Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / python / cpython3 / cb09dcacab59fd1b19346e73ee85318b18c16519 / . / Lib / http / server.py
blob: 2a9502870580f3247064ff55864b505b862ad73f [file] [log] [blame]
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1"""HTTP server classes.
2
3Note: BaseHTTPRequestHandler doesn't implement any HTTP request; see
4SimpleHTTPRequestHandler for simple implementations of GET, HEAD and POST,
5and CGIHTTPRequestHandler for CGI scripts.
6
7It does, however, optionally implement HTTP/1.1 persistent connections,
8as of version 0.3.
9
10Notes on CGIHTTPRequestHandler
11------------------------------
12
13This class implements GET and POST requests to cgi-bin scripts.
14
15If the os.fork() function is not present (e.g. on Windows),
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]16subprocess.Popen() is used as a fallback, with slightly altered semantics.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]17
18In all cases, the implementation is intentionally naive -- all
19requests are executed synchronously.
20
21SECURITY WARNING: DON'T USE THIS CODE UNLESS YOU ARE INSIDE A FIREWALL
22-- it may execute arbitrary Python code or external programs.
23
24Note that status code 200 is sent prior to execution of a CGI script, so
25scripts cannot send other status codes such as 302 (redirect).
26
27XXX To do:
28
29- log requests even later (to capture byte count)
30- log user-agent header and other interesting goodies
31- send error log to separate file
32"""
33
34
35# See also:
36#
37# HTTP Working Group T. Berners-Lee
38# INTERNET-DRAFT R. T. Fielding
39# <draft-ietf-http-v10-spec-00.txt> H. Frystyk Nielsen
40# Expires September 8, 1995 March 8, 1995
41#
42# URL: http://www.ics.uci.edu/pub/ietf/http/draft-ietf-http-v10-spec-00.txt
43#
44# and
45#
46# Network Working Group R. Fielding
47# Request for Comments: 2616 et al
48# Obsoletes: 2068 June 1999
49# Category: Standards Track
50#
51# URL: http://www.faqs.org/rfcs/rfc2616.html
52
53# Log files
54# ---------
55#
56# Here's a quote from the NCSA httpd docs about log file format.
57#
58# | The logfile format is as follows. Each line consists of:
59# |
60# | host rfc931 authuser [DD/Mon/YYYY:hh:mm:ss] "request" ddd bbbb
61# |
62# | host: Either the DNS name or the IP number of the remote client
63# | rfc931: Any information returned by identd for this person,
64# | - otherwise.
65# | authuser: If user sent a userid for authentication, the user name,
66# | - otherwise.
67# | DD: Day
68# | Mon: Month (calendar name)
69# | YYYY: Year
70# | hh: hour (24-hour format, the machine's timezone)
71# | mm: minutes
72# | ss: seconds
73# | request: The first line of the HTTP request as sent by the client.
74# | ddd: the status code returned by the server, - if not available.
75# | bbbb: the total number of bytes sent,
76# | *not including the HTTP/1.0 header*, - if not available
77# |
78# | You can determine the name of the file accessed through request.
79#
80# (Actually, the latter is only true if you know the server configuration
81# at the time the request was made!)
82
83__version__ = "0.6"
84
85__all__ = ["HTTPServer", "BaseHTTPRequestHandler"]
86
Georg Brandl1f7fffb2010-10-15 15:57:45 +0000[diff] [blame]87import html
Jeremy Hylton914ab452009-03-27 17:16:06 +0000[diff] [blame]88import http.client
89import io
90import mimetypes
91import os
92import posixpath
93import select
94import shutil
95import socket # For gethostbyaddr()
96import socketserver
97import sys
98import time
99import urllib.parse
Senthil Kumaran42713722010-10-03 17:55:45 +0000[diff] [blame]100import copy
Senthil Kumaran1251faf2012-06-03 16:15:54 +0800[diff] [blame]101import argparse
102
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]103
104# Default error message template
105DEFAULT_ERROR_MESSAGE = """\
Senthil Kumaran1b407fe2011-03-20 10:44:30 +0800[diff] [blame]106<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
107 "http://www.w3.org/TR/html4/strict.dtd">
Ezio Melottica897e92011-11-02 19:33:29 +0200[diff] [blame]108<html>
Senthil Kumaranb253c9f2011-03-17 16:43:22 +0800[diff] [blame]109 <head>
Senthil Kumaran1b407fe2011-03-20 10:44:30 +0800[diff] [blame]110 <meta http-equiv="Content-Type" content="text/html;charset=utf-8">
Senthil Kumaranb253c9f2011-03-17 16:43:22 +0800[diff] [blame]111 <title>Error response</title>
112 </head>
113 <body>
114 <h1>Error response</h1>
115 <p>Error code: %(code)d</p>
116 <p>Message: %(message)s.</p>
117 <p>Error code explanation: %(code)s - %(explain)s.</p>
118 </body>
119</html>
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]120"""
121
122DEFAULT_ERROR_CONTENT_TYPE = "text/html;charset=utf-8"
123
124def _quote_html(html):
125 return html.replace("&", "&amp;").replace("<", "&lt;").replace(">", "&gt;")
126
127class HTTPServer(socketserver.TCPServer):
128
129 allow_reuse_address = 1 # Seems to make sense in testing environment
130
131 def server_bind(self):
132 """Override server_bind to store the server name."""
133 socketserver.TCPServer.server_bind(self)
134 host, port = self.socket.getsockname()[:2]
135 self.server_name = socket.getfqdn(host)
136 self.server_port = port
137
138
139class BaseHTTPRequestHandler(socketserver.StreamRequestHandler):
140
141 """HTTP request handler base class.
142
143 The following explanation of HTTP serves to guide you through the
144 code as well as to expose any misunderstandings I may have about
145 HTTP (so you don't need to read the code to figure out I'm wrong
146 :-).
147
148 HTTP (HyperText Transfer Protocol) is an extensible protocol on
149 top of a reliable stream transport (e.g. TCP/IP). The protocol
150 recognizes three parts to a request:
151
152 1. One line identifying the request type and path
153 2. An optional set of RFC-822-style headers
154 3. An optional data part
155
156 The headers and data are separated by a blank line.
157
158 The first line of the request has the form
159
160 <command> <path> <version>
161
162 where <command> is a (case-sensitive) keyword such as GET or POST,
163 <path> is a string containing path information for the request,
164 and <version> should be the string "HTTP/1.0" or "HTTP/1.1".
165 <path> is encoded using the URL encoding scheme (using %xx to signify
166 the ASCII character with hex code xx).
167
168 The specification specifies that lines are separated by CRLF but
169 for compatibility with the widest range of clients recommends
170 servers also handle LF. Similarly, whitespace in the request line
171 is treated sensibly (allowing multiple spaces between components
172 and allowing trailing whitespace).
173
174 Similarly, for output, lines ought to be separated by CRLF pairs
175 but most clients grok LF characters just fine.
176
177 If the first line of the request has the form
178
179 <command> <path>
180
181 (i.e. <version> is left out) then this is assumed to be an HTTP
182 0.9 request; this form has no optional headers and data part and
183 the reply consists of just the data.
184
185 The reply form of the HTTP 1.x protocol again has three parts:
186
187 1. One line giving the response code
188 2. An optional set of RFC-822-style headers
189 3. The data
190
191 Again, the headers and data are separated by a blank line.
192
193 The response code line has the form
194
195 <version> <responsecode> <responsestring>
196
197 where <version> is the protocol version ("HTTP/1.0" or "HTTP/1.1"),
198 <responsecode> is a 3-digit response code indicating success or
199 failure of the request, and <responsestring> is an optional
200 human-readable string explaining what the response code means.
201
202 This server parses the request and the headers, and then calls a
203 function specific to the request type (<command>). Specifically,
204 a request SPAM will be handled by a method do_SPAM(). If no
205 such method exists the server sends an error response to the
206 client. If it exists, it is called with no arguments:
207
208 do_SPAM()
209
210 Note that the request name is case sensitive (i.e. SPAM and spam
211 are different requests).
212
213 The various request details are stored in instance variables:
214
215 - client_address is the client IP address in the form (host,
216 port);
217
218 - command, path and version are the broken-down request line;
219
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]220 - headers is an instance of email.message.Message (or a derived
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]221 class) containing the header information;
222
223 - rfile is a file object open for reading positioned at the
224 start of the optional input data part;
225
226 - wfile is a file object open for writing.
227
228 IT IS IMPORTANT TO ADHERE TO THE PROTOCOL FOR WRITING!
229
230 The first thing to be written must be the response line. Then
231 follow 0 or more header lines, then a blank line, and then the
232 actual data (if any). The meaning of the header lines depends on
233 the command executed by the server; in most cases, when data is
234 returned, there should be at least one header line of the form
235
236 Content-type: <type>/<subtype>
237
238 where <type> and <subtype> should be registered MIME types,
239 e.g. "text/html" or "text/plain".
240
241 """
242
243 # The Python system version, truncated to its first component.
244 sys_version = "Python/" + sys.version.split()[0]
245
246 # The server software version. You may want to override this.
247 # The format is multiple whitespace-separated strings,
248 # where each string is of the form name[/version].
249 server_version = "BaseHTTP/" + __version__
250
251 error_message_format = DEFAULT_ERROR_MESSAGE
252 error_content_type = DEFAULT_ERROR_CONTENT_TYPE
253
254 # The default request version. This only affects responses up until
255 # the point where the request line is parsed, so it mainly decides what
256 # the client gets back when sending a malformed request line.
257 # Most web servers default to HTTP 0.9, i.e. don't send a status line.
258 default_request_version = "HTTP/0.9"
259
260 def parse_request(self):
261 """Parse a request (internal).
262
263 The request should be stored in self.raw_requestline; the results
264 are in self.command, self.path, self.request_version and
265 self.headers.
266
267 Return True for success, False for failure; on failure, an
268 error is sent back.
269
270 """
271 self.command = None # set in case of error on the first line
272 self.request_version = version = self.default_request_version
273 self.close_connection = 1
274 requestline = str(self.raw_requestline, 'iso-8859-1')
Senthil Kumaran30755492011-12-23 17:03:41 +0800[diff] [blame]275 requestline = requestline.rstrip('\r\n')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]276 self.requestline = requestline
277 words = requestline.split()
278 if len(words) == 3:
Senthil Kumaran30755492011-12-23 17:03:41 +0800[diff] [blame]279 command, path, version = words
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]280 if version[:5] != 'HTTP/':
281 self.send_error(400, "Bad request version (%r)" % version)
282 return False
283 try:
284 base_version_number = version.split('/', 1)[1]
285 version_number = base_version_number.split(".")
286 # RFC 2145 section 3.1 says there can be only one "." and
287 # - major and minor numbers MUST be treated as
288 # separate integers;
289 # - HTTP/2.4 is a lower version than HTTP/2.13, which in
290 # turn is lower than HTTP/12.3;
291 # - Leading zeros MUST be ignored by recipients.
292 if len(version_number) != 2:
293 raise ValueError
294 version_number = int(version_number[0]), int(version_number[1])
295 except (ValueError, IndexError):
296 self.send_error(400, "Bad request version (%r)" % version)
297 return False
298 if version_number >= (1, 1) and self.protocol_version >= "HTTP/1.1":
299 self.close_connection = 0
300 if version_number >= (2, 0):
301 self.send_error(505,
302 "Invalid HTTP Version (%s)" % base_version_number)
303 return False
304 elif len(words) == 2:
Senthil Kumaran30755492011-12-23 17:03:41 +0800[diff] [blame]305 command, path = words
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]306 self.close_connection = 1
307 if command != 'GET':
308 self.send_error(400,
309 "Bad HTTP/0.9 request type (%r)" % command)
310 return False
311 elif not words:
312 return False
313 else:
314 self.send_error(400, "Bad request syntax (%r)" % requestline)
315 return False
316 self.command, self.path, self.request_version = command, path, version
317
318 # Examine the headers and look for a Connection directive.
Senthil Kumaran5466bf12010-12-18 16:55:23 +0000[diff] [blame]319 try:
320 self.headers = http.client.parse_headers(self.rfile,
321 _class=self.MessageClass)
322 except http.client.LineTooLong:
323 self.send_error(400, "Line too long")
324 return False
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]325
326 conntype = self.headers.get('Connection', "")
327 if conntype.lower() == 'close':
328 self.close_connection = 1
329 elif (conntype.lower() == 'keep-alive' and
330 self.protocol_version >= "HTTP/1.1"):
331 self.close_connection = 0
Senthil Kumaran0f476d42010-09-30 06:09:18 +0000[diff] [blame]332 # Examine the headers and look for an Expect directive
333 expect = self.headers.get('Expect', "")
334 if (expect.lower() == "100-continue" and
335 self.protocol_version >= "HTTP/1.1" and
336 self.request_version >= "HTTP/1.1"):
337 if not self.handle_expect_100():
338 return False
339 return True
340
341 def handle_expect_100(self):
342 """Decide what to do with an "Expect: 100-continue" header.
343
344 If the client is expecting a 100 Continue response, we must
345 respond with either a 100 Continue or a final response before
346 waiting for the request body. The default is to always respond
347 with a 100 Continue. You can behave differently (for example,
348 reject unauthorized requests) by overriding this method.
349
350 This method should either return True (possibly after sending
351 a 100 Continue response) or send an error response and return
352 False.
353
354 """
355 self.send_response_only(100)
Benjamin Peterson04424232014-01-18 21:50:18 -0500[diff] [blame]356 self.end_headers()
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]357 return True
358
359 def handle_one_request(self):
360 """Handle a single HTTP request.
361
362 You normally don't need to override this method; see the class
363 __doc__ string for information on how to handle specific HTTP
364 commands such as GET and POST.
365
366 """
Kristján Valur Jónsson985fc6a2009-07-01 10:01:31 +0000[diff] [blame]367 try:
Antoine Pitrouc4924372010-12-16 16:48:36 +0000[diff] [blame]368 self.raw_requestline = self.rfile.readline(65537)
369 if len(self.raw_requestline) > 65536:
370 self.requestline = ''
371 self.request_version = ''
372 self.command = ''
373 self.send_error(414)
374 return
Kristján Valur Jónsson985fc6a2009-07-01 10:01:31 +0000[diff] [blame]375 if not self.raw_requestline:
376 self.close_connection = 1
377 return
378 if not self.parse_request():
379 # An error code has been sent, just exit
380 return
381 mname = 'do_' + self.command
382 if not hasattr(self, mname):
383 self.send_error(501, "Unsupported method (%r)" % self.command)
384 return
385 method = getattr(self, mname)
386 method()
387 self.wfile.flush() #actually send the response if not already done.
388 except socket.timeout as e:
389 #a read or a write timed out. Discard this connection
390 self.log_error("Request timed out: %r", e)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]391 self.close_connection = 1
392 return
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]393
394 def handle(self):
395 """Handle multiple requests if necessary."""
396 self.close_connection = 1
397
398 self.handle_one_request()
399 while not self.close_connection:
400 self.handle_one_request()
401
Senthil Kumaran26886442013-03-15 07:53:21 -0700[diff] [blame]402 def send_error(self, code, message=None, explain=None):
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]403 """Send and log an error reply.
404
Senthil Kumaran26886442013-03-15 07:53:21 -0700[diff] [blame]405 Arguments are
406 * code: an HTTP error code
407 3 digits
408 * message: a simple optional 1 line reason phrase.
409 *( HTAB / SP / VCHAR / %x80-FF )
410 defaults to short entry matching the response code
411 * explain: a detailed message defaults to the long entry
412 matching the response code.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]413
414 This sends an error response (so it must be called before any
415 output has been generated), logs the error, and finally sends
416 a piece of HTML explaining the error to the user.
417
418 """
419
420 try:
421 shortmsg, longmsg = self.responses[code]
422 except KeyError:
423 shortmsg, longmsg = '???', '???'
424 if message is None:
425 message = shortmsg
Senthil Kumaran26886442013-03-15 07:53:21 -0700[diff] [blame]426 if explain is None:
427 explain = longmsg
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]428 self.log_error("code %d, message %s", code, message)
429 # using _quote_html to prevent Cross Site Scripting attacks (see bug #1100201)
430 content = (self.error_message_format %
Senthil Kumaran26886442013-03-15 07:53:21 -0700[diff] [blame]431 {'code': code, 'message': _quote_html(message), 'explain': _quote_html(explain)})
Senthil Kumaran52d27202012-10-10 23:16:21 -0700[diff] [blame]432 body = content.encode('UTF-8', 'replace')
Senthil Kumaran1e7551d2013-03-05 02:25:58 -0800[diff] [blame]433 self.send_response(code, message)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]434 self.send_header("Content-Type", self.error_content_type)
435 self.send_header('Connection', 'close')
Senthil Kumaran52d27202012-10-10 23:16:21 -0700[diff] [blame]436 self.send_header('Content-Length', int(len(body)))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]437 self.end_headers()
438 if self.command != 'HEAD' and code >= 200 and code not in (204, 304):
Senthil Kumaran52d27202012-10-10 23:16:21 -0700[diff] [blame]439 self.wfile.write(body)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]440
441 def send_response(self, code, message=None):
Senthil Kumaranc7ae19b2011-05-09 23:25:02 +0800[diff] [blame]442 """Add the response header to the headers buffer and log the
443 response code.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]444
445 Also send two standard headers with the server software
446 version and the current date.
447
448 """
449 self.log_request(code)
Senthil Kumaran0f476d42010-09-30 06:09:18 +0000[diff] [blame]450 self.send_response_only(code, message)
451 self.send_header('Server', self.version_string())
452 self.send_header('Date', self.date_time_string())
453
454 def send_response_only(self, code, message=None):
455 """Send the response header only."""
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]456 if message is None:
457 if code in self.responses:
458 message = self.responses[code][0]
459 else:
460 message = ''
461 if self.request_version != 'HTTP/0.9':
Senthil Kumaranc7ae19b2011-05-09 23:25:02 +0800[diff] [blame]462 if not hasattr(self, '_headers_buffer'):
463 self._headers_buffer = []
464 self._headers_buffer.append(("%s %d %s\r\n" %
465 (self.protocol_version, code, message)).encode(
466 'latin-1', 'strict'))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]467
468 def send_header(self, keyword, value):
Senthil Kumaranc7ae19b2011-05-09 23:25:02 +0800[diff] [blame]469 """Send a MIME header to the headers buffer."""
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]470 if self.request_version != 'HTTP/0.9':
Senthil Kumarane4dad4f2010-11-21 14:36:14 +0000[diff] [blame]471 if not hasattr(self, '_headers_buffer'):
472 self._headers_buffer = []
473 self._headers_buffer.append(
Marc-André Lemburg8f36af72011-02-25 15:42:01 +0000[diff] [blame]474 ("%s: %s\r\n" % (keyword, value)).encode('latin-1', 'strict'))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]475
476 if keyword.lower() == 'connection':
477 if value.lower() == 'close':
478 self.close_connection = 1
479 elif value.lower() == 'keep-alive':
480 self.close_connection = 0
481
482 def end_headers(self):
483 """Send the blank line ending the MIME headers."""
484 if self.request_version != 'HTTP/0.9':
Senthil Kumarane4dad4f2010-11-21 14:36:14 +0000[diff] [blame]485 self._headers_buffer.append(b"\r\n")
Senthil Kumaranc7ae19b2011-05-09 23:25:02 +0800[diff] [blame]486 self.flush_headers()
487
488 def flush_headers(self):
489 if hasattr(self, '_headers_buffer'):
Senthil Kumarane4dad4f2010-11-21 14:36:14 +0000[diff] [blame]490 self.wfile.write(b"".join(self._headers_buffer))
491 self._headers_buffer = []
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]492
493 def log_request(self, code='-', size='-'):
494 """Log an accepted request.
495
496 This is called by send_response().
497
498 """
499
500 self.log_message('"%s" %s %s',
501 self.requestline, str(code), str(size))
502
503 def log_error(self, format, *args):
504 """Log an error.
505
506 This is called when a request cannot be fulfilled. By
507 default it passes the message on to log_message().
508
509 Arguments are the same as for log_message().
510
511 XXX This should go to the separate error log.
512
513 """
514
515 self.log_message(format, *args)
516
517 def log_message(self, format, *args):
518 """Log an arbitrary message.
519
520 This is used by all other logging functions. Override
521 it if you have specific logging wishes.
522
523 The first argument, FORMAT, is a format string for the
524 message to be logged. If the format string contains
525 any % escapes requiring parameters, they should be
526 specified as subsequent arguments (it's just like
527 printf!).
528
Senthil Kumarandb727b42012-04-29 13:41:03 +0800[diff] [blame]529 The client ip and current date/time are prefixed to
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]530 every message.
531
532 """
533
534 sys.stderr.write("%s - - [%s] %s\n" %
535 (self.address_string(),
536 self.log_date_time_string(),
537 format%args))
538
539 def version_string(self):
540 """Return the server software version string."""
541 return self.server_version + ' ' + self.sys_version
542
543 def date_time_string(self, timestamp=None):
544 """Return the current date and time formatted for a message header."""
545 if timestamp is None:
546 timestamp = time.time()
547 year, month, day, hh, mm, ss, wd, y, z = time.gmtime(timestamp)
548 s = "%s, %02d %3s %4d %02d:%02d:%02d GMT" % (
549 self.weekdayname[wd],
550 day, self.monthname[month], year,
551 hh, mm, ss)
552 return s
553
554 def log_date_time_string(self):
555 """Return the current time formatted for logging."""
556 now = time.time()
557 year, month, day, hh, mm, ss, x, y, z = time.localtime(now)
558 s = "%02d/%3s/%04d %02d:%02d:%02d" % (
559 day, self.monthname[month], year, hh, mm, ss)
560 return s
561
562 weekdayname = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
563
564 monthname = [None,
565 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun',
566 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
567
568 def address_string(self):
Senthil Kumaran1aacba42012-04-29 12:51:54 +0800[diff] [blame]569 """Return the client address."""
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]570
Senthil Kumaran1aacba42012-04-29 12:51:54 +0800[diff] [blame]571 return self.client_address[0]
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]572
573 # Essentially static class variables
574
575 # The version of the HTTP protocol we support.
576 # Set this to HTTP/1.1 to enable automatic keepalive
577 protocol_version = "HTTP/1.0"
578
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]579 # MessageClass used to parse headers
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]580 MessageClass = http.client.HTTPMessage
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]581
582 # Table mapping response codes to messages; entries have the
583 # form {code: (shortmessage, longmessage)}.
Hynek Schlawack51b2ed52012-05-16 09:51:07 +0200[diff] [blame]584 # See RFC 2616 and 6585.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]585 responses = {
586 100: ('Continue', 'Request received, please continue'),
587 101: ('Switching Protocols',
588 'Switching to new protocol; obey Upgrade header'),
589
590 200: ('OK', 'Request fulfilled, document follows'),
591 201: ('Created', 'Document created, URL follows'),
592 202: ('Accepted',
593 'Request accepted, processing continues off-line'),
594 203: ('Non-Authoritative Information', 'Request fulfilled from cache'),
595 204: ('No Content', 'Request fulfilled, nothing follows'),
596 205: ('Reset Content', 'Clear input form for further input.'),
597 206: ('Partial Content', 'Partial content follows.'),
598
599 300: ('Multiple Choices',
600 'Object has several resources -- see URI list'),
601 301: ('Moved Permanently', 'Object moved permanently -- see URI list'),
602 302: ('Found', 'Object moved temporarily -- see URI list'),
603 303: ('See Other', 'Object moved -- see Method and URL list'),
604 304: ('Not Modified',
605 'Document has not changed since given time'),
606 305: ('Use Proxy',
607 'You must use proxy specified in Location to access this '
608 'resource.'),
609 307: ('Temporary Redirect',
610 'Object moved temporarily -- see URI list'),
611
612 400: ('Bad Request',
613 'Bad request syntax or unsupported method'),
614 401: ('Unauthorized',
615 'No permission -- see authorization schemes'),
616 402: ('Payment Required',
617 'No payment -- see charging schemes'),
618 403: ('Forbidden',
619 'Request forbidden -- authorization will not help'),
620 404: ('Not Found', 'Nothing matches the given URI'),
621 405: ('Method Not Allowed',
Senthil Kumaran7aa26212010-02-22 11:00:50 +0000[diff] [blame]622 'Specified method is invalid for this resource.'),
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]623 406: ('Not Acceptable', 'URI not available in preferred format.'),
624 407: ('Proxy Authentication Required', 'You must authenticate with '
625 'this proxy before proceeding.'),
626 408: ('Request Timeout', 'Request timed out; try again later.'),
627 409: ('Conflict', 'Request conflict.'),
628 410: ('Gone',
629 'URI no longer exists and has been permanently removed.'),
630 411: ('Length Required', 'Client must specify Content-Length.'),
631 412: ('Precondition Failed', 'Precondition in headers is false.'),
632 413: ('Request Entity Too Large', 'Entity is too large.'),
633 414: ('Request-URI Too Long', 'URI is too long.'),
634 415: ('Unsupported Media Type', 'Entity body in unsupported format.'),
635 416: ('Requested Range Not Satisfiable',
636 'Cannot satisfy request range.'),
637 417: ('Expectation Failed',
638 'Expect condition could not be satisfied.'),
Hynek Schlawack51b2ed52012-05-16 09:51:07 +0200[diff] [blame]639 428: ('Precondition Required',
640 'The origin server requires the request to be conditional.'),
641 429: ('Too Many Requests', 'The user has sent too many requests '
642 'in a given amount of time ("rate limiting").'),
643 431: ('Request Header Fields Too Large', 'The server is unwilling to '
644 'process the request because its header fields are too large.'),
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]645
646 500: ('Internal Server Error', 'Server got itself in trouble'),
647 501: ('Not Implemented',
648 'Server does not support this operation'),
649 502: ('Bad Gateway', 'Invalid responses from another server/proxy.'),
650 503: ('Service Unavailable',
651 'The server cannot process the request due to a high load'),
652 504: ('Gateway Timeout',
653 'The gateway server did not receive a timely response'),
654 505: ('HTTP Version Not Supported', 'Cannot fulfill request.'),
Hynek Schlawack51b2ed52012-05-16 09:51:07 +0200[diff] [blame]655 511: ('Network Authentication Required',
656 'The client needs to authenticate to gain network access.'),
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]657 }
658
659
660class SimpleHTTPRequestHandler(BaseHTTPRequestHandler):
661
662 """Simple HTTP request handler with GET and HEAD commands.
663
664 This serves files from the current directory and any of its
665 subdirectories. The MIME type for files is determined by
666 calling the .guess_type() method.
667
668 The GET and HEAD requests are identical except that the HEAD
669 request omits the actual contents of the file.
670
671 """
672
673 server_version = "SimpleHTTP/" + __version__
674
675 def do_GET(self):
676 """Serve a GET request."""
677 f = self.send_head()
678 if f:
Serhiy Storchaka91b0bc22014-01-25 19:43:02 +0200[diff] [blame]679 try:
680 self.copyfile(f, self.wfile)
681 finally:
682 f.close()
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]683
684 def do_HEAD(self):
685 """Serve a HEAD request."""
686 f = self.send_head()
687 if f:
688 f.close()
689
690 def send_head(self):
691 """Common code for GET and HEAD commands.
692
693 This sends the response code and MIME headers.
694
695 Return value is either a file object (which has to be copied
696 to the outputfile by the caller unless the command was HEAD,
697 and must be closed by the caller under all circumstances), or
698 None, in which case the caller has nothing further to do.
699
700 """
701 path = self.translate_path(self.path)
702 f = None
703 if os.path.isdir(path):
704 if not self.path.endswith('/'):
705 # redirect browser - doing basically what apache does
706 self.send_response(301)
707 self.send_header("Location", self.path + "/")
708 self.end_headers()
709 return None
710 for index in "index.html", "index.htm":
711 index = os.path.join(path, index)
712 if os.path.exists(index):
713 path = index
714 break
715 else:
716 return self.list_directory(path)
717 ctype = self.guess_type(path)
718 try:
719 f = open(path, 'rb')
Andrew Svetlovf7a17b42012-12-25 16:47:37 +0200[diff] [blame]720 except OSError:
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]721 self.send_error(404, "File not found")
722 return None
Serhiy Storchaka91b0bc22014-01-25 19:43:02 +0200[diff] [blame]723 try:
724 self.send_response(200)
725 self.send_header("Content-type", ctype)
726 fs = os.fstat(f.fileno())
727 self.send_header("Content-Length", str(fs[6]))
728 self.send_header("Last-Modified", self.date_time_string(fs.st_mtime))
729 self.end_headers()
730 return f
731 except:
732 f.close()
733 raise
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]734
735 def list_directory(self, path):
736 """Helper to produce a directory listing (absent index.html).
737
738 Return value is either a file object, or None (indicating an
739 error). In either case, the headers are sent, making the
740 interface the same as for send_head().
741
742 """
743 try:
744 list = os.listdir(path)
Andrew Svetlovad28c7f2012-12-18 22:02:39 +0200[diff] [blame]745 except OSError:
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]746 self.send_error(404, "No permission to list directory")
747 return None
748 list.sort(key=lambda a: a.lower())
749 r = []
Georg Brandl1f7fffb2010-10-15 15:57:45 +0000[diff] [blame]750 displaypath = html.escape(urllib.parse.unquote(self.path))
Ezio Melottica897e92011-11-02 19:33:29 +0200[diff] [blame]751 enc = sys.getfilesystemencoding()
752 title = 'Directory listing for %s' % displaypath
753 r.append('<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" '
754 '"http://www.w3.org/TR/html4/strict.dtd">')
755 r.append('<html>\n<head>')
756 r.append('<meta http-equiv="Content-Type" '
757 'content="text/html; charset=%s">' % enc)
758 r.append('<title>%s</title>\n</head>' % title)
759 r.append('<body>\n<h1>%s</h1>' % title)
760 r.append('<hr>\n<ul>')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]761 for name in list:
762 fullname = os.path.join(path, name)
763 displayname = linkname = name
764 # Append / for directories or @ for symbolic links
765 if os.path.isdir(fullname):
766 displayname = name + "/"
767 linkname = name + "/"
768 if os.path.islink(fullname):
769 displayname = name + "@"
770 # Note: a link to a directory displays with @ and links with /
Ezio Melottica897e92011-11-02 19:33:29 +0200[diff] [blame]771 r.append('<li><a href="%s">%s</a></li>'
Georg Brandl1f7fffb2010-10-15 15:57:45 +0000[diff] [blame]772 % (urllib.parse.quote(linkname), html.escape(displayname)))
Ezio Melottica897e92011-11-02 19:33:29 +0200[diff] [blame]773 r.append('</ul>\n<hr>\n</body>\n</html>\n')
774 encoded = '\n'.join(r).encode(enc)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]775 f = io.BytesIO()
776 f.write(encoded)
777 f.seek(0)
778 self.send_response(200)
779 self.send_header("Content-type", "text/html; charset=%s" % enc)
780 self.send_header("Content-Length", str(len(encoded)))
781 self.end_headers()
782 return f
783
784 def translate_path(self, path):
785 """Translate a /-separated PATH to the local filename syntax.
786
787 Components that mean special things to the local file system
788 (e.g. drive or directory names) are ignored. (XXX They should
789 probably be diagnosed.)
790
791 """
792 # abandon query parameters
793 path = path.split('?',1)[0]
794 path = path.split('#',1)[0]
Senthil Kumaran72c238e2013-09-13 00:21:18 -0700[diff] [blame]795 # Don't forget explicit trailing slash when normalizing. Issue17324
Senthil Kumaran600b7352013-09-29 18:59:04 -0700[diff] [blame]796 trailing_slash = path.rstrip().endswith('/')
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]797 path = posixpath.normpath(urllib.parse.unquote(path))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]798 words = path.split('/')
799 words = filter(None, words)
800 path = os.getcwd()
801 for word in words:
802 drive, word = os.path.splitdrive(word)
803 head, word = os.path.split(word)
804 if word in (os.curdir, os.pardir): continue
805 path = os.path.join(path, word)
Senthil Kumaran72c238e2013-09-13 00:21:18 -0700[diff] [blame]806 if trailing_slash:
807 path += '/'
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]808 return path
809
810 def copyfile(self, source, outputfile):
811 """Copy all data between two file objects.
812
813 The SOURCE argument is a file object open for reading
814 (or anything with a read() method) and the DESTINATION
815 argument is a file object open for writing (or
816 anything with a write() method).
817
818 The only reason for overriding this would be to change
819 the block size or perhaps to replace newlines by CRLF
820 -- note however that this the default server uses this
821 to copy binary data as well.
822
823 """
824 shutil.copyfileobj(source, outputfile)
825
826 def guess_type(self, path):
827 """Guess the type of a file.
828
829 Argument is a PATH (a filename).
830
831 Return value is a string of the form type/subtype,
832 usable for a MIME Content-type header.
833
834 The default implementation looks the file's extension
835 up in the table self.extensions_map, using application/octet-stream
836 as a default; however it would be permissible (if
837 slow) to look inside the data to make a better guess.
838
839 """
840
841 base, ext = posixpath.splitext(path)
842 if ext in self.extensions_map:
843 return self.extensions_map[ext]
844 ext = ext.lower()
845 if ext in self.extensions_map:
846 return self.extensions_map[ext]
847 else:
848 return self.extensions_map['']
849
850 if not mimetypes.inited:
851 mimetypes.init() # try to read system mime.types
852 extensions_map = mimetypes.types_map.copy()
853 extensions_map.update({
854 '': 'application/octet-stream', # Default
855 '.py': 'text/plain',
856 '.c': 'text/plain',
857 '.h': 'text/plain',
858 })
859
860
861# Utilities for CGIHTTPRequestHandler
862
Senthil Kumarand70846b2012-04-12 02:34:32 +0800[diff] [blame]863def _url_collapse_path(path):
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]864 """
865 Given a URL path, remove extra '/'s and '.' path elements and collapse
Senthil Kumarand70846b2012-04-12 02:34:32 +0800[diff] [blame]866 any '..' references and returns a colllapsed path.
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]867
868 Implements something akin to RFC-2396 5.2 step 6 to parse relative paths.
Senthil Kumarand70846b2012-04-12 02:34:32 +0800[diff] [blame]869 The utility of this function is limited to is_cgi method and helps
870 preventing some security attacks.
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]871
872 Returns: A tuple of (head, tail) where tail is everything after the final /
873 and head is everything before it. Head will always start with a '/' and,
874 if it contains anything else, never have a trailing '/'.
875
876 Raises: IndexError if too many '..' occur within the path.
Senthil Kumarand70846b2012-04-12 02:34:32 +0800[diff] [blame]877
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]878 """
879 # Similar to os.path.split(os.path.normpath(path)) but specific to URL
880 # path semantics rather than local operating system semantics.
Senthil Kumarand70846b2012-04-12 02:34:32 +0800[diff] [blame]881 path_parts = path.split('/')
882 head_parts = []
883 for part in path_parts[:-1]:
884 if part == '..':
885 head_parts.pop() # IndexError if more '..' than prior parts
886 elif part and part != '.':
887 head_parts.append( part )
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]888 if path_parts:
Senthil Kumarandbb369d2012-04-11 03:15:28 +0800[diff] [blame]889 tail_part = path_parts.pop()
Senthil Kumarand70846b2012-04-12 02:34:32 +0800[diff] [blame]890 if tail_part:
891 if tail_part == '..':
892 head_parts.pop()
893 tail_part = ''
894 elif tail_part == '.':
895 tail_part = ''
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]896 else:
897 tail_part = ''
Senthil Kumarand70846b2012-04-12 02:34:32 +0800[diff] [blame]898
899 splitpath = ('/' + '/'.join(head_parts), tail_part)
900 collapsed_path = "/".join(splitpath)
901
902 return collapsed_path
903
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]904
905
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]906nobody = None
907
908def nobody_uid():
909 """Internal routine to get nobody's uid"""
910 global nobody
911 if nobody:
912 return nobody
913 try:
914 import pwd
Brett Cannoncd171c82013-07-04 17:43:24 -0400[diff] [blame]915 except ImportError:
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]916 return -1
917 try:
918 nobody = pwd.getpwnam('nobody')[2]
919 except KeyError:
Georg Brandlcbd2ab12010-12-04 10:39:14 +0000[diff] [blame]920 nobody = 1 + max(x[2] for x in pwd.getpwall())
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]921 return nobody
922
923
924def executable(path):
925 """Test for executable file."""
Victor Stinnerfb25ba92011-06-20 17:45:54 +0200[diff] [blame]926 return os.access(path, os.X_OK)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]927
928
929class CGIHTTPRequestHandler(SimpleHTTPRequestHandler):
930
931 """Complete HTTP server with GET, HEAD and POST commands.
932
933 GET and HEAD also support running CGI scripts.
934
935 The POST command is *only* implemented for CGI scripts.
936
937 """
938
939 # Determine platform specifics
940 have_fork = hasattr(os, 'fork')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]941
942 # Make rfile unbuffered -- we need to read one line and then pass
943 # the rest to a subprocess, so we can't use buffered input.
944 rbufsize = 0
945
946 def do_POST(self):
947 """Serve a POST request.
948
949 This is only implemented for CGI scripts.
950
951 """
952
953 if self.is_cgi():
954 self.run_cgi()
955 else:
956 self.send_error(501, "Can only POST to CGI scripts")
957
958 def send_head(self):
959 """Version of send_head that support CGI scripts"""
960 if self.is_cgi():
961 return self.run_cgi()
962 else:
963 return SimpleHTTPRequestHandler.send_head(self)
964
965 def is_cgi(self):
966 """Test whether self.path corresponds to a CGI script.
967
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]968 Returns True and updates the cgi_info attribute to the tuple
969 (dir, rest) if self.path requires running a CGI script.
970 Returns False otherwise.
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]971
Benjamin Petersona7deeee2009-05-08 20:54:42 +0000[diff] [blame]972 If any exception is raised, the caller should assume that
973 self.path was rejected as invalid and act accordingly.
974
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]975 The default implementation tests whether the normalized url
976 path begins with one of the strings in self.cgi_directories
977 (and the next character is a '/' or the end of the string).
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]978
979 """
Benjamin Peterson73b8b1c2014-06-14 18:36:29 -0700[diff] [blame]980 collapsed_path = _url_collapse_path(urllib.parse.unquote(self.path))
Senthil Kumarand70846b2012-04-12 02:34:32 +0800[diff] [blame]981 dir_sep = collapsed_path.find('/', 1)
982 head, tail = collapsed_path[:dir_sep], collapsed_path[dir_sep+1:]
Senthil Kumarandbb369d2012-04-11 03:15:28 +0800[diff] [blame]983 if head in self.cgi_directories:
984 self.cgi_info = head, tail
Benjamin Petersonad71f0f2009-04-11 20:12:10 +0000[diff] [blame]985 return True
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]986 return False
987
Senthil Kumarand70846b2012-04-12 02:34:32 +0800[diff] [blame]988
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]989 cgi_directories = ['/cgi-bin', '/htbin']
990
991 def is_executable(self, path):
992 """Test whether argument path is an executable file."""
993 return executable(path)
994
995 def is_python(self, path):
996 """Test whether argument path is a Python script."""
997 head, tail = os.path.splitext(path)
998 return tail.lower() in (".py", ".pyw")
999
1000 def run_cgi(self):
1001 """Execute a CGI script."""
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1002 dir, rest = self.cgi_info
Ned Deily915a30f2014-07-12 22:06:26 -0700[diff] [blame]1003 path = dir + '/' + rest
1004 i = path.find('/', len(dir)+1)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1005 while i >= 0:
Ned Deily915a30f2014-07-12 22:06:26 -0700[diff] [blame]1006 nextdir = path[:i]
1007 nextrest = path[i+1:]
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1008
1009 scriptdir = self.translate_path(nextdir)
1010 if os.path.isdir(scriptdir):
1011 dir, rest = nextdir, nextrest
Ned Deily915a30f2014-07-12 22:06:26 -0700[diff] [blame]1012 i = path.find('/', len(dir)+1)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1013 else:
1014 break
1015
1016 # find an explicit query string, if present.
1017 i = rest.rfind('?')
1018 if i >= 0:
1019 rest, query = rest[:i], rest[i+1:]
1020 else:
1021 query = ''
1022
1023 # dissect the part after the directory name into a script name &
1024 # a possible additional path, to be stored in PATH_INFO.
1025 i = rest.find('/')
1026 if i >= 0:
1027 script, rest = rest[:i], rest[i:]
1028 else:
1029 script, rest = rest, ''
1030
1031 scriptname = dir + '/' + script
1032 scriptfile = self.translate_path(scriptname)
1033 if not os.path.exists(scriptfile):
1034 self.send_error(404, "No such CGI script (%r)" % scriptname)
1035 return
1036 if not os.path.isfile(scriptfile):
1037 self.send_error(403, "CGI script is not a plain file (%r)" %
1038 scriptname)
1039 return
1040 ispy = self.is_python(scriptname)
Victor Stinnerfb25ba92011-06-20 17:45:54 +0200[diff] [blame]1041 if self.have_fork or not ispy:
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1042 if not self.is_executable(scriptfile):
1043 self.send_error(403, "CGI script is not executable (%r)" %
1044 scriptname)
1045 return
1046
1047 # Reference: http://hoohoo.ncsa.uiuc.edu/cgi/env.html
1048 # XXX Much of the following could be prepared ahead of time!
Senthil Kumaran42713722010-10-03 17:55:45 +0000[diff] [blame]1049 env = copy.deepcopy(os.environ)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1050 env['SERVER_SOFTWARE'] = self.version_string()
1051 env['SERVER_NAME'] = self.server.server_name
1052 env['GATEWAY_INTERFACE'] = 'CGI/1.1'
1053 env['SERVER_PROTOCOL'] = self.protocol_version
1054 env['SERVER_PORT'] = str(self.server.server_port)
1055 env['REQUEST_METHOD'] = self.command
Jeremy Hylton1afc1692008-06-18 20:49:58 +0000[diff] [blame]1056 uqrest = urllib.parse.unquote(rest)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1057 env['PATH_INFO'] = uqrest
1058 env['PATH_TRANSLATED'] = self.translate_path(uqrest)
1059 env['SCRIPT_NAME'] = scriptname
1060 if query:
1061 env['QUERY_STRING'] = query
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1062 env['REMOTE_ADDR'] = self.client_address[0]
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1063 authorization = self.headers.get("authorization")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1064 if authorization:
1065 authorization = authorization.split()
1066 if len(authorization) == 2:
1067 import base64, binascii
1068 env['AUTH_TYPE'] = authorization[0]
1069 if authorization[0].lower() == "basic":
1070 try:
1071 authorization = authorization[1].encode('ascii')
Georg Brandl706824f2009-06-04 09:42:55 +0000[diff] [blame]1072 authorization = base64.decodebytes(authorization).\
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1073 decode('ascii')
1074 except (binascii.Error, UnicodeError):
1075 pass
1076 else:
1077 authorization = authorization.split(':')
1078 if len(authorization) == 2:
1079 env['REMOTE_USER'] = authorization[0]
1080 # XXX REMOTE_IDENT
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1081 if self.headers.get('content-type') is None:
1082 env['CONTENT_TYPE'] = self.headers.get_content_type()
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1083 else:
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1084 env['CONTENT_TYPE'] = self.headers['content-type']
1085 length = self.headers.get('content-length')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1086 if length:
1087 env['CONTENT_LENGTH'] = length
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1088 referer = self.headers.get('referer')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1089 if referer:
1090 env['HTTP_REFERER'] = referer
1091 accept = []
1092 for line in self.headers.getallmatchingheaders('accept'):
1093 if line[:1] in "\t\n\r ":
1094 accept.append(line.strip())
1095 else:
1096 accept = accept + line[7:].split(',')
1097 env['HTTP_ACCEPT'] = ','.join(accept)
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1098 ua = self.headers.get('user-agent')
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1099 if ua:
1100 env['HTTP_USER_AGENT'] = ua
Barry Warsaw820c1202008-06-12 04:06:45 +0000[diff] [blame]1101 co = filter(None, self.headers.get_all('cookie', []))
Georg Brandl62e2ca22010-07-31 21:54:24 +0000[diff] [blame]1102 cookie_str = ', '.join(co)
1103 if cookie_str:
1104 env['HTTP_COOKIE'] = cookie_str
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1105 # XXX Other HTTP_* headers
1106 # Since we're setting the env in the parent, provide empty
1107 # values to override previously set values
1108 for k in ('QUERY_STRING', 'REMOTE_HOST', 'CONTENT_LENGTH',
1109 'HTTP_USER_AGENT', 'HTTP_COOKIE', 'HTTP_REFERER'):
1110 env.setdefault(k, "")
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1111
1112 self.send_response(200, "Script output follows")
Senthil Kumaranc7ae19b2011-05-09 23:25:02 +0800[diff] [blame]1113 self.flush_headers()
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1114
1115 decoded_query = query.replace('+', ' ')
1116
1117 if self.have_fork:
1118 # Unix -- fork as we should
1119 args = [script]
1120 if '=' not in decoded_query:
1121 args.append(decoded_query)
1122 nobody = nobody_uid()
1123 self.wfile.flush() # Always flush before forking
1124 pid = os.fork()
1125 if pid != 0:
1126 # Parent
1127 pid, sts = os.waitpid(pid, 0)
1128 # throw away additional data [see bug #427345]
1129 while select.select([self.rfile], [], [], 0)[0]:
1130 if not self.rfile.read(1):
1131 break
1132 if sts:
1133 self.log_error("CGI script exit status %#x", sts)
1134 return
1135 # Child
1136 try:
1137 try:
1138 os.setuid(nobody)
Andrew Svetlovad28c7f2012-12-18 22:02:39 +0200[diff] [blame]1139 except OSError:
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1140 pass
1141 os.dup2(self.rfile.fileno(), 0)
1142 os.dup2(self.wfile.fileno(), 1)
Senthil Kumaran42713722010-10-03 17:55:45 +0000[diff] [blame]1143 os.execve(scriptfile, args, env)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1144 except:
1145 self.server.handle_error(self.request, self.client_address)
1146 os._exit(127)
1147
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1148 else:
1149 # Non-Unix -- use subprocess
1150 import subprocess
Senthil Kumarane29cd162009-11-11 04:17:53 +0000[diff] [blame]1151 cmdline = [scriptfile]
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1152 if self.is_python(scriptfile):
1153 interp = sys.executable
1154 if interp.lower().endswith("w.exe"):
1155 # On Windows, use python.exe, not pythonw.exe
1156 interp = interp[:-5] + interp[-4:]
Senthil Kumarane29cd162009-11-11 04:17:53 +0000[diff] [blame]1157 cmdline = [interp, '-u'] + cmdline
1158 if '=' not in query:
1159 cmdline.append(query)
1160 self.log_message("command: %s", subprocess.list2cmdline(cmdline))
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1161 try:
1162 nbytes = int(length)
1163 except (TypeError, ValueError):
1164 nbytes = 0
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1165 p = subprocess.Popen(cmdline,
1166 stdin=subprocess.PIPE,
1167 stdout=subprocess.PIPE,
Senthil Kumaran42713722010-10-03 17:55:45 +0000[diff] [blame]1168 stderr=subprocess.PIPE,
1169 env = env
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1170 )
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1171 if self.command.lower() == "post" and nbytes > 0:
1172 data = self.rfile.read(nbytes)
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1173 else:
1174 data = None
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1175 # throw away additional data [see bug #427345]
1176 while select.select([self.rfile._sock], [], [], 0)[0]:
1177 if not self.rfile._sock.recv(1):
1178 break
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1179 stdout, stderr = p.communicate(data)
1180 self.wfile.write(stdout)
1181 if stderr:
1182 self.log_error('%s', stderr)
Brian Curtincbad4df2010-11-05 15:04:48 +0000[diff] [blame]1183 p.stderr.close()
1184 p.stdout.close()
Amaury Forgeot d'Arccb0d2d72008-06-18 22:19:22 +0000[diff] [blame]1185 status = p.returncode
1186 if status:
1187 self.log_error("CGI script exit status %#x", status)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1188 else:
1189 self.log_message("CGI script exited OK")
1190
1191
Senthil Kumarandefe7f42013-09-15 09:37:27 -0700[diff] [blame]1192def test(HandlerClass=BaseHTTPRequestHandler,
1193 ServerClass=HTTPServer, protocol="HTTP/1.0", port=8000, bind=""):
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1194 """Test the HTTP request handler class.
1195
1196 This runs an HTTP server on port 8000 (or the first command line
1197 argument).
1198
1199 """
Senthil Kumarandefe7f42013-09-15 09:37:27 -0700[diff] [blame]1200 server_address = (bind, port)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1201
1202 HandlerClass.protocol_version = protocol
1203 httpd = ServerClass(server_address, HandlerClass)
1204
1205 sa = httpd.socket.getsockname()
1206 print("Serving HTTP on", sa[0], "port", sa[1], "...")
Alexandre Vassalottib5292a22009-04-03 07:16:55 +0000[diff] [blame]1207 try:
1208 httpd.serve_forever()
1209 except KeyboardInterrupt:
1210 print("\nKeyboard interrupt received, exiting.")
1211 httpd.server_close()
1212 sys.exit(0)
Georg Brandl24420152008-05-26 16:32:26 +0000[diff] [blame]1213
1214if __name__ == '__main__':
Senthil Kumaran1251faf2012-06-03 16:15:54 +0800[diff] [blame]1215 parser = argparse.ArgumentParser()
1216 parser.add_argument('--cgi', action='store_true',
1217 help='Run as CGI Server')
Senthil Kumarandefe7f42013-09-15 09:37:27 -0700[diff] [blame]1218 parser.add_argument('--bind', '-b', default='', metavar='ADDRESS',
1219 help='Specify alternate bind address '
1220 '[default: all interfaces]')
Senthil Kumaran1251faf2012-06-03 16:15:54 +0800[diff] [blame]1221 parser.add_argument('port', action='store',
1222 default=8000, type=int,
1223 nargs='?',
1224 help='Specify alternate port [default: 8000]')
1225 args = parser.parse_args()
1226 if args.cgi:
Senthil Kumarandefe7f42013-09-15 09:37:27 -0700[diff] [blame]1227 handler_class = CGIHTTPRequestHandler
Senthil Kumaran1251faf2012-06-03 16:15:54 +0800[diff] [blame]1228 else:
Senthil Kumarandefe7f42013-09-15 09:37:27 -0700[diff] [blame]1229 handler_class = SimpleHTTPRequestHandler
1230 test(HandlerClass=handler_class, port=args.port, bind=args.bind)