Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / python / cpython3 / ff1bbba92aad261df1ebd8fd8cc189c104e113b0 / . / Lib / http / server.py
blob: 8de604a739e3457ad2467d74dcf3af512e440eaa [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 = """\
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.
Antoine Pitrouff1bbba2010-12-18 18:04:38 +0000[diff] [blame^]317 try:
318 self.headers = http.client.parse_headers(self.rfile,
319 _class=self.MessageClass)
320 except http.client.LineTooLong:
321 self.send_error(400, "Line too long")
322 return False
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]323
324 conntype = self.headers.get('Connection', "")
325 if conntype.lower() == 'close':
326 self.close_connection = 1
327 elif (conntype.lower() == 'keep-alive' and
328 self.protocol_version >= "HTTP/1.1"):
329 self.close_connection = 0
330 return True
331
332 def handle_one_request(self):
333 """Handle a single HTTP request.
334
335 You normally don't need to override this method; see the class
336 __doc__ string for information on how to handle specific HTTP
337 commands such as GET and POST.
338
339 """
Antoine Pitrou3022ce12010-12-16 17:03:16 +0000[diff] [blame]340 self.raw_requestline = self.rfile.readline(65537)
341 if len(self.raw_requestline) > 65536:
342 self.requestline = ''
343 self.request_version = ''
344 self.command = ''
345 self.send_error(414)
346 return
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]347 if not self.raw_requestline:
348 self.close_connection = 1
349 return
350 if not self.parse_request(): # An error code has been sent, just exit
351 return
352 mname = 'do_' + self.command
353 if not hasattr(self, mname):
354 self.send_error(501, "Unsupported method (%r)" % self.command)
355 return
356 method = getattr(self, mname)
357 method()
358
359 def handle(self):
360 """Handle multiple requests if necessary."""
361 self.close_connection = 1
362
363 self.handle_one_request()
364 while not self.close_connection:
365 self.handle_one_request()
366
367 def send_error(self, code, message=None):
368 """Send and log an error reply.
369
370 Arguments are the error code, and a detailed message.
371 The detailed message defaults to the short entry matching the
372 response code.
373
374 This sends an error response (so it must be called before any
375 output has been generated), logs the error, and finally sends
376 a piece of HTML explaining the error to the user.
377
378 """
379
380 try:
381 shortmsg, longmsg = self.responses[code]
382 except KeyError:
383 shortmsg, longmsg = '???', '???'
384 if message is None:
385 message = shortmsg
386 explain = longmsg
387 self.log_error("code %d, message %s", code, message)
388 # using _quote_html to prevent Cross Site Scripting attacks (see bug #1100201)
389 content = (self.error_message_format %
390 {'code': code, 'message': _quote_html(message), 'explain': explain})
391 self.send_response(code, message)
392 self.send_header("Content-Type", self.error_content_type)
393 self.send_header('Connection', 'close')
394 self.end_headers()
395 if self.command != 'HEAD' and code >= 200 and code not in (204, 304):
396 self.wfile.write(content.encode('UTF-8', 'replace'))
397
398 def send_response(self, code, message=None):
399 """Send the response header and log the response code.
400
401 Also send two standard headers with the server software
402 version and the current date.
403
404 """
405 self.log_request(code)
406 if message is None:
407 if code in self.responses:
408 message = self.responses[code][0]
409 else:
410 message = ''
411 if self.request_version != 'HTTP/0.9':
412 self.wfile.write(("%s %d %s\r\n" %
413 (self.protocol_version, code, message)).encode('ASCII', 'strict'))
414 # print (self.protocol_version, code, message)
415 self.send_header('Server', self.version_string())
416 self.send_header('Date', self.date_time_string())
417
418 def send_header(self, keyword, value):
419 """Send a MIME header."""
420 if self.request_version != 'HTTP/0.9':
421 self.wfile.write(("%s: %s\r\n" % (keyword, value)).encode('ASCII', 'strict'))
422
423 if keyword.lower() == 'connection':
424 if value.lower() == 'close':
425 self.close_connection = 1
426 elif value.lower() == 'keep-alive':
427 self.close_connection = 0
428
429 def end_headers(self):
430 """Send the blank line ending the MIME headers."""
431 if self.request_version != 'HTTP/0.9':
432 self.wfile.write(b"\r\n")
433
434 def log_request(self, code='-', size='-'):
435 """Log an accepted request.
436
437 This is called by send_response().
438
439 """
440
441 self.log_message('"%s" %s %s',
442 self.requestline, str(code), str(size))
443
444 def log_error(self, format, *args):
445 """Log an error.
446
447 This is called when a request cannot be fulfilled. By
448 default it passes the message on to log_message().
449
450 Arguments are the same as for log_message().
451
452 XXX This should go to the separate error log.
453
454 """
455
456 self.log_message(format, *args)
457
458 def log_message(self, format, *args):
459 """Log an arbitrary message.
460
461 This is used by all other logging functions. Override
462 it if you have specific logging wishes.
463
464 The first argument, FORMAT, is a format string for the
465 message to be logged. If the format string contains
466 any % escapes requiring parameters, they should be
467 specified as subsequent arguments (it's just like
468 printf!).
469
470 The client host and current date/time are prefixed to
471 every message.
472
473 """
474
475 sys.stderr.write("%s - - [%s] %s\n" %
476 (self.address_string(),
477 self.log_date_time_string(),
478 format%args))
479
480 def version_string(self):
481 """Return the server software version string."""
482 return self.server_version + ' ' + self.sys_version
483
484 def date_time_string(self, timestamp=None):
485 """Return the current date and time formatted for a message header."""
486 if timestamp is None:
487 timestamp = time.time()
488 year, month, day, hh, mm, ss, wd, y, z = time.gmtime(timestamp)
489 s = "%s, %02d %3s %4d %02d:%02d:%02d GMT" % (
490 self.weekdayname[wd],
491 day, self.monthname[month], year,
492 hh, mm, ss)
493 return s
494
495 def log_date_time_string(self):
496 """Return the current time formatted for logging."""
497 now = time.time()
498 year, month, day, hh, mm, ss, x, y, z = time.localtime(now)
499 s = "%02d/%3s/%04d %02d:%02d:%02d" % (
500 day, self.monthname[month], year, hh, mm, ss)
501 return s
502
503 weekdayname = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
504
505 monthname = [None,
506 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun',
507 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
508
509 def address_string(self):
510 """Return the client address formatted for logging.
511
512 This version looks up the full hostname using gethostbyaddr(),
513 and tries to find a name that contains at least one dot.
514
515 """
516
517 host, port = self.client_address[:2]
518 return socket.getfqdn(host)
519
520 # Essentially static class variables
521
522 # The version of the HTTP protocol we support.
523 # Set this to HTTP/1.1 to enable automatic keepalive
524 protocol_version = "HTTP/1.0"
525
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]526 # MessageClass used to parse headers
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]527 MessageClass = http.client.HTTPMessage
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]528
529 # Table mapping response codes to messages; entries have the
530 # form {code: (shortmessage, longmessage)}.
531 # See RFC 2616.
532 responses = {
533 100: ('Continue', 'Request received, please continue'),
534 101: ('Switching Protocols',
535 'Switching to new protocol; obey Upgrade header'),
536
537 200: ('OK', 'Request fulfilled, document follows'),
538 201: ('Created', 'Document created, URL follows'),
539 202: ('Accepted',
540 'Request accepted, processing continues off-line'),
541 203: ('Non-Authoritative Information', 'Request fulfilled from cache'),
542 204: ('No Content', 'Request fulfilled, nothing follows'),
543 205: ('Reset Content', 'Clear input form for further input.'),
544 206: ('Partial Content', 'Partial content follows.'),
545
546 300: ('Multiple Choices',
547 'Object has several resources -- see URI list'),
548 301: ('Moved Permanently', 'Object moved permanently -- see URI list'),
549 302: ('Found', 'Object moved temporarily -- see URI list'),
550 303: ('See Other', 'Object moved -- see Method and URL list'),
551 304: ('Not Modified',
552 'Document has not changed since given time'),
553 305: ('Use Proxy',
554 'You must use proxy specified in Location to access this '
555 'resource.'),
556 307: ('Temporary Redirect',
557 'Object moved temporarily -- see URI list'),
558
559 400: ('Bad Request',
560 'Bad request syntax or unsupported method'),
561 401: ('Unauthorized',
562 'No permission -- see authorization schemes'),
563 402: ('Payment Required',
564 'No payment -- see charging schemes'),
565 403: ('Forbidden',
566 'Request forbidden -- authorization will not help'),
567 404: ('Not Found', 'Nothing matches the given URI'),
568 405: ('Method Not Allowed',
Senthil Kumaran613c61c2010-02-22 11:02:53 +0000[diff] [blame]569 'Specified method is invalid for this resource.'),
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]570 406: ('Not Acceptable', 'URI not available in preferred format.'),
571 407: ('Proxy Authentication Required', 'You must authenticate with '
572 'this proxy before proceeding.'),
573 408: ('Request Timeout', 'Request timed out; try again later.'),
574 409: ('Conflict', 'Request conflict.'),
575 410: ('Gone',
576 'URI no longer exists and has been permanently removed.'),
577 411: ('Length Required', 'Client must specify Content-Length.'),
578 412: ('Precondition Failed', 'Precondition in headers is false.'),
579 413: ('Request Entity Too Large', 'Entity is too large.'),
580 414: ('Request-URI Too Long', 'URI is too long.'),
581 415: ('Unsupported Media Type', 'Entity body in unsupported format.'),
582 416: ('Requested Range Not Satisfiable',
583 'Cannot satisfy request range.'),
584 417: ('Expectation Failed',
585 'Expect condition could not be satisfied.'),
586
587 500: ('Internal Server Error', 'Server got itself in trouble'),
588 501: ('Not Implemented',
589 'Server does not support this operation'),
590 502: ('Bad Gateway', 'Invalid responses from another server/proxy.'),
591 503: ('Service Unavailable',
592 'The server cannot process the request due to a high load'),
593 504: ('Gateway Timeout',
594 'The gateway server did not receive a timely response'),
595 505: ('HTTP Version Not Supported', 'Cannot fulfill request.'),
596 }
597
598
599class SimpleHTTPRequestHandler(BaseHTTPRequestHandler):
600
601 """Simple HTTP request handler with GET and HEAD commands.
602
603 This serves files from the current directory and any of its
604 subdirectories. The MIME type for files is determined by
605 calling the .guess_type() method.
606
607 The GET and HEAD requests are identical except that the HEAD
608 request omits the actual contents of the file.
609
610 """
611
612 server_version = "SimpleHTTP/" + __version__
613
614 def do_GET(self):
615 """Serve a GET request."""
616 f = self.send_head()
617 if f:
618 self.copyfile(f, self.wfile)
619 f.close()
620
621 def do_HEAD(self):
622 """Serve a HEAD request."""
623 f = self.send_head()
624 if f:
625 f.close()
626
627 def send_head(self):
628 """Common code for GET and HEAD commands.
629
630 This sends the response code and MIME headers.
631
632 Return value is either a file object (which has to be copied
633 to the outputfile by the caller unless the command was HEAD,
634 and must be closed by the caller under all circumstances), or
635 None, in which case the caller has nothing further to do.
636
637 """
638 path = self.translate_path(self.path)
639 f = None
640 if os.path.isdir(path):
641 if not self.path.endswith('/'):
642 # redirect browser - doing basically what apache does
643 self.send_response(301)
644 self.send_header("Location", self.path + "/")
645 self.end_headers()
646 return None
647 for index in "index.html", "index.htm":
648 index = os.path.join(path, index)
649 if os.path.exists(index):
650 path = index
651 break
652 else:
653 return self.list_directory(path)
654 ctype = self.guess_type(path)
655 try:
656 f = open(path, 'rb')
657 except IOError:
658 self.send_error(404, "File not found")
659 return None
660 self.send_response(200)
661 self.send_header("Content-type", ctype)
662 fs = os.fstat(f.fileno())
663 self.send_header("Content-Length", str(fs[6]))
664 self.send_header("Last-Modified", self.date_time_string(fs.st_mtime))
665 self.end_headers()
666 return f
667
668 def list_directory(self, path):
669 """Helper to produce a directory listing (absent index.html).
670
671 Return value is either a file object, or None (indicating an
672 error). In either case, the headers are sent, making the
673 interface the same as for send_head().
674
675 """
676 try:
677 list = os.listdir(path)
678 except os.error:
679 self.send_error(404, "No permission to list directory")
680 return None
681 list.sort(key=lambda a: a.lower())
682 r = []
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]683 displaypath = cgi.escape(urllib.parse.unquote(self.path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]684 r.append('<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">')
685 r.append("<html>\n<title>Directory listing for %s</title>\n" % displaypath)
686 r.append("<body>\n<h2>Directory listing for %s</h2>\n" % displaypath)
687 r.append("<hr>\n<ul>\n")
688 for name in list:
689 fullname = os.path.join(path, name)
690 displayname = linkname = name
691 # Append / for directories or @ for symbolic links
692 if os.path.isdir(fullname):
693 displayname = name + "/"
694 linkname = name + "/"
695 if os.path.islink(fullname):
696 displayname = name + "@"
697 # Note: a link to a directory displays with @ and links with /
698 r.append('<li><a href="%s">%s</a>\n'
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]699 % (urllib.parse.quote(linkname), cgi.escape(displayname)))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]700 r.append("</ul>\n<hr>\n</body>\n</html>\n")
701 enc = sys.getfilesystemencoding()
702 encoded = ''.join(r).encode(enc)
703 f = io.BytesIO()
704 f.write(encoded)
705 f.seek(0)
706 self.send_response(200)
707 self.send_header("Content-type", "text/html; charset=%s" % enc)
708 self.send_header("Content-Length", str(len(encoded)))
709 self.end_headers()
710 return f
711
712 def translate_path(self, path):
713 """Translate a /-separated PATH to the local filename syntax.
714
715 Components that mean special things to the local file system
716 (e.g. drive or directory names) are ignored. (XXX They should
717 probably be diagnosed.)
718
719 """
720 # abandon query parameters
721 path = path.split('?',1)[0]
722 path = path.split('#',1)[0]
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]723 path = posixpath.normpath(urllib.parse.unquote(path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]724 words = path.split('/')
725 words = filter(None, words)
726 path = os.getcwd()
727 for word in words:
728 drive, word = os.path.splitdrive(word)
729 head, word = os.path.split(word)
730 if word in (os.curdir, os.pardir): continue
731 path = os.path.join(path, word)
732 return path
733
734 def copyfile(self, source, outputfile):
735 """Copy all data between two file objects.
736
737 The SOURCE argument is a file object open for reading
738 (or anything with a read() method) and the DESTINATION
739 argument is a file object open for writing (or
740 anything with a write() method).
741
742 The only reason for overriding this would be to change
743 the block size or perhaps to replace newlines by CRLF
744 -- note however that this the default server uses this
745 to copy binary data as well.
746
747 """
748 shutil.copyfileobj(source, outputfile)
749
750 def guess_type(self, path):
751 """Guess the type of a file.
752
753 Argument is a PATH (a filename).
754
755 Return value is a string of the form type/subtype,
756 usable for a MIME Content-type header.
757
758 The default implementation looks the file's extension
759 up in the table self.extensions_map, using application/octet-stream
760 as a default; however it would be permissible (if
761 slow) to look inside the data to make a better guess.
762
763 """
764
765 base, ext = posixpath.splitext(path)
766 if ext in self.extensions_map:
767 return self.extensions_map[ext]
768 ext = ext.lower()
769 if ext in self.extensions_map:
770 return self.extensions_map[ext]
771 else:
772 return self.extensions_map['']
773
774 if not mimetypes.inited:
775 mimetypes.init() # try to read system mime.types
776 extensions_map = mimetypes.types_map.copy()
777 extensions_map.update({
778 '': 'application/octet-stream', # Default
779 '.py': 'text/plain',
780 '.c': 'text/plain',
781 '.h': 'text/plain',
782 })
783
784
785# Utilities for CGIHTTPRequestHandler
786
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]787# TODO(gregory.p.smith): Move this into an appropriate library.
788def _url_collapse_path_split(path):
789 """
790 Given a URL path, remove extra '/'s and '.' path elements and collapse
791 any '..' references.
792
793 Implements something akin to RFC-2396 5.2 step 6 to parse relative paths.
794
795 Returns: A tuple of (head, tail) where tail is everything after the final /
796 and head is everything before it. Head will always start with a '/' and,
797 if it contains anything else, never have a trailing '/'.
798
799 Raises: IndexError if too many '..' occur within the path.
800 """
801 # Similar to os.path.split(os.path.normpath(path)) but specific to URL
802 # path semantics rather than local operating system semantics.
803 path_parts = []
804 for part in path.split('/'):
805 if part == '.':
806 path_parts.append('')
807 else:
808 path_parts.append(part)
809 # Filter out blank non trailing parts before consuming the '..'.
810 path_parts = [part for part in path_parts[:-1] if part] + path_parts[-1:]
811 if path_parts:
812 tail_part = path_parts.pop()
813 else:
814 tail_part = ''
815 head_parts = []
816 for part in path_parts:
817 if part == '..':
818 head_parts.pop()
819 else:
820 head_parts.append(part)
821 if tail_part and tail_part == '..':
822 head_parts.pop()
823 tail_part = ''
824 return ('/' + '/'.join(head_parts), tail_part)
825
826
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]827nobody = None
828
829def nobody_uid():
830 """Internal routine to get nobody's uid"""
831 global nobody
832 if nobody:
833 return nobody
834 try:
835 import pwd
836 except ImportError:
837 return -1
838 try:
839 nobody = pwd.getpwnam('nobody')[2]
840 except KeyError:
841 nobody = 1 + max(map(lambda x: x[2], pwd.getpwall()))
842 return nobody
843
844
845def executable(path):
846 """Test for executable file."""
847 try:
848 st = os.stat(path)
849 except os.error:
850 return False
851 return st.st_mode & 0o111 != 0
852
853
854class CGIHTTPRequestHandler(SimpleHTTPRequestHandler):
855
856 """Complete HTTP server with GET, HEAD and POST commands.
857
858 GET and HEAD also support running CGI scripts.
859
860 The POST command is *only* implemented for CGI scripts.
861
862 """
863
864 # Determine platform specifics
865 have_fork = hasattr(os, 'fork')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]866
867 # Make rfile unbuffered -- we need to read one line and then pass
868 # the rest to a subprocess, so we can't use buffered input.
869 rbufsize = 0
870
871 def do_POST(self):
872 """Serve a POST request.
873
874 This is only implemented for CGI scripts.
875
876 """
877
878 if self.is_cgi():
879 self.run_cgi()
880 else:
881 self.send_error(501, "Can only POST to CGI scripts")
882
883 def send_head(self):
884 """Version of send_head that support CGI scripts"""
885 if self.is_cgi():
886 return self.run_cgi()
887 else:
888 return SimpleHTTPRequestHandler.send_head(self)
889
890 def is_cgi(self):
891 """Test whether self.path corresponds to a CGI script.
892
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]893 Returns True and updates the cgi_info attribute to the tuple
894 (dir, rest) if self.path requires running a CGI script.
895 Returns False otherwise.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]896
Benjamin Petersona7deeee2009-05-08 20:54:42 +0000[diff] [blame]897 If any exception is raised, the caller should assume that
898 self.path was rejected as invalid and act accordingly.
899
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]900 The default implementation tests whether the normalized url
901 path begins with one of the strings in self.cgi_directories
902 (and the next character is a '/' or the end of the string).
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]903
904 """
905
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]906 splitpath = _url_collapse_path_split(self.path)
907 if splitpath[0] in self.cgi_directories:
908 self.cgi_info = splitpath
909 return True
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]910 return False
911
912 cgi_directories = ['/cgi-bin', '/htbin']
913
914 def is_executable(self, path):
915 """Test whether argument path is an executable file."""
916 return executable(path)
917
918 def is_python(self, path):
919 """Test whether argument path is a Python script."""
920 head, tail = os.path.splitext(path)
921 return tail.lower() in (".py", ".pyw")
922
923 def run_cgi(self):
924 """Execute a CGI script."""
925 path = self.path
926 dir, rest = self.cgi_info
927
928 i = path.find('/', len(dir) + 1)
929 while i >= 0:
930 nextdir = path[:i]
931 nextrest = path[i+1:]
932
933 scriptdir = self.translate_path(nextdir)
934 if os.path.isdir(scriptdir):
935 dir, rest = nextdir, nextrest
936 i = path.find('/', len(dir) + 1)
937 else:
938 break
939
940 # find an explicit query string, if present.
941 i = rest.rfind('?')
942 if i >= 0:
943 rest, query = rest[:i], rest[i+1:]
944 else:
945 query = ''
946
947 # dissect the part after the directory name into a script name &
948 # a possible additional path, to be stored in PATH_INFO.
949 i = rest.find('/')
950 if i >= 0:
951 script, rest = rest[:i], rest[i:]
952 else:
953 script, rest = rest, ''
954
955 scriptname = dir + '/' + script
956 scriptfile = self.translate_path(scriptname)
957 if not os.path.exists(scriptfile):
958 self.send_error(404, "No such CGI script (%r)" % scriptname)
959 return
960 if not os.path.isfile(scriptfile):
961 self.send_error(403, "CGI script is not a plain file (%r)" %
962 scriptname)
963 return
964 ispy = self.is_python(scriptname)
965 if not ispy:
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]966 if not self.is_executable(scriptfile):
967 self.send_error(403, "CGI script is not executable (%r)" %
968 scriptname)
969 return
970
971 # Reference: http://hoohoo.ncsa.uiuc.edu/cgi/env.html
972 # XXX Much of the following could be prepared ahead of time!
Senthil Kumaran5e8826c2010-10-03 18:04:52 +0000[diff] [blame]973 env = copy.deepcopy(os.environ)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]974 env['SERVER_SOFTWARE'] = self.version_string()
975 env['SERVER_NAME'] = self.server.server_name
976 env['GATEWAY_INTERFACE'] = 'CGI/1.1'
977 env['SERVER_PROTOCOL'] = self.protocol_version
978 env['SERVER_PORT'] = str(self.server.server_port)
979 env['REQUEST_METHOD'] = self.command
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]980 uqrest = urllib.parse.unquote(rest)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]981 env['PATH_INFO'] = uqrest
982 env['PATH_TRANSLATED'] = self.translate_path(uqrest)
983 env['SCRIPT_NAME'] = scriptname
984 if query:
985 env['QUERY_STRING'] = query
986 host = self.address_string()
987 if host != self.client_address[0]:
988 env['REMOTE_HOST'] = host
989 env['REMOTE_ADDR'] = self.client_address[0]
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]990 authorization = self.headers.get("authorization")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]991 if authorization:
992 authorization = authorization.split()
993 if len(authorization) == 2:
994 import base64, binascii
995 env['AUTH_TYPE'] = authorization[0]
996 if authorization[0].lower() == "basic":
997 try:
998 authorization = authorization[1].encode('ascii')
Georg Brandl706824f2009-06-04 09:42:55 +0000[diff] [blame]999 authorization = base64.decodebytes(authorization).\
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1000 decode('ascii')
1001 except (binascii.Error, UnicodeError):
1002 pass
1003 else:
1004 authorization = authorization.split(':')
1005 if len(authorization) == 2:
1006 env['REMOTE_USER'] = authorization[0]
1007 # XXX REMOTE_IDENT
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1008 if self.headers.get('content-type') is None:
1009 env['CONTENT_TYPE'] = self.headers.get_content_type()
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1010 else:
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1011 env['CONTENT_TYPE'] = self.headers['content-type']
1012 length = self.headers.get('content-length')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1013 if length:
1014 env['CONTENT_LENGTH'] = length
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1015 referer = self.headers.get('referer')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1016 if referer:
1017 env['HTTP_REFERER'] = referer
1018 accept = []
1019 for line in self.headers.getallmatchingheaders('accept'):
1020 if line[:1] in "\t\n\r ":
1021 accept.append(line.strip())
1022 else:
1023 accept = accept + line[7:].split(',')
1024 env['HTTP_ACCEPT'] = ','.join(accept)
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1025 ua = self.headers.get('user-agent')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1026 if ua:
1027 env['HTTP_USER_AGENT'] = ua
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1028 co = filter(None, self.headers.get_all('cookie', []))
Georg Brandlcaa78fe2010-08-01 19:07:28 +0000[diff] [blame]1029 cookie_str = ', '.join(co)
1030 if cookie_str:
1031 env['HTTP_COOKIE'] = cookie_str
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1032 # XXX Other HTTP_* headers
1033 # Since we're setting the env in the parent, provide empty
1034 # values to override previously set values
1035 for k in ('QUERY_STRING', 'REMOTE_HOST', 'CONTENT_LENGTH',
1036 'HTTP_USER_AGENT', 'HTTP_COOKIE', 'HTTP_REFERER'):
1037 env.setdefault(k, "")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1038
1039 self.send_response(200, "Script output follows")
1040
1041 decoded_query = query.replace('+', ' ')
1042
1043 if self.have_fork:
1044 # Unix -- fork as we should
1045 args = [script]
1046 if '=' not in decoded_query:
1047 args.append(decoded_query)
1048 nobody = nobody_uid()
1049 self.wfile.flush() # Always flush before forking
1050 pid = os.fork()
1051 if pid != 0:
1052 # Parent
1053 pid, sts = os.waitpid(pid, 0)
1054 # throw away additional data [see bug #427345]
1055 while select.select([self.rfile], [], [], 0)[0]:
1056 if not self.rfile.read(1):
1057 break
1058 if sts:
1059 self.log_error("CGI script exit status %#x", sts)
1060 return
1061 # Child
1062 try:
1063 try:
1064 os.setuid(nobody)
1065 except os.error:
1066 pass
1067 os.dup2(self.rfile.fileno(), 0)
1068 os.dup2(self.wfile.fileno(), 1)
Senthil Kumaran5e8826c2010-10-03 18:04:52 +0000[diff] [blame]1069 os.execve(scriptfile, args, env)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1070 except:
1071 self.server.handle_error(self.request, self.client_address)
1072 os._exit(127)
1073
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1074 else:
1075 # Non-Unix -- use subprocess
1076 import subprocess
Senthil Kumaranca5130c2009-11-11 04:21:22 +0000[diff] [blame]1077 cmdline = [scriptfile]
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1078 if self.is_python(scriptfile):
1079 interp = sys.executable
1080 if interp.lower().endswith("w.exe"):
1081 # On Windows, use python.exe, not pythonw.exe
1082 interp = interp[:-5] + interp[-4:]
Senthil Kumaranca5130c2009-11-11 04:21:22 +0000[diff] [blame]1083 cmdline = [interp, '-u'] + cmdline
1084 if '=' not in query:
1085 cmdline.append(query)
1086 self.log_message("command: %s", subprocess.list2cmdline(cmdline))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1087 try:
1088 nbytes = int(length)
1089 except (TypeError, ValueError):
1090 nbytes = 0
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1091 p = subprocess.Popen(cmdline,
1092 stdin=subprocess.PIPE,
1093 stdout=subprocess.PIPE,
Senthil Kumaran5e8826c2010-10-03 18:04:52 +0000[diff] [blame]1094 stderr=subprocess.PIPE,
1095 env = env
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1096 )
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1097 if self.command.lower() == "post" and nbytes > 0:
1098 data = self.rfile.read(nbytes)
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1099 else:
1100 data = None
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1101 # throw away additional data [see bug #427345]
1102 while select.select([self.rfile._sock], [], [], 0)[0]:
1103 if not self.rfile._sock.recv(1):
1104 break
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1105 stdout, stderr = p.communicate(data)
1106 self.wfile.write(stdout)
1107 if stderr:
1108 self.log_error('%s', stderr)
Brian Curtin938ece72010-11-05 15:08:19 +0000[diff] [blame]1109 p.stderr.close()
1110 p.stdout.close()
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1111 status = p.returncode
1112 if status:
1113 self.log_error("CGI script exit status %#x", status)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1114 else:
1115 self.log_message("CGI script exited OK")
1116
1117
1118def test(HandlerClass = BaseHTTPRequestHandler,
1119 ServerClass = HTTPServer, protocol="HTTP/1.0"):
1120 """Test the HTTP request handler class.
1121
1122 This runs an HTTP server on port 8000 (or the first command line
1123 argument).
1124
1125 """
1126
1127 if sys.argv[1:]:
1128 port = int(sys.argv[1])
1129 else:
1130 port = 8000
1131 server_address = ('', port)
1132
1133 HandlerClass.protocol_version = protocol
1134 httpd = ServerClass(server_address, HandlerClass)
1135
1136 sa = httpd.socket.getsockname()
1137 print("Serving HTTP on", sa[0], "port", sa[1], "...")
Alexandre Vassalottib5292a22009-04-03 07:16:55 +0000[diff] [blame]1138 try:
1139 httpd.serve_forever()
1140 except KeyboardInterrupt:
1141 print("\nKeyboard interrupt received, exiting.")
1142 httpd.server_close()
1143 sys.exit(0)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1144
1145if __name__ == '__main__':
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1146 test(HandlerClass=SimpleHTTPRequestHandler)