Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / python / cpython3 / 5e8826cd981b7bfba8f5f28cddc602bb211fb2f8 / . / Lib / http / server.py
blob: a4acff0dfa18d245beda9cbd03b0e1bf82968f70 [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.
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
326 return True
327
328 def handle_one_request(self):
329 """Handle a single HTTP request.
330
331 You normally don't need to override this method; see the class
332 __doc__ string for information on how to handle specific HTTP
333 commands such as GET and POST.
334
335 """
336 self.raw_requestline = self.rfile.readline()
337 if not self.raw_requestline:
338 self.close_connection = 1
339 return
340 if not self.parse_request(): # An error code has been sent, just exit
341 return
342 mname = 'do_' + self.command
343 if not hasattr(self, mname):
344 self.send_error(501, "Unsupported method (%r)" % self.command)
345 return
346 method = getattr(self, mname)
347 method()
348
349 def handle(self):
350 """Handle multiple requests if necessary."""
351 self.close_connection = 1
352
353 self.handle_one_request()
354 while not self.close_connection:
355 self.handle_one_request()
356
357 def send_error(self, code, message=None):
358 """Send and log an error reply.
359
360 Arguments are the error code, and a detailed message.
361 The detailed message defaults to the short entry matching the
362 response code.
363
364 This sends an error response (so it must be called before any
365 output has been generated), logs the error, and finally sends
366 a piece of HTML explaining the error to the user.
367
368 """
369
370 try:
371 shortmsg, longmsg = self.responses[code]
372 except KeyError:
373 shortmsg, longmsg = '???', '???'
374 if message is None:
375 message = shortmsg
376 explain = longmsg
377 self.log_error("code %d, message %s", code, message)
378 # using _quote_html to prevent Cross Site Scripting attacks (see bug #1100201)
379 content = (self.error_message_format %
380 {'code': code, 'message': _quote_html(message), 'explain': explain})
381 self.send_response(code, message)
382 self.send_header("Content-Type", self.error_content_type)
383 self.send_header('Connection', 'close')
384 self.end_headers()
385 if self.command != 'HEAD' and code >= 200 and code not in (204, 304):
386 self.wfile.write(content.encode('UTF-8', 'replace'))
387
388 def send_response(self, code, message=None):
389 """Send the response header and log the response code.
390
391 Also send two standard headers with the server software
392 version and the current date.
393
394 """
395 self.log_request(code)
396 if message is None:
397 if code in self.responses:
398 message = self.responses[code][0]
399 else:
400 message = ''
401 if self.request_version != 'HTTP/0.9':
402 self.wfile.write(("%s %d %s\r\n" %
403 (self.protocol_version, code, message)).encode('ASCII', 'strict'))
404 # print (self.protocol_version, code, message)
405 self.send_header('Server', self.version_string())
406 self.send_header('Date', self.date_time_string())
407
408 def send_header(self, keyword, value):
409 """Send a MIME header."""
410 if self.request_version != 'HTTP/0.9':
411 self.wfile.write(("%s: %s\r\n" % (keyword, value)).encode('ASCII', 'strict'))
412
413 if keyword.lower() == 'connection':
414 if value.lower() == 'close':
415 self.close_connection = 1
416 elif value.lower() == 'keep-alive':
417 self.close_connection = 0
418
419 def end_headers(self):
420 """Send the blank line ending the MIME headers."""
421 if self.request_version != 'HTTP/0.9':
422 self.wfile.write(b"\r\n")
423
424 def log_request(self, code='-', size='-'):
425 """Log an accepted request.
426
427 This is called by send_response().
428
429 """
430
431 self.log_message('"%s" %s %s',
432 self.requestline, str(code), str(size))
433
434 def log_error(self, format, *args):
435 """Log an error.
436
437 This is called when a request cannot be fulfilled. By
438 default it passes the message on to log_message().
439
440 Arguments are the same as for log_message().
441
442 XXX This should go to the separate error log.
443
444 """
445
446 self.log_message(format, *args)
447
448 def log_message(self, format, *args):
449 """Log an arbitrary message.
450
451 This is used by all other logging functions. Override
452 it if you have specific logging wishes.
453
454 The first argument, FORMAT, is a format string for the
455 message to be logged. If the format string contains
456 any % escapes requiring parameters, they should be
457 specified as subsequent arguments (it's just like
458 printf!).
459
460 The client host and current date/time are prefixed to
461 every message.
462
463 """
464
465 sys.stderr.write("%s - - [%s] %s\n" %
466 (self.address_string(),
467 self.log_date_time_string(),
468 format%args))
469
470 def version_string(self):
471 """Return the server software version string."""
472 return self.server_version + ' ' + self.sys_version
473
474 def date_time_string(self, timestamp=None):
475 """Return the current date and time formatted for a message header."""
476 if timestamp is None:
477 timestamp = time.time()
478 year, month, day, hh, mm, ss, wd, y, z = time.gmtime(timestamp)
479 s = "%s, %02d %3s %4d %02d:%02d:%02d GMT" % (
480 self.weekdayname[wd],
481 day, self.monthname[month], year,
482 hh, mm, ss)
483 return s
484
485 def log_date_time_string(self):
486 """Return the current time formatted for logging."""
487 now = time.time()
488 year, month, day, hh, mm, ss, x, y, z = time.localtime(now)
489 s = "%02d/%3s/%04d %02d:%02d:%02d" % (
490 day, self.monthname[month], year, hh, mm, ss)
491 return s
492
493 weekdayname = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
494
495 monthname = [None,
496 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun',
497 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
498
499 def address_string(self):
500 """Return the client address formatted for logging.
501
502 This version looks up the full hostname using gethostbyaddr(),
503 and tries to find a name that contains at least one dot.
504
505 """
506
507 host, port = self.client_address[:2]
508 return socket.getfqdn(host)
509
510 # Essentially static class variables
511
512 # The version of the HTTP protocol we support.
513 # Set this to HTTP/1.1 to enable automatic keepalive
514 protocol_version = "HTTP/1.0"
515
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]516 # MessageClass used to parse headers
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]517 MessageClass = http.client.HTTPMessage
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]518
519 # Table mapping response codes to messages; entries have the
520 # form {code: (shortmessage, longmessage)}.
521 # See RFC 2616.
522 responses = {
523 100: ('Continue', 'Request received, please continue'),
524 101: ('Switching Protocols',
525 'Switching to new protocol; obey Upgrade header'),
526
527 200: ('OK', 'Request fulfilled, document follows'),
528 201: ('Created', 'Document created, URL follows'),
529 202: ('Accepted',
530 'Request accepted, processing continues off-line'),
531 203: ('Non-Authoritative Information', 'Request fulfilled from cache'),
532 204: ('No Content', 'Request fulfilled, nothing follows'),
533 205: ('Reset Content', 'Clear input form for further input.'),
534 206: ('Partial Content', 'Partial content follows.'),
535
536 300: ('Multiple Choices',
537 'Object has several resources -- see URI list'),
538 301: ('Moved Permanently', 'Object moved permanently -- see URI list'),
539 302: ('Found', 'Object moved temporarily -- see URI list'),
540 303: ('See Other', 'Object moved -- see Method and URL list'),
541 304: ('Not Modified',
542 'Document has not changed since given time'),
543 305: ('Use Proxy',
544 'You must use proxy specified in Location to access this '
545 'resource.'),
546 307: ('Temporary Redirect',
547 'Object moved temporarily -- see URI list'),
548
549 400: ('Bad Request',
550 'Bad request syntax or unsupported method'),
551 401: ('Unauthorized',
552 'No permission -- see authorization schemes'),
553 402: ('Payment Required',
554 'No payment -- see charging schemes'),
555 403: ('Forbidden',
556 'Request forbidden -- authorization will not help'),
557 404: ('Not Found', 'Nothing matches the given URI'),
558 405: ('Method Not Allowed',
Senthil Kumaran613c61c2010-02-22 11:02:53 +0000[diff] [blame]559 'Specified method is invalid for this resource.'),
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]560 406: ('Not Acceptable', 'URI not available in preferred format.'),
561 407: ('Proxy Authentication Required', 'You must authenticate with '
562 'this proxy before proceeding.'),
563 408: ('Request Timeout', 'Request timed out; try again later.'),
564 409: ('Conflict', 'Request conflict.'),
565 410: ('Gone',
566 'URI no longer exists and has been permanently removed.'),
567 411: ('Length Required', 'Client must specify Content-Length.'),
568 412: ('Precondition Failed', 'Precondition in headers is false.'),
569 413: ('Request Entity Too Large', 'Entity is too large.'),
570 414: ('Request-URI Too Long', 'URI is too long.'),
571 415: ('Unsupported Media Type', 'Entity body in unsupported format.'),
572 416: ('Requested Range Not Satisfiable',
573 'Cannot satisfy request range.'),
574 417: ('Expectation Failed',
575 'Expect condition could not be satisfied.'),
576
577 500: ('Internal Server Error', 'Server got itself in trouble'),
578 501: ('Not Implemented',
579 'Server does not support this operation'),
580 502: ('Bad Gateway', 'Invalid responses from another server/proxy.'),
581 503: ('Service Unavailable',
582 'The server cannot process the request due to a high load'),
583 504: ('Gateway Timeout',
584 'The gateway server did not receive a timely response'),
585 505: ('HTTP Version Not Supported', 'Cannot fulfill request.'),
586 }
587
588
589class SimpleHTTPRequestHandler(BaseHTTPRequestHandler):
590
591 """Simple HTTP request handler with GET and HEAD commands.
592
593 This serves files from the current directory and any of its
594 subdirectories. The MIME type for files is determined by
595 calling the .guess_type() method.
596
597 The GET and HEAD requests are identical except that the HEAD
598 request omits the actual contents of the file.
599
600 """
601
602 server_version = "SimpleHTTP/" + __version__
603
604 def do_GET(self):
605 """Serve a GET request."""
606 f = self.send_head()
607 if f:
608 self.copyfile(f, self.wfile)
609 f.close()
610
611 def do_HEAD(self):
612 """Serve a HEAD request."""
613 f = self.send_head()
614 if f:
615 f.close()
616
617 def send_head(self):
618 """Common code for GET and HEAD commands.
619
620 This sends the response code and MIME headers.
621
622 Return value is either a file object (which has to be copied
623 to the outputfile by the caller unless the command was HEAD,
624 and must be closed by the caller under all circumstances), or
625 None, in which case the caller has nothing further to do.
626
627 """
628 path = self.translate_path(self.path)
629 f = None
630 if os.path.isdir(path):
631 if not self.path.endswith('/'):
632 # redirect browser - doing basically what apache does
633 self.send_response(301)
634 self.send_header("Location", self.path + "/")
635 self.end_headers()
636 return None
637 for index in "index.html", "index.htm":
638 index = os.path.join(path, index)
639 if os.path.exists(index):
640 path = index
641 break
642 else:
643 return self.list_directory(path)
644 ctype = self.guess_type(path)
645 try:
646 f = open(path, 'rb')
647 except IOError:
648 self.send_error(404, "File not found")
649 return None
650 self.send_response(200)
651 self.send_header("Content-type", ctype)
652 fs = os.fstat(f.fileno())
653 self.send_header("Content-Length", str(fs[6]))
654 self.send_header("Last-Modified", self.date_time_string(fs.st_mtime))
655 self.end_headers()
656 return f
657
658 def list_directory(self, path):
659 """Helper to produce a directory listing (absent index.html).
660
661 Return value is either a file object, or None (indicating an
662 error). In either case, the headers are sent, making the
663 interface the same as for send_head().
664
665 """
666 try:
667 list = os.listdir(path)
668 except os.error:
669 self.send_error(404, "No permission to list directory")
670 return None
671 list.sort(key=lambda a: a.lower())
672 r = []
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]673 displaypath = cgi.escape(urllib.parse.unquote(self.path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]674 r.append('<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">')
675 r.append("<html>\n<title>Directory listing for %s</title>\n" % displaypath)
676 r.append("<body>\n<h2>Directory listing for %s</h2>\n" % displaypath)
677 r.append("<hr>\n<ul>\n")
678 for name in list:
679 fullname = os.path.join(path, name)
680 displayname = linkname = name
681 # Append / for directories or @ for symbolic links
682 if os.path.isdir(fullname):
683 displayname = name + "/"
684 linkname = name + "/"
685 if os.path.islink(fullname):
686 displayname = name + "@"
687 # Note: a link to a directory displays with @ and links with /
688 r.append('<li><a href="%s">%s</a>\n'
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]689 % (urllib.parse.quote(linkname), cgi.escape(displayname)))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]690 r.append("</ul>\n<hr>\n</body>\n</html>\n")
691 enc = sys.getfilesystemencoding()
692 encoded = ''.join(r).encode(enc)
693 f = io.BytesIO()
694 f.write(encoded)
695 f.seek(0)
696 self.send_response(200)
697 self.send_header("Content-type", "text/html; charset=%s" % enc)
698 self.send_header("Content-Length", str(len(encoded)))
699 self.end_headers()
700 return f
701
702 def translate_path(self, path):
703 """Translate a /-separated PATH to the local filename syntax.
704
705 Components that mean special things to the local file system
706 (e.g. drive or directory names) are ignored. (XXX They should
707 probably be diagnosed.)
708
709 """
710 # abandon query parameters
711 path = path.split('?',1)[0]
712 path = path.split('#',1)[0]
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]713 path = posixpath.normpath(urllib.parse.unquote(path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]714 words = path.split('/')
715 words = filter(None, words)
716 path = os.getcwd()
717 for word in words:
718 drive, word = os.path.splitdrive(word)
719 head, word = os.path.split(word)
720 if word in (os.curdir, os.pardir): continue
721 path = os.path.join(path, word)
722 return path
723
724 def copyfile(self, source, outputfile):
725 """Copy all data between two file objects.
726
727 The SOURCE argument is a file object open for reading
728 (or anything with a read() method) and the DESTINATION
729 argument is a file object open for writing (or
730 anything with a write() method).
731
732 The only reason for overriding this would be to change
733 the block size or perhaps to replace newlines by CRLF
734 -- note however that this the default server uses this
735 to copy binary data as well.
736
737 """
738 shutil.copyfileobj(source, outputfile)
739
740 def guess_type(self, path):
741 """Guess the type of a file.
742
743 Argument is a PATH (a filename).
744
745 Return value is a string of the form type/subtype,
746 usable for a MIME Content-type header.
747
748 The default implementation looks the file's extension
749 up in the table self.extensions_map, using application/octet-stream
750 as a default; however it would be permissible (if
751 slow) to look inside the data to make a better guess.
752
753 """
754
755 base, ext = posixpath.splitext(path)
756 if ext in self.extensions_map:
757 return self.extensions_map[ext]
758 ext = ext.lower()
759 if ext in self.extensions_map:
760 return self.extensions_map[ext]
761 else:
762 return self.extensions_map['']
763
764 if not mimetypes.inited:
765 mimetypes.init() # try to read system mime.types
766 extensions_map = mimetypes.types_map.copy()
767 extensions_map.update({
768 '': 'application/octet-stream', # Default
769 '.py': 'text/plain',
770 '.c': 'text/plain',
771 '.h': 'text/plain',
772 })
773
774
775# Utilities for CGIHTTPRequestHandler
776
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]777# TODO(gregory.p.smith): Move this into an appropriate library.
778def _url_collapse_path_split(path):
779 """
780 Given a URL path, remove extra '/'s and '.' path elements and collapse
781 any '..' references.
782
783 Implements something akin to RFC-2396 5.2 step 6 to parse relative paths.
784
785 Returns: A tuple of (head, tail) where tail is everything after the final /
786 and head is everything before it. Head will always start with a '/' and,
787 if it contains anything else, never have a trailing '/'.
788
789 Raises: IndexError if too many '..' occur within the path.
790 """
791 # Similar to os.path.split(os.path.normpath(path)) but specific to URL
792 # path semantics rather than local operating system semantics.
793 path_parts = []
794 for part in path.split('/'):
795 if part == '.':
796 path_parts.append('')
797 else:
798 path_parts.append(part)
799 # Filter out blank non trailing parts before consuming the '..'.
800 path_parts = [part for part in path_parts[:-1] if part] + path_parts[-1:]
801 if path_parts:
802 tail_part = path_parts.pop()
803 else:
804 tail_part = ''
805 head_parts = []
806 for part in path_parts:
807 if part == '..':
808 head_parts.pop()
809 else:
810 head_parts.append(part)
811 if tail_part and tail_part == '..':
812 head_parts.pop()
813 tail_part = ''
814 return ('/' + '/'.join(head_parts), tail_part)
815
816
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]817nobody = None
818
819def nobody_uid():
820 """Internal routine to get nobody's uid"""
821 global nobody
822 if nobody:
823 return nobody
824 try:
825 import pwd
826 except ImportError:
827 return -1
828 try:
829 nobody = pwd.getpwnam('nobody')[2]
830 except KeyError:
831 nobody = 1 + max(map(lambda x: x[2], pwd.getpwall()))
832 return nobody
833
834
835def executable(path):
836 """Test for executable file."""
837 try:
838 st = os.stat(path)
839 except os.error:
840 return False
841 return st.st_mode & 0o111 != 0
842
843
844class CGIHTTPRequestHandler(SimpleHTTPRequestHandler):
845
846 """Complete HTTP server with GET, HEAD and POST commands.
847
848 GET and HEAD also support running CGI scripts.
849
850 The POST command is *only* implemented for CGI scripts.
851
852 """
853
854 # Determine platform specifics
855 have_fork = hasattr(os, 'fork')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]856
857 # Make rfile unbuffered -- we need to read one line and then pass
858 # the rest to a subprocess, so we can't use buffered input.
859 rbufsize = 0
860
861 def do_POST(self):
862 """Serve a POST request.
863
864 This is only implemented for CGI scripts.
865
866 """
867
868 if self.is_cgi():
869 self.run_cgi()
870 else:
871 self.send_error(501, "Can only POST to CGI scripts")
872
873 def send_head(self):
874 """Version of send_head that support CGI scripts"""
875 if self.is_cgi():
876 return self.run_cgi()
877 else:
878 return SimpleHTTPRequestHandler.send_head(self)
879
880 def is_cgi(self):
881 """Test whether self.path corresponds to a CGI script.
882
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]883 Returns True and updates the cgi_info attribute to the tuple
884 (dir, rest) if self.path requires running a CGI script.
885 Returns False otherwise.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]886
Benjamin Petersona7deeee2009-05-08 20:54:42 +0000[diff] [blame]887 If any exception is raised, the caller should assume that
888 self.path was rejected as invalid and act accordingly.
889
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]890 The default implementation tests whether the normalized url
891 path begins with one of the strings in self.cgi_directories
892 (and the next character is a '/' or the end of the string).
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]893
894 """
895
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]896 splitpath = _url_collapse_path_split(self.path)
897 if splitpath[0] in self.cgi_directories:
898 self.cgi_info = splitpath
899 return True
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]900 return False
901
902 cgi_directories = ['/cgi-bin', '/htbin']
903
904 def is_executable(self, path):
905 """Test whether argument path is an executable file."""
906 return executable(path)
907
908 def is_python(self, path):
909 """Test whether argument path is a Python script."""
910 head, tail = os.path.splitext(path)
911 return tail.lower() in (".py", ".pyw")
912
913 def run_cgi(self):
914 """Execute a CGI script."""
915 path = self.path
916 dir, rest = self.cgi_info
917
918 i = path.find('/', len(dir) + 1)
919 while i >= 0:
920 nextdir = path[:i]
921 nextrest = path[i+1:]
922
923 scriptdir = self.translate_path(nextdir)
924 if os.path.isdir(scriptdir):
925 dir, rest = nextdir, nextrest
926 i = path.find('/', len(dir) + 1)
927 else:
928 break
929
930 # find an explicit query string, if present.
931 i = rest.rfind('?')
932 if i >= 0:
933 rest, query = rest[:i], rest[i+1:]
934 else:
935 query = ''
936
937 # dissect the part after the directory name into a script name &
938 # a possible additional path, to be stored in PATH_INFO.
939 i = rest.find('/')
940 if i >= 0:
941 script, rest = rest[:i], rest[i:]
942 else:
943 script, rest = rest, ''
944
945 scriptname = dir + '/' + script
946 scriptfile = self.translate_path(scriptname)
947 if not os.path.exists(scriptfile):
948 self.send_error(404, "No such CGI script (%r)" % scriptname)
949 return
950 if not os.path.isfile(scriptfile):
951 self.send_error(403, "CGI script is not a plain file (%r)" %
952 scriptname)
953 return
954 ispy = self.is_python(scriptname)
955 if not ispy:
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]956 if not self.is_executable(scriptfile):
957 self.send_error(403, "CGI script is not executable (%r)" %
958 scriptname)
959 return
960
961 # Reference: http://hoohoo.ncsa.uiuc.edu/cgi/env.html
962 # XXX Much of the following could be prepared ahead of time!
Senthil Kumaran5e8826c2010-10-03 18:04:52 +0000[diff] [blame^]963 env = copy.deepcopy(os.environ)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]964 env['SERVER_SOFTWARE'] = self.version_string()
965 env['SERVER_NAME'] = self.server.server_name
966 env['GATEWAY_INTERFACE'] = 'CGI/1.1'
967 env['SERVER_PROTOCOL'] = self.protocol_version
968 env['SERVER_PORT'] = str(self.server.server_port)
969 env['REQUEST_METHOD'] = self.command
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]970 uqrest = urllib.parse.unquote(rest)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]971 env['PATH_INFO'] = uqrest
972 env['PATH_TRANSLATED'] = self.translate_path(uqrest)
973 env['SCRIPT_NAME'] = scriptname
974 if query:
975 env['QUERY_STRING'] = query
976 host = self.address_string()
977 if host != self.client_address[0]:
978 env['REMOTE_HOST'] = host
979 env['REMOTE_ADDR'] = self.client_address[0]
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]980 authorization = self.headers.get("authorization")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]981 if authorization:
982 authorization = authorization.split()
983 if len(authorization) == 2:
984 import base64, binascii
985 env['AUTH_TYPE'] = authorization[0]
986 if authorization[0].lower() == "basic":
987 try:
988 authorization = authorization[1].encode('ascii')
Georg Brandl706824f2009-06-04 09:42:55 +0000[diff] [blame]989 authorization = base64.decodebytes(authorization).\
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]990 decode('ascii')
991 except (binascii.Error, UnicodeError):
992 pass
993 else:
994 authorization = authorization.split(':')
995 if len(authorization) == 2:
996 env['REMOTE_USER'] = authorization[0]
997 # XXX REMOTE_IDENT
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]998 if self.headers.get('content-type') is None:
999 env['CONTENT_TYPE'] = self.headers.get_content_type()
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1000 else:
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1001 env['CONTENT_TYPE'] = self.headers['content-type']
1002 length = self.headers.get('content-length')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1003 if length:
1004 env['CONTENT_LENGTH'] = length
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1005 referer = self.headers.get('referer')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1006 if referer:
1007 env['HTTP_REFERER'] = referer
1008 accept = []
1009 for line in self.headers.getallmatchingheaders('accept'):
1010 if line[:1] in "\t\n\r ":
1011 accept.append(line.strip())
1012 else:
1013 accept = accept + line[7:].split(',')
1014 env['HTTP_ACCEPT'] = ','.join(accept)
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1015 ua = self.headers.get('user-agent')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1016 if ua:
1017 env['HTTP_USER_AGENT'] = ua
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1018 co = filter(None, self.headers.get_all('cookie', []))
Georg Brandlcaa78fe2010-08-01 19:07:28 +0000[diff] [blame]1019 cookie_str = ', '.join(co)
1020 if cookie_str:
1021 env['HTTP_COOKIE'] = cookie_str
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1022 # XXX Other HTTP_* headers
1023 # Since we're setting the env in the parent, provide empty
1024 # values to override previously set values
1025 for k in ('QUERY_STRING', 'REMOTE_HOST', 'CONTENT_LENGTH',
1026 'HTTP_USER_AGENT', 'HTTP_COOKIE', 'HTTP_REFERER'):
1027 env.setdefault(k, "")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1028
1029 self.send_response(200, "Script output follows")
1030
1031 decoded_query = query.replace('+', ' ')
1032
1033 if self.have_fork:
1034 # Unix -- fork as we should
1035 args = [script]
1036 if '=' not in decoded_query:
1037 args.append(decoded_query)
1038 nobody = nobody_uid()
1039 self.wfile.flush() # Always flush before forking
1040 pid = os.fork()
1041 if pid != 0:
1042 # Parent
1043 pid, sts = os.waitpid(pid, 0)
1044 # throw away additional data [see bug #427345]
1045 while select.select([self.rfile], [], [], 0)[0]:
1046 if not self.rfile.read(1):
1047 break
1048 if sts:
1049 self.log_error("CGI script exit status %#x", sts)
1050 return
1051 # Child
1052 try:
1053 try:
1054 os.setuid(nobody)
1055 except os.error:
1056 pass
1057 os.dup2(self.rfile.fileno(), 0)
1058 os.dup2(self.wfile.fileno(), 1)
Senthil Kumaran5e8826c2010-10-03 18:04:52 +0000[diff] [blame^]1059 os.execve(scriptfile, args, env)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1060 except:
1061 self.server.handle_error(self.request, self.client_address)
1062 os._exit(127)
1063
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1064 else:
1065 # Non-Unix -- use subprocess
1066 import subprocess
Senthil Kumaranca5130c2009-11-11 04:21:22 +0000[diff] [blame]1067 cmdline = [scriptfile]
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1068 if self.is_python(scriptfile):
1069 interp = sys.executable
1070 if interp.lower().endswith("w.exe"):
1071 # On Windows, use python.exe, not pythonw.exe
1072 interp = interp[:-5] + interp[-4:]
Senthil Kumaranca5130c2009-11-11 04:21:22 +0000[diff] [blame]1073 cmdline = [interp, '-u'] + cmdline
1074 if '=' not in query:
1075 cmdline.append(query)
1076 self.log_message("command: %s", subprocess.list2cmdline(cmdline))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1077 try:
1078 nbytes = int(length)
1079 except (TypeError, ValueError):
1080 nbytes = 0
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1081 p = subprocess.Popen(cmdline,
1082 stdin=subprocess.PIPE,
1083 stdout=subprocess.PIPE,
Senthil Kumaran5e8826c2010-10-03 18:04:52 +0000[diff] [blame^]1084 stderr=subprocess.PIPE,
1085 env = env
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1086 )
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1087 if self.command.lower() == "post" and nbytes > 0:
1088 data = self.rfile.read(nbytes)
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1089 else:
1090 data = None
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1091 # throw away additional data [see bug #427345]
1092 while select.select([self.rfile._sock], [], [], 0)[0]:
1093 if not self.rfile._sock.recv(1):
1094 break
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1095 stdout, stderr = p.communicate(data)
1096 self.wfile.write(stdout)
1097 if stderr:
1098 self.log_error('%s', stderr)
1099 status = p.returncode
1100 if status:
1101 self.log_error("CGI script exit status %#x", status)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1102 else:
1103 self.log_message("CGI script exited OK")
1104
1105
1106def test(HandlerClass = BaseHTTPRequestHandler,
1107 ServerClass = HTTPServer, protocol="HTTP/1.0"):
1108 """Test the HTTP request handler class.
1109
1110 This runs an HTTP server on port 8000 (or the first command line
1111 argument).
1112
1113 """
1114
1115 if sys.argv[1:]:
1116 port = int(sys.argv[1])
1117 else:
1118 port = 8000
1119 server_address = ('', port)
1120
1121 HandlerClass.protocol_version = protocol
1122 httpd = ServerClass(server_address, HandlerClass)
1123
1124 sa = httpd.socket.getsockname()
1125 print("Serving HTTP on", sa[0], "port", sa[1], "...")
Alexandre Vassalottib5292a22009-04-03 07:16:55 +0000[diff] [blame]1126 try:
1127 httpd.serve_forever()
1128 except KeyboardInterrupt:
1129 print("\nKeyboard interrupt received, exiting.")
1130 httpd.server_close()
1131 sys.exit(0)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1132
1133if __name__ == '__main__':
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1134 test(HandlerClass=SimpleHTTPRequestHandler)