1
"""Generic socket server classes.
3
This module tries to capture the various aspects of defining a server:
5
For socket-based servers:
8
- AF_INET{,6}: IP (Internet Protocol) sockets (default)
9
- AF_UNIX: Unix domain sockets
10
- others, e.g. AF_DECNET are conceivable (see <socket.h>
12
- SOCK_STREAM (reliable stream, e.g. TCP)
13
- SOCK_DGRAM (datagrams, e.g. UDP)
15
For request-based servers (including socket-based):
17
- client address verification before further looking at the request
18
(This is actually a hook for any processing that needs to look
19
at the request before anything else, e.g. logging)
20
- how to handle multiple requests:
21
- synchronous (one request is handled at a time)
22
- forking (each request is handled by a new process)
23
- threading (each request is handled by a new thread)
25
The classes in this module favor the server type that is simplest to
26
write: a synchronous TCP/IP server. This is bad class design, but
27
save some typing. (There's also the issue that a deep class hierarchy
28
slows down method lookups.)
30
There are five classes in an inheritance diagram, four of which represent
31
synchronous servers of four types:
38
+-----------+ +------------------+
39
| TCPServer |------->| UnixStreamServer |
40
+-----------+ +------------------+
43
+-----------+ +--------------------+
44
| UDPServer |------->| UnixDatagramServer |
45
+-----------+ +--------------------+
47
Note that UnixDatagramServer derives from UDPServer, not from
48
UnixStreamServer -- the only difference between an IP and a Unix
49
stream server is the address family, which is simply repeated in both
52
Forking and threading versions of each type of server can be created
53
using the ForkingMixIn and ThreadingMixIn mix-in classes. For
54
instance, a threading UDP server class is created as follows:
56
class ThreadingUDPServer(ThreadingMixIn, UDPServer): pass
58
The Mix-in class must come first, since it overrides a method defined
59
in UDPServer! Setting the various member variables also changes
60
the behavior of the underlying server mechanism.
62
To implement a service, you must derive a class from
63
BaseRequestHandler and redefine its handle() method. You can then run
64
various versions of the service by combining one of the server classes
65
with your request handler class.
67
The request handler class must be different for datagram or stream
68
services. This can be hidden by using the request handler
69
subclasses StreamRequestHandler or DatagramRequestHandler.
71
Of course, you still have to use your head!
73
For instance, it makes no sense to use a forking server if the service
74
contains state in memory that can be modified by requests (since the
75
modifications in the child process would never reach the initial state
76
kept in the parent process and passed to each child). In this case,
77
you can use a threading server, but you will probably have to use
78
locks to avoid two requests that come in nearly simultaneous to apply
79
conflicting changes to the server state.
81
On the other hand, if you are building e.g. an HTTP server, where all
82
data is stored externally (e.g. in the file system), a synchronous
83
class will essentially render the service "deaf" while one request is
84
being handled -- which may be for a very long time if a client is slow
85
to reqd all the data it has requested. Here a threading or forking
86
server is appropriate.
88
In some cases, it may be appropriate to process part of a request
89
synchronously, but to finish processing in a forked child depending on
90
the request data. This can be implemented by using a synchronous
91
server and doing an explicit fork in the request handler class
94
Another approach to handling multiple simultaneous requests in an
95
environment that supports neither threads nor fork (or where these are
96
too expensive or inappropriate for the service) is to maintain an
97
explicit table of partially finished requests and to use select() to
98
decide which request to work on next (or whether to handle a new
99
incoming request). This is particularly important for stream services
100
where each client can potentially be connected for a long time (if
101
threads or subprocesses cannot be used).
104
- Standard classes for Sun RPC (which uses either UDP or TCP)
105
- Standard mix-in classes to implement various authentication
106
and encryption schemes
107
- Standard framework for select-based multiplexing
110
- What to do with out-of-band data?
113
- split generic "request" functionality out into BaseServer class.
114
Copyright (C) 2000 Luke Kenneth Casson Leighton <lkcl@samba.org>
116
example: read entries from a SQL database (requires overriding
117
get_request() to return a table entry from the database).
118
entry is processed by a RequestHandlerClass.
122
# Author of the BaseServer patch: Luke Kenneth Casson Leighton
125
# There is a test suite for this module, but it cannot be run by the
126
# standard regression test.
127
# To run it manually, run Lib/test/test_socketserver.py.
139
import dummy_threading as threading
141
__all__ = ["TCPServer","UDPServer","ForkingUDPServer","ForkingTCPServer",
142
"ThreadingUDPServer","ThreadingTCPServer","BaseRequestHandler",
143
"StreamRequestHandler","DatagramRequestHandler",
144
"ThreadingMixIn", "ForkingMixIn"]
145
if hasattr(socket, "AF_UNIX"):
146
__all__.extend(["UnixStreamServer","UnixDatagramServer",
147
"ThreadingUnixStreamServer",
148
"ThreadingUnixDatagramServer"])
152
"""Base class for server classes.
154
Methods for the caller:
156
- __init__(server_address, RequestHandlerClass)
157
- serve_forever(poll_interval=0.5)
159
- handle_request() # if you do not use serve_forever()
160
- fileno() -> int # for select()
162
Methods that may be overridden:
166
- get_request() -> request, client_address
168
- verify_request(request, client_address)
170
- process_request(request, client_address)
171
- close_request(request)
174
Methods for derived classes:
176
- finish_request(request, client_address)
178
Class variables that may be overridden by derived classes or
184
- allow_reuse_address
188
- RequestHandlerClass
195
def __init__(self, server_address, RequestHandlerClass):
196
"""Constructor. May be extended, do not override."""
197
self.server_address = server_address
198
self.RequestHandlerClass = RequestHandlerClass
199
self.__is_shut_down = threading.Event()
200
self.__serving = False
202
def server_activate(self):
203
"""Called by constructor to activate the server.
210
def serve_forever(self, poll_interval=0.5):
211
"""Handle one request at a time until shutdown.
213
Polls for shutdown every poll_interval seconds. Ignores
214
self.timeout. If you need to do periodic tasks, do them in
217
self.__serving = True
218
self.__is_shut_down.clear()
219
while self.__serving:
220
# XXX: Consider using another file descriptor or
221
# connecting to the socket to wake this up instead of
222
# polling. Polling reduces our responsiveness to a
223
# shutdown request and wastes cpu at all other times.
224
r, w, e = select.select([self], [], [], poll_interval)
226
self._handle_request_noblock()
227
self.__is_shut_down.set()
230
"""Stops the serve_forever loop.
232
Blocks until the loop has finished. This must be called while
233
serve_forever() is running in another thread, or it will
236
self.__serving = False
237
self.__is_shut_down.wait()
239
# The distinction between handling, getting, processing and
240
# finishing a request is fairly arbitrary. Remember:
242
# - handle_request() is the top-level call. It calls
243
# select, get_request(), verify_request() and process_request()
244
# - get_request() is different for stream or datagram sockets
245
# - process_request() is the place that may fork a new process
246
# or create a new thread to finish the request
247
# - finish_request() instantiates the request handler class;
248
# this constructor will handle the request all by itself
250
def handle_request(self):
251
"""Handle one request, possibly blocking.
253
Respects self.timeout.
255
# Support people who used socket.settimeout() to escape
256
# handle_request before self.timeout was available.
257
timeout = self.socket.gettimeout()
259
timeout = self.timeout
260
elif self.timeout is not None:
261
timeout = min(timeout, self.timeout)
262
fd_sets = select.select([self], [], [], timeout)
264
self.handle_timeout()
266
self._handle_request_noblock()
268
def _handle_request_noblock(self):
269
"""Handle one request, without blocking.
271
I assume that select.select has returned that the socket is
272
readable before this function was called, so there should be
273
no risk of blocking in get_request().
276
request, client_address = self.get_request()
279
if self.verify_request(request, client_address):
281
self.process_request(request, client_address)
283
self.handle_error(request, client_address)
284
self.close_request(request)
286
def handle_timeout(self):
287
"""Called if no new request arrives within self.timeout.
289
Overridden by ForkingMixIn.
293
def verify_request(self, request, client_address):
294
"""Verify the request. May be overridden.
296
Return True if we should proceed with this request.
301
def process_request(self, request, client_address):
302
"""Call finish_request.
304
Overridden by ForkingMixIn and ThreadingMixIn.
307
self.finish_request(request, client_address)
308
self.close_request(request)
310
def server_close(self):
311
"""Called to clean-up the server.
318
def finish_request(self, request, client_address):
319
"""Finish one request by instantiating RequestHandlerClass."""
320
self.RequestHandlerClass(request, client_address, self)
322
def close_request(self, request):
323
"""Called to clean up an individual request."""
326
def handle_error(self, request, client_address):
327
"""Handle an error gracefully. May be overridden.
329
The default is to print a traceback and continue.
333
print 'Exception happened during processing of request from',
336
traceback.print_exc() # XXX But this goes to stderr!
340
class TCPServer(BaseServer):
342
"""Base class for various socket-based server classes.
344
Defaults to synchronous IP stream (i.e., TCP).
346
Methods for the caller:
348
- __init__(server_address, RequestHandlerClass, bind_and_activate=True)
349
- serve_forever(poll_interval=0.5)
351
- handle_request() # if you don't use serve_forever()
352
- fileno() -> int # for select()
354
Methods that may be overridden:
358
- get_request() -> request, client_address
360
- verify_request(request, client_address)
361
- process_request(request, client_address)
362
- close_request(request)
365
Methods for derived classes:
367
- finish_request(request, client_address)
369
Class variables that may be overridden by derived classes or
375
- request_queue_size (only for stream sockets)
376
- allow_reuse_address
381
- RequestHandlerClass
386
address_family = socket.AF_INET
388
socket_type = socket.SOCK_STREAM
390
request_queue_size = 5
392
allow_reuse_address = False
394
def __init__(self, server_address, RequestHandlerClass, bind_and_activate=True):
395
"""Constructor. May be extended, do not override."""
396
BaseServer.__init__(self, server_address, RequestHandlerClass)
397
self.socket = socket.socket(self.address_family,
399
if bind_and_activate:
401
self.server_activate()
403
def server_bind(self):
404
"""Called by constructor to bind the socket.
409
if self.allow_reuse_address:
410
self.socket.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
411
self.socket.bind(self.server_address)
412
self.server_address = self.socket.getsockname()
414
def server_activate(self):
415
"""Called by constructor to activate the server.
420
self.socket.listen(self.request_queue_size)
422
def server_close(self):
423
"""Called to clean-up the server.
431
"""Return socket file number.
433
Interface required by select().
436
return self.socket.fileno()
438
def get_request(self):
439
"""Get the request and client address from the socket.
444
return self.socket.accept()
446
def close_request(self, request):
447
"""Called to clean up an individual request."""
451
class UDPServer(TCPServer):
453
"""UDP server class."""
455
allow_reuse_address = False
457
socket_type = socket.SOCK_DGRAM
459
max_packet_size = 8192
461
def get_request(self):
462
data, client_addr = self.socket.recvfrom(self.max_packet_size)
463
return (data, self.socket), client_addr
465
def server_activate(self):
466
# No need to call listen() for UDP.
469
def close_request(self, request):
470
# No need to close anything.
475
"""Mix-in class to handle each request in a new process."""
478
active_children = None
481
def collect_children(self):
482
"""Internal routine to wait for children that have exited."""
483
if self.active_children is None: return
484
while len(self.active_children) >= self.max_children:
485
# XXX: This will wait for any child process, not just ones
486
# spawned by this library. This could confuse other
487
# libraries that expect to be able to wait for their own
490
pid, status = os.waitpid(0, options=0)
493
if pid not in self.active_children: continue
494
self.active_children.remove(pid)
496
# XXX: This loop runs more system calls than it ought
497
# to. There should be a way to put the active_children into a
498
# process group and then use os.waitpid(-pgid) to wait for any
499
# of that set, but I couldn't find a way to allocate pgids
500
# that couldn't collide.
501
for child in self.active_children:
503
pid, status = os.waitpid(child, os.WNOHANG)
508
self.active_children.remove(pid)
509
except ValueError, e:
510
raise ValueError('%s. x=%d and list=%r' % (e.message, pid,
511
self.active_children))
513
def handle_timeout(self):
514
"""Wait for zombies after self.timeout seconds of inactivity.
516
May be extended, do not override.
518
self.collect_children()
520
def process_request(self, request, client_address):
521
"""Fork a new subprocess to process the request."""
522
self.collect_children()
526
if self.active_children is None:
527
self.active_children = []
528
self.active_children.append(pid)
529
self.close_request(request)
533
# This must never return, hence os._exit()!
535
self.finish_request(request, client_address)
539
self.handle_error(request, client_address)
544
class ThreadingMixIn:
545
"""Mix-in class to handle each request in a new thread."""
547
# Decides how threads will act upon termination of the
549
daemon_threads = False
551
def process_request_thread(self, request, client_address):
552
"""Same as in BaseServer but as a thread.
554
In addition, exception handling is done here.
558
self.finish_request(request, client_address)
559
self.close_request(request)
561
self.handle_error(request, client_address)
562
self.close_request(request)
564
def process_request(self, request, client_address):
565
"""Start a new thread to process the request."""
566
t = threading.Thread(target = self.process_request_thread,
567
args = (request, client_address))
568
if self.daemon_threads:
573
class ForkingUDPServer(ForkingMixIn, UDPServer): pass
574
class ForkingTCPServer(ForkingMixIn, TCPServer): pass
576
class ThreadingUDPServer(ThreadingMixIn, UDPServer): pass
577
class ThreadingTCPServer(ThreadingMixIn, TCPServer): pass
579
if hasattr(socket, 'AF_UNIX'):
581
class UnixStreamServer(TCPServer):
582
address_family = socket.AF_UNIX
584
class UnixDatagramServer(UDPServer):
585
address_family = socket.AF_UNIX
587
class ThreadingUnixStreamServer(ThreadingMixIn, UnixStreamServer): pass
589
class ThreadingUnixDatagramServer(ThreadingMixIn, UnixDatagramServer): pass
591
class BaseRequestHandler:
593
"""Base class for request handler classes.
595
This class is instantiated for each request to be handled. The
596
constructor sets the instance variables request, client_address
597
and server, and then calls the handle() method. To implement a
598
specific service, all you need to do is to derive a class which
599
defines a handle() method.
601
The handle() method can find the request as self.request, the
602
client address as self.client_address, and the server (in case it
603
needs access to per-server information) as self.server. Since a
604
separate instance is created for each request, the handle() method
605
can define arbitrary other instance variariables.
609
def __init__(self, request, client_address, server):
610
self.request = request
611
self.client_address = client_address
618
sys.exc_traceback = None # Help garbage collection
630
# The following two classes make it possible to use the same service
631
# class for stream or datagram servers.
632
# Each class sets up these instance variables:
633
# - rfile: a file object from which receives the request is read
634
# - wfile: a file object to which the reply is written
635
# When the handle() method returns, wfile is flushed properly
638
class StreamRequestHandler(BaseRequestHandler):
640
"""Define self.rfile and self.wfile for stream sockets."""
642
# Default buffer sizes for rfile, wfile.
643
# We default rfile to buffered because otherwise it could be
644
# really slow for large data (a getc() call per byte); we make
645
# wfile unbuffered because (a) often after a write() we want to
646
# read and we need to flush the line; (b) big writes to unbuffered
647
# files are typically optimized by stdio even when big reads
653
self.connection = self.request
654
self.rfile = self.connection.makefile('rb', self.rbufsize)
655
self.wfile = self.connection.makefile('wb', self.wbufsize)
658
if not self.wfile.closed:
664
class DatagramRequestHandler(BaseRequestHandler):
666
# XXX Regrettably, I cannot get this working on Linux;
667
# s.recvfrom() doesn't return a meaningful client address.
669
"""Define self.rfile and self.wfile for datagram sockets."""
673
from cStringIO import StringIO
675
from StringIO import StringIO
676
self.packet, self.socket = self.request
677
self.rfile = StringIO(self.packet)
678
self.wfile = StringIO()
681
self.socket.sendto(self.wfile.getvalue(), self.client_address)