritative Information', 'Request fulfilled from cache'),
567 204: ('No Content', 'Request fulfilled, nothing follows'),
568 205: ('Reset Content', 'Clear input form for further input.'),
569 206: ('Partial Content', 'Partial content follows.'),
570
571 300: ('Multiple Choices',
572 'Object has several resources -- see URI list'),
573 301: ('Moved Permanently', 'Object moved permanently -- see URI list'),
574 302: ('Found', 'Object moved temporarily -- see URI list'),
575 303: ('See Other', 'Object moved -- see Method and URL list'),
576 304: ('Not Modified',
577 'Document has not changed since given time'),
578 305: ('Use Proxy',
579 'You must use proxy specified in Location to access this '
580 'resource.'),
581 307: ('Temporary Redirect',
582 'Object moved temporarily -- see URI list'),
583
584 400: ('Bad Request',
585 'Bad request syntax or unsupported method'),
586 401: ('Unauthorized',
587 'No permission -- see authorization schemes'),
588 402: ('Payment Required',
589 'No payment -- see charging schemes'),
590 403: ('Forbidden',
591 'Request forbidden -- authorization will not help'),
592 404: ('Not Found', 'Nothing matches the given URI'),
593 405: ('Method Not Allowed',
Senthil Kumaran7aa26212010-02-22 11:00:50 +0000[diff] [blame]594 'Specified method is invalid for this resource.'),
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]595 406: ('Not Acceptable', 'URI not available in preferred format.'),
596 407: ('Proxy Authentication Required', 'You must authenticate with '
597 'this proxy before proceeding.'),
598 408: ('Request Timeout', 'Request timed out; try again later.'),
599 409: ('Conflict', 'Request conflict.'),
600 410: ('Gone',
601 'URI no longer exists and has been permanently removed.'),
602 411: ('Length Required', 'Client must specify Content-Length.'),
603 412: ('Precondition Failed', 'Precondition in headers is false.'),
604 413: ('Request Entity Too Large', 'Entity is too large.'),
605 414: ('Request-URI Too Long', 'URI is too long.'),
606 415: ('Unsupported Media Type', 'Entity body in unsupported format.'),
607 416: ('Requested Range Not Satisfiable',
608 'Cannot satisfy request range.'),
609 417: ('Expectation Failed',
610 'Expect condition could not be satisfied.'),
611
612 500: ('Internal Server Error', 'Server got itself in trouble'),
613 501: ('Not Implemented',
614 'Server does not support this operation'),
615 502: ('Bad Gateway', 'Invalid responses from another server/proxy.'),
616 503: ('Service Unavailable',
617 'The server cannot process the request due to a high load'),
618 504: ('Gateway Timeout',
619 'The gateway server did not receive a timely response'),
620 505: ('HTTP Version Not Supported', 'Cannot fulfill request.'),
621 }
622
623
624class SimpleHTTPRequestHandler(BaseHTTPRequestHandler):
625
626 """Simple HTTP request handler with GET and HEAD commands.
627
628 This serves files from the current directory and any of its
629 subdirectories. The MIME type for files is determined by
630 calling the .guess_type() method.
631
632 The GET and HEAD requests are identical except that the HEAD
633 request omits the actual contents of the file.
634
635 """
636
637 server_version = "SimpleHTTP/" + __version__
638
639 def do_GET(self):
640 """Serve a GET request."""
641 f = self.send_head()
642 if f:
643 self.copyfile(f, self.wfile)
644 f.close()
645
646 def do_HEAD(self):
647 """Serve a HEAD request."""
648 f = self.send_head()
649 if f:
650 f.close()
651
652 def send_head(self):
653 """Common code for GET and HEAD commands.
654
655 This sends the response code and MIME headers.
656
657 Return value is either a file object (which has to be copied
658 to the outputfile by the caller unless the command was HEAD,
659 and must be closed by the caller under all circumstances), or
660 None, in which case the caller has nothing further to do.
661
662 """
663 path = self.translate_path(self.path)
664 f = None
665 if os.path.isdir(path):
666 if not self.path.endswith('/'):
667 # redirect browser - doing basically what apache does
668 self.send_response(301)
669 self.send_header("Location", self.path + "/")
670 self.end_headers()
671 return None
672 for index in "index.html", "index.htm":
673 index = os.path.join(path, index)
674 if os.path.exists(index):
675 path = index
676 break
677 else:
678 return self.list_directory(path)
679 ctype = self.guess_type(path)
680 try:
681 f = open(path, 'rb')
682 except IOError:
683 self.send_error(404, "File not found")
684 return None
685 self.send_response(200)
686 self.send_header("Content-type", ctype)
687 fs = os.fstat(f.fileno())
688 self.send_header("Content-Length", str(fs[6]))
689 self.send_header("Last-Modified", self.date_time_string(fs.st_mtime))
690 self.end_headers()
691 return f
692
693 def list_directory(self, path):
694 """Helper to produce a directory listing (absent index.html).
695
696 Return value is either a file object, or None (indicating an
697 error). In either case, the headers are sent, making the
698 interface the same as for send_head().
699
700 """
701 try:
702 list = os.listdir(path)
703 except os.error:
704 self.send_error(404, "No permission to list directory")
705 return None
706 list.sort(key=lambda a: a.lower())
707 r = []
Georg Brandl1f7fffb2010-10-15 15:57:45 +0000[diff] [blame]708 displaypath = html.escape(urllib.parse.unquote(self.path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]709 r.append('<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">')
710 r.append("<html>\n<title>Directory listing for %s</title>\n" % displaypath)
711 r.append("<body>\n<h2>Directory listing for %s</h2>\n" % displaypath)
712 r.append("<hr>\n<ul>\n")
713 for name in list:
714 fullname = os.path.join(path, name)
715 displayname = linkname = name
716 # Append / for directories or @ for symbolic links
717 if os.path.isdir(fullname):
718 displayname = name + "/"
719 linkname = name + "/"
720 if os.path.islink(fullname):
721 displayname = name + "@"
722 # Note: a link to a directory displays with @ and links with /
723 r.append('<li><a href="%s">%s</a>\n'
Georg Brandl1f7fffb2010-10-15 15:57:45 +0000[diff] [blame]724 % (urllib.parse.quote(linkname), html.escape(displayname)))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]725 r.append("</ul>\n<hr>\n</body>\n</html>\n")
726 enc = sys.getfilesystemencoding()
727 encoded = ''.join(r).encode(enc)
728 f = io.BytesIO()
729 f.write(encoded)
730 f.seek(0)
731 self.send_response(200)
732 self.send_header("Content-type", "text/html; charset=%s" % enc)
733 self.send_header("Content-Length", str(len(encoded)))
734 self.end_headers()
735 return f
736
737 def translate_path(self, path):
738 """Translate a /-separated PATH to the local filename syntax.
739
740 Components that mean special things to the local file system
741 (e.g. drive or directory names) are ignored. (XXX They should
742 probably be diagnosed.)
743
744 """
745 # abandon query parameters
746 path = path.split('?',1)[0]
747 path = path.split('#',1)[0]
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]748 path = posixpath.normpath(urllib.parse.unquote(path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]749 words = path.split('/')
750 words = filter(None, words)
751 path = os.getcwd()
752 for word in words:
753 drive, word = os.path.splitdrive(word)
754 head, word = os.path.split(word)
755 if word in (os.curdir, os.pardir): continue
756 path = os.path.join(path, word)
757 return path
758
759 def copyfile(self, source, outputfile):
760 """Copy all data between two file objects.
761
762 The SOURCE argument is a file object open for reading
763 (or anything with a read() method) and the DESTINATION
764 argument is a file object open for writing (or
765 anything with a write() method).
766
767 The only reason for overriding this would be to change
768 the block size or perhaps to replace newlines by CRLF
769 -- note however that this the default server uses this
770 to copy binary data as well.
771
772 """
773 shutil.copyfileobj(source, outputfile)
774
775 def guess_type(self, path):
776 """Guess the type of a file.
777
778 Argument is a PATH (a filename).
779
780 Return value is a string of the form type/subtype,
781 usable for a MIME Content-type header.
782
783 The default implementation looks the file's extension
784 up in the table self.extensions_map, using application/octet-stream
785 as a default; however it would be permissible (if
786 slow) to look inside the data to make a better guess.
787
788 """
789
790 base, ext = posixpath.splitext(path)
791 if ext in self.extensions_map:
792 return self.extensions_map[ext]
793 ext = ext.lower()
794 if ext in self.extensions_map:
795 return self.extensions_map[ext]
796 else:
797 return self.extensions_map['']
798
799 if not mimetypes.inited:
800 mimetypes.init() # try to read system mime.types
801 extensions_map = mimetypes.types_map.copy()
802 extensions_map.update({
803 '': 'application/octet-stream', # Default
804 '.py': 'text/plain',
805 '.c': 'text/plain',
806 '.h': 'text/plain',
807 })
808
809
810# Utilities for CGIHTTPRequestHandler
811
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]812# TODO(gregory.p.smith): Move this into an appropriate library.
813def _url_collapse_path_split(path):
814 """
815 Given a URL path, remove extra '/'s and '.' path elements and collapse
816 any '..' references.
817
818 Implements something akin to RFC-2396 5.2 step 6 to parse relative paths.
819
820 Returns: A tuple of (head, tail) where tail is everything after the final /
821 and head is everything before it. Head will always start with a '/' and,
822 if it contains anything else, never have a trailing '/'.
823
824 Raises: IndexError if too many '..' occur within the path.
825 """
826 # Similar to os.path.split(os.path.normpath(path)) but specific to URL
827 # path semantics rather than local operating system semantics.
828 path_parts = []
829 for part in path.split('/'):
830 if part == '.':
831 path_parts.append('')
832 else:
833 path_parts.append(part)
834 # Filter out blank non trailing parts before consuming the '..'.
835 path_parts = [part for part in path_parts[:-1] if part] + path_parts[-1:]
836 if path_parts:
837 tail_part = path_parts.pop()
838 else:
839 tail_part = ''
840 head_parts = []
841 for part in path_parts:
842 if part == '..':
843 head_parts.pop()
844 else:
845 head_parts.append(part)
846 if tail_part and tail_part == '..':
847 head_parts.pop()
848 tail_part = ''
849 return ('/' + '/'.join(head_parts), tail_part)
850
851
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]852nobody = None
853
854def nobody_uid():
855 """Internal routine to get nobody's uid"""
856 global nobody
857 if nobody:
858 return nobody
859 try:
860 import pwd
861 except ImportError:
862 return -1
863 try:
864 nobody = pwd.getpwnam('nobody')[2]
865 except KeyError:
866 nobody = 1 + max(map(lambda x: x[2], pwd.getpwall()))
867 return nobody
868
869
870def executable(path):
871 """Test for executable file."""
872 try:
873 st = os.stat(path)
874 except os.error:
875 return False
876 return st.st_mode & 0o111 != 0
877
878
879class CGIHTTPRequestHandler(SimpleHTTPRequestHandler):
880
881 """Complete HTTP server with GET, HEAD and POST commands.
882
883 GET and HEAD also support running CGI scripts.
884
885 The POST command is *only* implemented for CGI scripts.
886
887 """
888
889 # Determine platform specifics
890 have_fork = hasattr(os, 'fork')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]891
892 # Make rfile unbuffered -- we need to read one line and then pass
893 # the rest to a subprocess, so we can't use buffered input.
894 rbufsize = 0
895
896 def do_POST(self):
897 """Serve a POST request.
898
899 This is only implemented for CGI scripts.
900
901 """
902
903 if self.is_cgi():
904 self.run_cgi()
905 else:
906 self.send_error(501, "Can only POST to CGI scripts")
907
908 def send_head(self):
909 """Version of send_head that support CGI scripts"""
910 if self.is_cgi():
911 return self.run_cgi()
912 else:
913 return SimpleHTTPRequestHandler.send_head(self)
914
915 def is_cgi(self):
916 """Test whether self.path corresponds to a CGI script.
917
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]918 Returns True and updates the cgi_info attribute to the tuple
919 (dir, rest) if self.path requires running a CGI script.
920 Returns False otherwise.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]921
Benjamin Petersona7deeee2009-05-08 20:54:42 +0000[diff] [blame]922 If any exception is raised, the caller should assume that
923 self.path was rejected as invalid and act accordingly.
924
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]925 The default implementation tests whether the normalized url
926 path begins with one of the strings in self.cgi_directories
927 (and the next character is a '/' or the end of the string).
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]928
929 """
930
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]931 splitpath = _url_collapse_path_split(self.path)
932 if splitpath[0] in self.cgi_directories:
933 self.cgi_info = splitpath
934 return True
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]935 return False
936
937 cgi_directories = ['/cgi-bin', '/htbin']
938
939 def is_executable(self, path):
940 """Test whether argument path is an executable file."""
941 return executable(path)
942
943 def is_python(self, path):
944 """Test whether argument path is a Python script."""
945 head, tail = os.path.splitext(path)
946 return tail.lower() in (".py", ".pyw")
947
948 def run_cgi(self):
949 """Execute a CGI script."""
950 path = self.path
951 dir, rest = self.cgi_info
952
953 i = path.find('/', len(dir) + 1)
954 while i >= 0:
955 nextdir = path[:i]
956 nextrest = path[i+1:]
957
958 scriptdir = self.translate_path(nextdir)
959 if os.path.isdir(scriptdir):
960 dir, rest = nextdir, nextrest
961 i = path.find('/', len(dir) + 1)
962 else:
963 break
964
965 # find an explicit query string, if present.
966 i = rest.rfind('?')
967 if i >= 0:
968 rest, query = rest[:i], rest[i+1:]
969 else:
970 query = ''
971
972 # dissect the part after the directory name into a script name &
973 # a possible additional path, to be stored in PATH_INFO.
974 i = rest.find('/')
975 if i >= 0:
976 script, rest = rest[:i], rest[i:]
977 else:
978 script, rest = rest, ''
979
980 scriptname = dir + '/' + script
981 scriptfile = self.translate_path(scriptname)
982 if not os.path.exists(scriptfile):
983 self.send_error(404, "No such CGI script (%r)" % scriptname)
984 return
985 if not os.path.isfile(scriptfile):
986 self.send_error(403, "CGI script is not a plain file (%r)" %
987 scriptname)
988 return
989 ispy = self.is_python(scriptname)
990 if not ispy:
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]991 if not self.is_executable(scriptfile):
992 self.send_error(403, "CGI script is not executable (%r)" %
993 scriptname)
994 return
995
996 # Reference: http://hoohoo.ncsa.uiuc.edu/cgi/env.html
997 # XXX Much of the following could be prepared ahead of time!
Senthil Kumaran42713722010-10-03 17:55:45 +0000[diff] [blame]998 env = copy.deepcopy(os.environ)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]999 env['SERVER_SOFTWARE'] = self.version_string()
1000 env['SERVER_NAME'] = self.server.server_name
1001 env['GATEWAY_INTERFACE'] = 'CGI/1.1'
1002 env['SERVER_PROTOCOL'] = self.protocol_version
1003 env['SERVER_PORT'] = str(self.server.server_port)
1004 env['REQUEST_METHOD'] = self.command
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]1005 uqrest = urllib.parse.unquote(rest)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1006 env['PATH_INFO'] = uqrest
1007 env['PATH_TRANSLATED'] = self.translate_path(uqrest)
1008 env['SCRIPT_NAME'] = scriptname
1009 if query:
1010 env['QUERY_STRING'] = query
1011 host = self.address_string()
1012 if host != self.client_address[0]:
1013 env['REMOTE_HOST'] = host
1014 env['REMOTE_ADDR'] = self.client_address[0]
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1015 authorization = self.headers.get("authorization")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1016 if authorization:
1017 authorization = authorization.split()
1018 if len(authorization) == 2:
1019 import base64, binascii
1020 env['AUTH_TYPE'] = authorization[0]
1021 if authorization[0].lower() == "basic":
1022 try:
1023 authorization = authorization[1].encode('ascii')
Georg Brandl706824f2009-06-04 09:42:55 +0000[diff] [blame]1024 authorization = base64.decodebytes(authorization).\
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1025 decode('ascii')
1026 except (binascii.Error, UnicodeError):
1027 pass
1028 else:
1029 authorization = authorization.split(':')
1030 if len(authorization) == 2:
1031 env['REMOTE_USER'] = authorization[0]
1032 # XXX REMOTE_IDENT
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1033 if self.headers.get('content-type') is None:
1034 env['CONTENT_TYPE'] = self.headers.get_content_type()
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1035 else:
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1036 env['CONTENT_TYPE'] = self.headers['content-type']
1037 length = self.headers.get('content-length')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1038 if length:
1039 env['CONTENT_LENGTH'] = length
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1040 referer = self.headers.get('referer')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1041 if referer:
1042 env['HTTP_REFERER'] = referer
1043 accept = []
1044 for line in self.headers.getallmatchingheaders('accept'):
1045 if line[:1] in "\t\n\r ":
1046 accept.append(line.strip())
1047 else:
1048 accept = accept + line[7:].split(',')
1049 env['HTTP_ACCEPT'] = ','.join(accept)
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1050 ua = self.headers.get('user-agent')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1051 if ua:
1052 env['HTTP_USER_AGENT'] = ua
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1053 co = filter(None, self.headers.get_all('cookie', []))
Georg Brandl62e2ca22010-07-31 21:54:24 +0000[diff] [blame]1054 cookie_str = ', '.join(co)
1055 if cookie_str:
1056 env['HTTP_COOKIE'] = cookie_str
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1057 # XXX Other HTTP_* headers
1058 # Since we're setting the env in the parent, provide empty
1059 # values to override previously set values
1060 for k in ('QUERY_STRING', 'REMOTE_HOST', 'CONTENT_LENGTH',
1061 'HTTP_USER_AGENT', 'HTTP_COOKIE', 'HTTP_REFERER'):
1062 env.setdefault(k, "")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]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)
Senthil Kumaran42713722010-10-03 17:55:45 +0000[diff] [blame]1094 os.execve(scriptfile, args, env)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]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 Kumaran42713722010-10-03 17:55:45 +0000[diff] [blame]1119 stderr=subprocess.PIPE,
1120 env = env
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1121 )
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1122 if self.command.lower() == "post" and nbytes > 0:
1123 data = self.rfile.read(nbytes)
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1124 else:
1125 data = None
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1126 # throw away additional data [see bug #427345]
1127 while select.select([self.rfile._sock], [], [], 0)[0]:
1128 if not self.rfile._sock.recv(1):
1129 break
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1130 stdout, stderr = p.communicate(data)
1131 self.wfile.write(stdout)
1132 if stderr:
1133 self.log_error('%s', stderr)
Brian Curtincbad4df2010-11-05 15:04:48 +0000[diff] [blame]1134 p.stderr.close()
1135 p.stdout.close()
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1136 status = p.returncode
1137 if status:
1138 self.log_error("CGI script exit status %#x", status)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1139 else:
1140 self.log_message("CGI script exited OK")
1141
1142
1143def test(HandlerClass = BaseHTTPRequestHandler,
1144 ServerClass = HTTPServer, protocol="HTTP/1.0"):
1145 """Test the HTTP request handler class.
1146
1147 This runs an HTTP server on port 8000 (or the first command line
1148 argument).
1149
1150 """
1151
1152 if sys.argv[1:]:
1153 port = int(sys.argv[1])
1154 else:
1155 port = 8000
1156 server_address = ('', port)
1157
1158 HandlerClass.protocol_version = protocol
1159 httpd = ServerClass(server_address, HandlerClass)
1160
1161 sa = httpd.socket.getsockname()
1162 print("Serving HTTP on", sa[0], "port", sa[1], "...")
Alexandre Vassalottib5292a22009-04-03 07:16:55 +0000[diff] [blame]1163 try:
1164 httpd.serve_forever()
1165 except KeyboardInterrupt:
1166 print("\nKeyboard interrupt received, exiting.")
1167 httpd.server_close()
1168 sys.exit(0)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1169
1170if __name__ == '__main__':
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1171 test(HandlerClass=SimpleHTTPRequestHandler)