Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / python / cpython3 / 9f0f960d4ce48062dcb286903721fa89d6037159 / . / Lib / http / server.py
blob: 35ade6c015245f96ca4462f477cc6e59972a1621 [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),
16os.popen2() is used as a fallback, with slightly altered semantics; if
17that function is not present either (e.g. on Macintosh), only Python
18scripts are supported, and they are executed by the current process.
19
20In all cases, the implementation is intentionally naive -- all
21requests are executed synchronously.
22
23SECURITY WARNING: DON'T USE THIS CODE UNLESS YOU ARE INSIDE A FIREWALL
24-- it may execute arbitrary Python code or external programs.
25
26Note that status code 200 is sent prior to execution of a CGI script, so
27scripts cannot send other status codes such as 302 (redirect).
28
29XXX To do:
30
31- log requests even later (to capture byte count)
32- log user-agent header and other interesting goodies
33- send error log to separate file
34"""
35
36
37# See also:
38#
39# HTTP Working Group T. Berners-Lee
40# INTERNET-DRAFT R. T. Fielding
41# <draft-ietf-http-v10-spec-00.txt> H. Frystyk Nielsen
42# Expires September 8, 1995 March 8, 1995
43#
44# URL: http://www.ics.uci.edu/pub/ietf/http/draft-ietf-http-v10-spec-00.txt
45#
46# and
47#
48# Network Working Group R. Fielding
49# Request for Comments: 2616 et al
50# Obsoletes: 2068 June 1999
51# Category: Standards Track
52#
53# URL: http://www.faqs.org/rfcs/rfc2616.html
54
55# Log files
56# ---------
57#
58# Here's a quote from the NCSA httpd docs about log file format.
59#
60# | The logfile format is as follows. Each line consists of:
61# |
62# | host rfc931 authuser [DD/Mon/YYYY:hh:mm:ss] "request" ddd bbbb
63# |
64# | host: Either the DNS name or the IP number of the remote client
65# | rfc931: Any information returned by identd for this person,
66# | - otherwise.
67# | authuser: If user sent a userid for authentication, the user name,
68# | - otherwise.
69# | DD: Day
70# | Mon: Month (calendar name)
71# | YYYY: Year
72# | hh: hour (24-hour format, the machine's timezone)
73# | mm: minutes
74# | ss: seconds
75# | request: The first line of the HTTP request as sent by the client.
76# | ddd: the status code returned by the server, - if not available.
77# | bbbb: the total number of bytes sent,
78# | *not including the HTTP/1.0 header*, - if not available
79# |
80# | You can determine the name of the file accessed through request.
81#
82# (Actually, the latter is only true if you know the server configuration
83# at the time the request was made!)
84
85__version__ = "0.6"
86
87__all__ = ["HTTPServer", "BaseHTTPRequestHandler"]
88
89import io
90import os
91import sys
92import cgi
93import time
94import socket # For gethostbyaddr()
95import shutil
96import urllib
97import select
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]98import mimetypes
99import posixpath
100import socketserver
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]101import email.message
102import email.parser
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.
317
Georg Brandl9f0f9602008-06-12 22:23:59 +0000[diff] [blame^]318 # MessageClass wants to see strings rather than bytes.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]319 # But a TextIOWrapper around self.rfile would buffer too many bytes
320 # from the stream, bytes which we later need to read as bytes.
321 # So we read the correct bytes here, as bytes, then use StringIO
322 # to make them look like strings for MessageClass to parse.
323 headers = []
324 while True:
325 line = self.rfile.readline()
326 headers.append(line)
327 if line in (b'\r\n', b'\n', b''):
328 break
329 hfile = io.StringIO(b''.join(headers).decode('iso-8859-1'))
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]330 self.headers = email.parser.Parser(_class=self.MessageClass).parse(hfile)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]331
332 conntype = self.headers.get('Connection', "")
333 if conntype.lower() == 'close':
334 self.close_connection = 1
335 elif (conntype.lower() == 'keep-alive' and
336 self.protocol_version >= "HTTP/1.1"):
337 self.close_connection = 0
338 return True
339
340 def handle_one_request(self):
341 """Handle a single HTTP request.
342
343 You normally don't need to override this method; see the class
344 __doc__ string for information on how to handle specific HTTP
345 commands such as GET and POST.
346
347 """
348 self.raw_requestline = self.rfile.readline()
349 if not self.raw_requestline:
350 self.close_connection = 1
351 return
352 if not self.parse_request(): # An error code has been sent, just exit
353 return
354 mname = 'do_' + self.command
355 if not hasattr(self, mname):
356 self.send_error(501, "Unsupported method (%r)" % self.command)
357 return
358 method = getattr(self, mname)
359 method()
360
361 def handle(self):
362 """Handle multiple requests if necessary."""
363 self.close_connection = 1
364
365 self.handle_one_request()
366 while not self.close_connection:
367 self.handle_one_request()
368
369 def send_error(self, code, message=None):
370 """Send and log an error reply.
371
372 Arguments are the error code, and a detailed message.
373 The detailed message defaults to the short entry matching the
374 response code.
375
376 This sends an error response (so it must be called before any
377 output has been generated), logs the error, and finally sends
378 a piece of HTML explaining the error to the user.
379
380 """
381
382 try:
383 shortmsg, longmsg = self.responses[code]
384 except KeyError:
385 shortmsg, longmsg = '???', '???'
386 if message is None:
387 message = shortmsg
388 explain = longmsg
389 self.log_error("code %d, message %s", code, message)
390 # using _quote_html to prevent Cross Site Scripting attacks (see bug #1100201)
391 content = (self.error_message_format %
392 {'code': code, 'message': _quote_html(message), 'explain': explain})
393 self.send_response(code, message)
394 self.send_header("Content-Type", self.error_content_type)
395 self.send_header('Connection', 'close')
396 self.end_headers()
397 if self.command != 'HEAD' and code >= 200 and code not in (204, 304):
398 self.wfile.write(content.encode('UTF-8', 'replace'))
399
400 def send_response(self, code, message=None):
401 """Send the response header and log the response code.
402
403 Also send two standard headers with the server software
404 version and the current date.
405
406 """
407 self.log_request(code)
408 if message is None:
409 if code in self.responses:
410 message = self.responses[code][0]
411 else:
412 message = ''
413 if self.request_version != 'HTTP/0.9':
414 self.wfile.write(("%s %d %s\r\n" %
415 (self.protocol_version, code, message)).encode('ASCII', 'strict'))
416 # print (self.protocol_version, code, message)
417 self.send_header('Server', self.version_string())
418 self.send_header('Date', self.date_time_string())
419
420 def send_header(self, keyword, value):
421 """Send a MIME header."""
422 if self.request_version != 'HTTP/0.9':
423 self.wfile.write(("%s: %s\r\n" % (keyword, value)).encode('ASCII', 'strict'))
424
425 if keyword.lower() == 'connection':
426 if value.lower() == 'close':
427 self.close_connection = 1
428 elif value.lower() == 'keep-alive':
429 self.close_connection = 0
430
431 def end_headers(self):
432 """Send the blank line ending the MIME headers."""
433 if self.request_version != 'HTTP/0.9':
434 self.wfile.write(b"\r\n")
435
436 def log_request(self, code='-', size='-'):
437 """Log an accepted request.
438
439 This is called by send_response().
440
441 """
442
443 self.log_message('"%s" %s %s',
444 self.requestline, str(code), str(size))
445
446 def log_error(self, format, *args):
447 """Log an error.
448
449 This is called when a request cannot be fulfilled. By
450 default it passes the message on to log_message().
451
452 Arguments are the same as for log_message().
453
454 XXX This should go to the separate error log.
455
456 """
457
458 self.log_message(format, *args)
459
460 def log_message(self, format, *args):
461 """Log an arbitrary message.
462
463 This is used by all other logging functions. Override
464 it if you have specific logging wishes.
465
466 The first argument, FORMAT, is a format string for the
467 message to be logged. If the format string contains
468 any % escapes requiring parameters, they should be
469 specified as subsequent arguments (it's just like
470 printf!).
471
472 The client host and current date/time are prefixed to
473 every message.
474
475 """
476
477 sys.stderr.write("%s - - [%s] %s\n" %
478 (self.address_string(),
479 self.log_date_time_string(),
480 format%args))
481
482 def version_string(self):
483 """Return the server software version string."""
484 return self.server_version + ' ' + self.sys_version
485
486 def date_time_string(self, timestamp=None):
487 """Return the current date and time formatted for a message header."""
488 if timestamp is None:
489 timestamp = time.time()
490 year, month, day, hh, mm, ss, wd, y, z = time.gmtime(timestamp)
491 s = "%s, %02d %3s %4d %02d:%02d:%02d GMT" % (
492 self.weekdayname[wd],
493 day, self.monthname[month], year,
494 hh, mm, ss)
495 return s
496
497 def log_date_time_string(self):
498 """Return the current time formatted for logging."""
499 now = time.time()
500 year, month, day, hh, mm, ss, x, y, z = time.localtime(now)
501 s = "%02d/%3s/%04d %02d:%02d:%02d" % (
502 day, self.monthname[month], year, hh, mm, ss)
503 return s
504
505 weekdayname = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
506
507 monthname = [None,
508 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun',
509 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
510
511 def address_string(self):
512 """Return the client address formatted for logging.
513
514 This version looks up the full hostname using gethostbyaddr(),
515 and tries to find a name that contains at least one dot.
516
517 """
518
519 host, port = self.client_address[:2]
520 return socket.getfqdn(host)
521
522 # Essentially static class variables
523
524 # The version of the HTTP protocol we support.
525 # Set this to HTTP/1.1 to enable automatic keepalive
526 protocol_version = "HTTP/1.0"
527
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]528 # MessageClass used to parse headers
529 import http.client
530 MessageClass = http.client.HTTPMessage
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]531
532 # Table mapping response codes to messages; entries have the
533 # form {code: (shortmessage, longmessage)}.
534 # See RFC 2616.
535 responses = {
536 100: ('Continue', 'Request received, please continue'),
537 101: ('Switching Protocols',
538 'Switching to new protocol; obey Upgrade header'),
539
540 200: ('OK', 'Request fulfilled, document follows'),
541 201: ('Created', 'Document created, URL follows'),
542 202: ('Accepted',
543 'Request accepted, processing continues off-line'),
544 203: ('Non-Authoritative Information', 'Request fulfilled from cache'),
545 204: ('No Content', 'Request fulfilled, nothing follows'),
546 205: ('Reset Content', 'Clear input form for further input.'),
547 206: ('Partial Content', 'Partial content follows.'),
548
549 300: ('Multiple Choices',
550 'Object has several resources -- see URI list'),
551 301: ('Moved Permanently', 'Object moved permanently -- see URI list'),
552 302: ('Found', 'Object moved temporarily -- see URI list'),
553 303: ('See Other', 'Object moved -- see Method and URL list'),
554 304: ('Not Modified',
555 'Document has not changed since given time'),
556 305: ('Use Proxy',
557 'You must use proxy specified in Location to access this '
558 'resource.'),
559 307: ('Temporary Redirect',
560 'Object moved temporarily -- see URI list'),
561
562 400: ('Bad Request',
563 'Bad request syntax or unsupported method'),
564 401: ('Unauthorized',
565 'No permission -- see authorization schemes'),
566 402: ('Payment Required',
567 'No payment -- see charging schemes'),
568 403: ('Forbidden',
569 'Request forbidden -- authorization will not help'),
570 404: ('Not Found', 'Nothing matches the given URI'),
571 405: ('Method Not Allowed',
572 'Specified method is invalid for this server.'),
573 406: ('Not Acceptable', 'URI not available in preferred format.'),
574 407: ('Proxy Authentication Required', 'You must authenticate with '
575 'this proxy before proceeding.'),
576 408: ('Request Timeout', 'Request timed out; try again later.'),
577 409: ('Conflict', 'Request conflict.'),
578 410: ('Gone',
579 'URI no longer exists and has been permanently removed.'),
580 411: ('Length Required', 'Client must specify Content-Length.'),
581 412: ('Precondition Failed', 'Precondition in headers is false.'),
582 413: ('Request Entity Too Large', 'Entity is too large.'),
583 414: ('Request-URI Too Long', 'URI is too long.'),
584 415: ('Unsupported Media Type', 'Entity body in unsupported format.'),
585 416: ('Requested Range Not Satisfiable',
586 'Cannot satisfy request range.'),
587 417: ('Expectation Failed',
588 'Expect condition could not be satisfied.'),
589
590 500: ('Internal Server Error', 'Server got itself in trouble'),
591 501: ('Not Implemented',
592 'Server does not support this operation'),
593 502: ('Bad Gateway', 'Invalid responses from another server/proxy.'),
594 503: ('Service Unavailable',
595 'The server cannot process the request due to a high load'),
596 504: ('Gateway Timeout',
597 'The gateway server did not receive a timely response'),
598 505: ('HTTP Version Not Supported', 'Cannot fulfill request.'),
599 }
600
601
602class SimpleHTTPRequestHandler(BaseHTTPRequestHandler):
603
604 """Simple HTTP request handler with GET and HEAD commands.
605
606 This serves files from the current directory and any of its
607 subdirectories. The MIME type for files is determined by
608 calling the .guess_type() method.
609
610 The GET and HEAD requests are identical except that the HEAD
611 request omits the actual contents of the file.
612
613 """
614
615 server_version = "SimpleHTTP/" + __version__
616
617 def do_GET(self):
618 """Serve a GET request."""
619 f = self.send_head()
620 if f:
621 self.copyfile(f, self.wfile)
622 f.close()
623
624 def do_HEAD(self):
625 """Serve a HEAD request."""
626 f = self.send_head()
627 if f:
628 f.close()
629
630 def send_head(self):
631 """Common code for GET and HEAD commands.
632
633 This sends the response code and MIME headers.
634
635 Return value is either a file object (which has to be copied
636 to the outputfile by the caller unless the command was HEAD,
637 and must be closed by the caller under all circumstances), or
638 None, in which case the caller has nothing further to do.
639
640 """
641 path = self.translate_path(self.path)
642 f = None
643 if os.path.isdir(path):
644 if not self.path.endswith('/'):
645 # redirect browser - doing basically what apache does
646 self.send_response(301)
647 self.send_header("Location", self.path + "/")
648 self.end_headers()
649 return None
650 for index in "index.html", "index.htm":
651 index = os.path.join(path, index)
652 if os.path.exists(index):
653 path = index
654 break
655 else:
656 return self.list_directory(path)
657 ctype = self.guess_type(path)
658 try:
659 f = open(path, 'rb')
660 except IOError:
661 self.send_error(404, "File not found")
662 return None
663 self.send_response(200)
664 self.send_header("Content-type", ctype)
665 fs = os.fstat(f.fileno())
666 self.send_header("Content-Length", str(fs[6]))
667 self.send_header("Last-Modified", self.date_time_string(fs.st_mtime))
668 self.end_headers()
669 return f
670
671 def list_directory(self, path):
672 """Helper to produce a directory listing (absent index.html).
673
674 Return value is either a file object, or None (indicating an
675 error). In either case, the headers are sent, making the
676 interface the same as for send_head().
677
678 """
679 try:
680 list = os.listdir(path)
681 except os.error:
682 self.send_error(404, "No permission to list directory")
683 return None
684 list.sort(key=lambda a: a.lower())
685 r = []
686 displaypath = cgi.escape(urllib.unquote(self.path))
687 r.append('<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">')
688 r.append("<html>\n<title>Directory listing for %s</title>\n" % displaypath)
689 r.append("<body>\n<h2>Directory listing for %s</h2>\n" % displaypath)
690 r.append("<hr>\n<ul>\n")
691 for name in list:
692 fullname = os.path.join(path, name)
693 displayname = linkname = name
694 # Append / for directories or @ for symbolic links
695 if os.path.isdir(fullname):
696 displayname = name + "/"
697 linkname = name + "/"
698 if os.path.islink(fullname):
699 displayname = name + "@"
700 # Note: a link to a directory displays with @ and links with /
701 r.append('<li><a href="%s">%s</a>\n'
702 % (urllib.quote(linkname), cgi.escape(displayname)))
703 r.append("</ul>\n<hr>\n</body>\n</html>\n")
704 enc = sys.getfilesystemencoding()
705 encoded = ''.join(r).encode(enc)
706 f = io.BytesIO()
707 f.write(encoded)
708 f.seek(0)
709 self.send_response(200)
710 self.send_header("Content-type", "text/html; charset=%s" % enc)
711 self.send_header("Content-Length", str(len(encoded)))
712 self.end_headers()
713 return f
714
715 def translate_path(self, path):
716 """Translate a /-separated PATH to the local filename syntax.
717
718 Components that mean special things to the local file system
719 (e.g. drive or directory names) are ignored. (XXX They should
720 probably be diagnosed.)
721
722 """
723 # abandon query parameters
724 path = path.split('?',1)[0]
725 path = path.split('#',1)[0]
726 path = posixpath.normpath(urllib.unquote(path))
727 words = path.split('/')
728 words = filter(None, words)
729 path = os.getcwd()
730 for word in words:
731 drive, word = os.path.splitdrive(word)
732 head, word = os.path.split(word)
733 if word in (os.curdir, os.pardir): continue
734 path = os.path.join(path, word)
735 return path
736
737 def copyfile(self, source, outputfile):
738 """Copy all data between two file objects.
739
740 The SOURCE argument is a file object open for reading
741 (or anything with a read() method) and the DESTINATION
742 argument is a file object open for writing (or
743 anything with a write() method).
744
745 The only reason for overriding this would be to change
746 the block size or perhaps to replace newlines by CRLF
747 -- note however that this the default server uses this
748 to copy binary data as well.
749
750 """
751 shutil.copyfileobj(source, outputfile)
752
753 def guess_type(self, path):
754 """Guess the type of a file.
755
756 Argument is a PATH (a filename).
757
758 Return value is a string of the form type/subtype,
759 usable for a MIME Content-type header.
760
761 The default implementation looks the file's extension
762 up in the table self.extensions_map, using application/octet-stream
763 as a default; however it would be permissible (if
764 slow) to look inside the data to make a better guess.
765
766 """
767
768 base, ext = posixpath.splitext(path)
769 if ext in self.extensions_map:
770 return self.extensions_map[ext]
771 ext = ext.lower()
772 if ext in self.extensions_map:
773 return self.extensions_map[ext]
774 else:
775 return self.extensions_map['']
776
777 if not mimetypes.inited:
778 mimetypes.init() # try to read system mime.types
779 extensions_map = mimetypes.types_map.copy()
780 extensions_map.update({
781 '': 'application/octet-stream', # Default
782 '.py': 'text/plain',
783 '.c': 'text/plain',
784 '.h': 'text/plain',
785 })
786
787
788# Utilities for CGIHTTPRequestHandler
789
790nobody = None
791
792def nobody_uid():
793 """Internal routine to get nobody's uid"""
794 global nobody
795 if nobody:
796 return nobody
797 try:
798 import pwd
799 except ImportError:
800 return -1
801 try:
802 nobody = pwd.getpwnam('nobody')[2]
803 except KeyError:
804 nobody = 1 + max(map(lambda x: x[2], pwd.getpwall()))
805 return nobody
806
807
808def executable(path):
809 """Test for executable file."""
810 try:
811 st = os.stat(path)
812 except os.error:
813 return False
814 return st.st_mode & 0o111 != 0
815
816
817class CGIHTTPRequestHandler(SimpleHTTPRequestHandler):
818
819 """Complete HTTP server with GET, HEAD and POST commands.
820
821 GET and HEAD also support running CGI scripts.
822
823 The POST command is *only* implemented for CGI scripts.
824
825 """
826
827 # Determine platform specifics
828 have_fork = hasattr(os, 'fork')
829 have_popen2 = hasattr(os, 'popen2')
830 have_popen3 = hasattr(os, 'popen3')
831
832 # Make rfile unbuffered -- we need to read one line and then pass
833 # the rest to a subprocess, so we can't use buffered input.
834 rbufsize = 0
835
836 def do_POST(self):
837 """Serve a POST request.
838
839 This is only implemented for CGI scripts.
840
841 """
842
843 if self.is_cgi():
844 self.run_cgi()
845 else:
846 self.send_error(501, "Can only POST to CGI scripts")
847
848 def send_head(self):
849 """Version of send_head that support CGI scripts"""
850 if self.is_cgi():
851 return self.run_cgi()
852 else:
853 return SimpleHTTPRequestHandler.send_head(self)
854
855 def is_cgi(self):
856 """Test whether self.path corresponds to a CGI script.
857
858 Return a tuple (dir, rest) if self.path requires running a
859 CGI script, None if not. Note that rest begins with a
860 slash if it is not empty.
861
862 The default implementation tests whether the path
863 begins with one of the strings in the list
864 self.cgi_directories (and the next character is a '/'
865 or the end of the string).
866
867 """
868
869 path = self.path
870
871 for x in self.cgi_directories:
872 i = len(x)
873 if path[:i] == x and (not path[i:] or path[i] == '/'):
874 self.cgi_info = path[:i], path[i+1:]
875 return True
876 return False
877
878 cgi_directories = ['/cgi-bin', '/htbin']
879
880 def is_executable(self, path):
881 """Test whether argument path is an executable file."""
882 return executable(path)
883
884 def is_python(self, path):
885 """Test whether argument path is a Python script."""
886 head, tail = os.path.splitext(path)
887 return tail.lower() in (".py", ".pyw")
888
889 def run_cgi(self):
890 """Execute a CGI script."""
891 path = self.path
892 dir, rest = self.cgi_info
893
894 i = path.find('/', len(dir) + 1)
895 while i >= 0:
896 nextdir = path[:i]
897 nextrest = path[i+1:]
898
899 scriptdir = self.translate_path(nextdir)
900 if os.path.isdir(scriptdir):
901 dir, rest = nextdir, nextrest
902 i = path.find('/', len(dir) + 1)
903 else:
904 break
905
906 # find an explicit query string, if present.
907 i = rest.rfind('?')
908 if i >= 0:
909 rest, query = rest[:i], rest[i+1:]
910 else:
911 query = ''
912
913 # dissect the part after the directory name into a script name &
914 # a possible additional path, to be stored in PATH_INFO.
915 i = rest.find('/')
916 if i >= 0:
917 script, rest = rest[:i], rest[i:]
918 else:
919 script, rest = rest, ''
920
921 scriptname = dir + '/' + script
922 scriptfile = self.translate_path(scriptname)
923 if not os.path.exists(scriptfile):
924 self.send_error(404, "No such CGI script (%r)" % scriptname)
925 return
926 if not os.path.isfile(scriptfile):
927 self.send_error(403, "CGI script is not a plain file (%r)" %
928 scriptname)
929 return
930 ispy = self.is_python(scriptname)
931 if not ispy:
932 if not (self.have_fork or self.have_popen2 or self.have_popen3):
933 self.send_error(403, "CGI script is not a Python script (%r)" %
934 scriptname)
935 return
936 if not self.is_executable(scriptfile):
937 self.send_error(403, "CGI script is not executable (%r)" %
938 scriptname)
939 return
940
941 # Reference: http://hoohoo.ncsa.uiuc.edu/cgi/env.html
942 # XXX Much of the following could be prepared ahead of time!
943 env = {}
944 env['SERVER_SOFTWARE'] = self.version_string()
945 env['SERVER_NAME'] = self.server.server_name
946 env['GATEWAY_INTERFACE'] = 'CGI/1.1'
947 env['SERVER_PROTOCOL'] = self.protocol_version
948 env['SERVER_PORT'] = str(self.server.server_port)
949 env['REQUEST_METHOD'] = self.command
950 uqrest = urllib.unquote(rest)
951 env['PATH_INFO'] = uqrest
952 env['PATH_TRANSLATED'] = self.translate_path(uqrest)
953 env['SCRIPT_NAME'] = scriptname
954 if query:
955 env['QUERY_STRING'] = query
956 host = self.address_string()
957 if host != self.client_address[0]:
958 env['REMOTE_HOST'] = host
959 env['REMOTE_ADDR'] = self.client_address[0]
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]960 authorization = self.headers.get("authorization")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]961 if authorization:
962 authorization = authorization.split()
963 if len(authorization) == 2:
964 import base64, binascii
965 env['AUTH_TYPE'] = authorization[0]
966 if authorization[0].lower() == "basic":
967 try:
968 authorization = authorization[1].encode('ascii')
969 authorization = base64.decodestring(authorization).\
970 decode('ascii')
971 except (binascii.Error, UnicodeError):
972 pass
973 else:
974 authorization = authorization.split(':')
975 if len(authorization) == 2:
976 env['REMOTE_USER'] = authorization[0]
977 # XXX REMOTE_IDENT
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]978 if self.headers.get('content-type') is None:
979 env['CONTENT_TYPE'] = self.headers.get_content_type()
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]980 else:
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]981 env['CONTENT_TYPE'] = self.headers['content-type']
982 length = self.headers.get('content-length')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]983 if length:
984 env['CONTENT_LENGTH'] = length
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]985 referer = self.headers.get('referer')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]986 if referer:
987 env['HTTP_REFERER'] = referer
988 accept = []
989 for line in self.headers.getallmatchingheaders('accept'):
990 if line[:1] in "\t\n\r ":
991 accept.append(line.strip())
992 else:
993 accept = accept + line[7:].split(',')
994 env['HTTP_ACCEPT'] = ','.join(accept)
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]995 ua = self.headers.get('user-agent')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]996 if ua:
997 env['HTTP_USER_AGENT'] = ua
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]998 co = filter(None, self.headers.get_all('cookie', []))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]999 if co:
1000 env['HTTP_COOKIE'] = ', '.join(co)
1001 # XXX Other HTTP_* headers
1002 # Since we're setting the env in the parent, provide empty
1003 # values to override previously set values
1004 for k in ('QUERY_STRING', 'REMOTE_HOST', 'CONTENT_LENGTH',
1005 'HTTP_USER_AGENT', 'HTTP_COOKIE', 'HTTP_REFERER'):
1006 env.setdefault(k, "")
1007 os.environ.update(env)
1008
1009 self.send_response(200, "Script output follows")
1010
1011 decoded_query = query.replace('+', ' ')
1012
1013 if self.have_fork:
1014 # Unix -- fork as we should
1015 args = [script]
1016 if '=' not in decoded_query:
1017 args.append(decoded_query)
1018 nobody = nobody_uid()
1019 self.wfile.flush() # Always flush before forking
1020 pid = os.fork()
1021 if pid != 0:
1022 # Parent
1023 pid, sts = os.waitpid(pid, 0)
1024 # throw away additional data [see bug #427345]
1025 while select.select([self.rfile], [], [], 0)[0]:
1026 if not self.rfile.read(1):
1027 break
1028 if sts:
1029 self.log_error("CGI script exit status %#x", sts)
1030 return
1031 # Child
1032 try:
1033 try:
1034 os.setuid(nobody)
1035 except os.error:
1036 pass
1037 os.dup2(self.rfile.fileno(), 0)
1038 os.dup2(self.wfile.fileno(), 1)
1039 os.execve(scriptfile, args, os.environ)
1040 except:
1041 self.server.handle_error(self.request, self.client_address)
1042 os._exit(127)
1043
1044 elif self.have_popen2 or self.have_popen3:
1045 # Windows -- use popen2 or popen3 to create a subprocess
1046 import shutil
1047 if self.have_popen3:
1048 popenx = os.popen3
1049 else:
1050 popenx = os.popen2
1051 cmdline = scriptfile
1052 if self.is_python(scriptfile):
1053 interp = sys.executable
1054 if interp.lower().endswith("w.exe"):
1055 # On Windows, use python.exe, not pythonw.exe
1056 interp = interp[:-5] + interp[-4:]
1057 cmdline = "%s -u %s" % (interp, cmdline)
1058 if '=' not in query and '"' not in query:
1059 cmdline = '%s "%s"' % (cmdline, query)
1060 self.log_message("command: %s", cmdline)
1061 try:
1062 nbytes = int(length)
1063 except (TypeError, ValueError):
1064 nbytes = 0
1065 files = popenx(cmdline, 'b')
1066 fi = files[0]
1067 fo = files[1]
1068 if self.have_popen3:
1069 fe = files[2]
1070 if self.command.lower() == "post" and nbytes > 0:
1071 data = self.rfile.read(nbytes)
1072 fi.write(data)
1073 # throw away additional data [see bug #427345]
1074 while select.select([self.rfile._sock], [], [], 0)[0]:
1075 if not self.rfile._sock.recv(1):
1076 break
1077 fi.close()
1078 shutil.copyfileobj(fo, self.wfile)
1079 if self.have_popen3:
1080 errors = fe.read()
1081 fe.close()
1082 if errors:
1083 self.log_error('%s', errors)
1084 sts = fo.close()
1085 if sts:
1086 self.log_error("CGI script exit status %#x", sts)
1087 else:
1088 self.log_message("CGI script exited OK")
1089
1090 else:
1091 # Other O.S. -- execute script in this process
1092 save_argv = sys.argv
1093 save_stdin = sys.stdin
1094 save_stdout = sys.stdout
1095 save_stderr = sys.stderr
1096 try:
1097 save_cwd = os.getcwd()
1098 try:
1099 sys.argv = [scriptfile]
1100 if '=' not in decoded_query:
1101 sys.argv.append(decoded_query)
1102 sys.stdout = self.wfile
1103 sys.stdin = self.rfile
1104 exec(open(scriptfile).read(), {"__name__": "__main__"})
1105 finally:
1106 sys.argv = save_argv
1107 sys.stdin = save_stdin
1108 sys.stdout = save_stdout
1109 sys.stderr = save_stderr
1110 os.chdir(save_cwd)
1111 except SystemExit as sts:
1112 self.log_error("CGI script exit status %s", str(sts))
1113 else:
1114 self.log_message("CGI script exited OK")
1115
1116
1117def test(HandlerClass = BaseHTTPRequestHandler,
1118 ServerClass = HTTPServer, protocol="HTTP/1.0"):
1119 """Test the HTTP request handler class.
1120
1121 This runs an HTTP server on port 8000 (or the first command line
1122 argument).
1123
1124 """
1125
1126 if sys.argv[1:]:
1127 port = int(sys.argv[1])
1128 else:
1129 port = 8000
1130 server_address = ('', port)
1131
1132 HandlerClass.protocol_version = protocol
1133 httpd = ServerClass(server_address, HandlerClass)
1134
1135 sa = httpd.socket.getsockname()
1136 print("Serving HTTP on", sa[0], "port", sa[1], "...")
1137 httpd.serve_forever()
1138
1139
1140if __name__ == '__main__':
1141 test(HandlerClass=BaseHTTPRequestHandler)
1142 test(HandlerClass=SimpleHTTPRequestHandler)
1143 test(HandlerClass=CGIHTTPRequestHandler)