blob: 65585129ceaaa9b40d1292fdee08ac70f1250521 [file] [log] [blame]
Guido van Rossum0039d7b1999-01-12 20:19:27 +00001# -*- Mode: Python; tab-width: 4 -*-
Tim Peters658cba62001-02-09 20:06:00 +00002# Id: asynchat.py,v 2.26 2000/09/07 22:29:26 rushing Exp
Tim Peters146965a2001-01-14 18:09:23 +00003# Author: Sam Rushing <rushing@nightmare.com>
Guido van Rossum0039d7b1999-01-12 20:19:27 +00004
5# ======================================================================
6# Copyright 1996 by Sam Rushing
Tim Peters146965a2001-01-14 18:09:23 +00007#
Guido van Rossum0039d7b1999-01-12 20:19:27 +00008# All Rights Reserved
Tim Peters146965a2001-01-14 18:09:23 +00009#
Guido van Rossum0039d7b1999-01-12 20:19:27 +000010# Permission to use, copy, modify, and distribute this software and
11# its documentation for any purpose and without fee is hereby
12# granted, provided that the above copyright notice appear in all
13# copies and that both that copyright notice and this permission
14# notice appear in supporting documentation, and that the name of Sam
15# Rushing not be used in advertising or publicity pertaining to
16# distribution of the software without specific, written prior
17# permission.
Tim Peters146965a2001-01-14 18:09:23 +000018#
Guido van Rossum0039d7b1999-01-12 20:19:27 +000019# SAM RUSHING DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
20# INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN
21# NO EVENT SHALL SAM RUSHING BE LIABLE FOR ANY SPECIAL, INDIRECT OR
22# CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS
23# OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT,
24# NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
25# CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
26# ======================================================================
27
Guido van Rossume4a1b6d2001-04-06 15:30:33 +000028r"""A class supporting chat-style (command/response) protocols.
Guido van Rossum4b8c6ea2000-02-04 15:39:30 +000029
30This class adds support for 'chat' style protocols - where one side
31sends a 'command', and the other sends a response (examples would be
32the common internet protocols - smtp, nntp, ftp, etc..).
33
34The handle_read() method looks at the input stream for the current
35'terminator' (usually '\r\n' for single-line responses, '\r\n.\r\n'
36for multi-line output), calling self.found_terminator() on its
37receipt.
38
39for example:
40Say you build an async nntp client using this class. At the start
41of the connection, you'll have self.terminator set to '\r\n', in
42order to process the single-line greeting. Just before issuing a
43'LIST' command you'll set it to '\r\n.\r\n'. The output of the LIST
44command will be accumulated (using your own 'collect_incoming_data'
45method) up to the terminator, and then control will be returned to
46you - by calling your self.found_terminator() method.
47"""
Guido van Rossum0039d7b1999-01-12 20:19:27 +000048import socket
49import asyncore
Raymond Hettingerac093c62004-02-07 03:19:10 +000050from collections import deque
Guido van Rossum0039d7b1999-01-12 20:19:27 +000051
Josiah Carlsond74900e2008-07-07 04:15:08 +000052def buffer(obj, start=None, stop=None):
53 # if memoryview objects gain slicing semantics,
54 # this function will change for the better
55 # memoryview used for the TypeError
56 memoryview(obj)
57 if start == None:
58 start = 0
59 if stop == None:
60 stop = len(obj)
61 x = obj[start:stop]
62 ## print("buffer type is: %s"%(type(x),))
63 return x
64
Guido van Rossum0039d7b1999-01-12 20:19:27 +000065class async_chat (asyncore.dispatcher):
Tim Peters146965a2001-01-14 18:09:23 +000066 """This is an abstract class. You must derive from this class, and add
67 the two methods collect_incoming_data() and found_terminator()"""
Guido van Rossum0039d7b1999-01-12 20:19:27 +000068
Tim Peters146965a2001-01-14 18:09:23 +000069 # these are overridable defaults
Guido van Rossum0039d7b1999-01-12 20:19:27 +000070
Tim Peters146965a2001-01-14 18:09:23 +000071 ac_in_buffer_size = 4096
72 ac_out_buffer_size = 4096
Guido van Rossum0039d7b1999-01-12 20:19:27 +000073
Josiah Carlsond74900e2008-07-07 04:15:08 +000074 # we don't want to enable the use of encoding by default, because that is a
75 # sign of an application bug that we don't want to pass silently
76
77 use_encoding = 0
78 encoding = 'latin1'
79
Josiah Carlson9f2f8332008-07-07 05:04:12 +000080 def __init__ (self, sock=None, map=None):
Josiah Carlsond74900e2008-07-07 04:15:08 +000081 # for string terminator matching
Guido van Rossum076da092007-07-12 07:58:54 +000082 self.ac_in_buffer = b''
Josiah Carlsond74900e2008-07-07 04:15:08 +000083
84 # we use a list here rather than cStringIO for a few reasons...
85 # del lst[:] is faster than sio.truncate(0)
86 # lst = [] is faster than sio.truncate(0)
87 # cStringIO will be gaining unicode support in py3k, which
88 # will negatively affect the performance of bytes compared to
89 # a ''.join() equivalent
90 self.incoming = []
91
92 # we toss the use of the "simple producer" and replace it with
93 # a pure deque, which the original fifo was a wrapping of
94 self.producer_fifo = deque()
Josiah Carlson9f2f8332008-07-07 05:04:12 +000095 asyncore.dispatcher.__init__ (self, sock, map)
Guido van Rossum0039d7b1999-01-12 20:19:27 +000096
Andrew M. Kuchling7dd5f3c2002-03-08 18:27:11 +000097 def collect_incoming_data(self, data):
Collin Winterce36ad82007-08-30 01:19:48 +000098 raise NotImplementedError("must be implemented in subclass")
Tim Peters863ac442002-04-16 01:38:40 +000099
Josiah Carlsond74900e2008-07-07 04:15:08 +0000100 def _collect_incoming_data(self, data):
101 self.incoming.append(data)
102
103 def _get_data(self):
104 d = b''.join(self.incoming)
105 del self.incoming[:]
106 return d
107
Andrew M. Kuchling7dd5f3c2002-03-08 18:27:11 +0000108 def found_terminator(self):
Collin Winterce36ad82007-08-30 01:19:48 +0000109 raise NotImplementedError("must be implemented in subclass")
Tim Peters863ac442002-04-16 01:38:40 +0000110
Tim Peters146965a2001-01-14 18:09:23 +0000111 def set_terminator (self, term):
112 "Set the input delimiter. Can be a fixed string of any length, an integer, or None"
Josiah Carlsond74900e2008-07-07 04:15:08 +0000113 if isinstance(term, str) and self.use_encoding:
114 term = bytes(term, self.encoding)
Tim Peters146965a2001-01-14 18:09:23 +0000115 self.terminator = term
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000116
Tim Peters146965a2001-01-14 18:09:23 +0000117 def get_terminator (self):
118 return self.terminator
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000119
Tim Peters146965a2001-01-14 18:09:23 +0000120 # grab some more data from the socket,
121 # throw it to the collector method,
122 # check for the terminator,
123 # if found, transition to the next state.
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000124
Tim Peters146965a2001-01-14 18:09:23 +0000125 def handle_read (self):
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000126
Tim Peters146965a2001-01-14 18:09:23 +0000127 try:
128 data = self.recv (self.ac_in_buffer_size)
Guido van Rossumb940e112007-01-10 16:19:56 +0000129 except socket.error as why:
Tim Peters146965a2001-01-14 18:09:23 +0000130 self.handle_error()
131 return
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000132
Josiah Carlsond74900e2008-07-07 04:15:08 +0000133 if isinstance(data, str) and self.use_encoding:
134 data = bytes(str, self.encoding)
135 self.ac_in_buffer = self.ac_in_buffer + data
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000136
Tim Peters146965a2001-01-14 18:09:23 +0000137 # Continue to search for self.terminator in self.ac_in_buffer,
138 # while calling self.collect_incoming_data. The while loop
139 # is necessary because we might read several data+terminator
Josiah Carlsond74900e2008-07-07 04:15:08 +0000140 # combos with a single recv(4096).
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000141
Tim Peters146965a2001-01-14 18:09:23 +0000142 while self.ac_in_buffer:
143 lb = len(self.ac_in_buffer)
144 terminator = self.get_terminator()
Andrew M. Kuchlingca69f022005-06-09 14:59:45 +0000145 if not terminator:
Tim Peters146965a2001-01-14 18:09:23 +0000146 # no terminator, collect it all
147 self.collect_incoming_data (self.ac_in_buffer)
Guido van Rossum806c2462007-08-06 23:33:07 +0000148 self.ac_in_buffer = b''
Josiah Carlsond74900e2008-07-07 04:15:08 +0000149 elif isinstance(terminator, int):
Tim Peters146965a2001-01-14 18:09:23 +0000150 # numeric terminator
151 n = terminator
152 if lb < n:
153 self.collect_incoming_data (self.ac_in_buffer)
Guido van Rossum076da092007-07-12 07:58:54 +0000154 self.ac_in_buffer = b''
Tim Peters146965a2001-01-14 18:09:23 +0000155 self.terminator = self.terminator - lb
156 else:
157 self.collect_incoming_data (self.ac_in_buffer[:n])
158 self.ac_in_buffer = self.ac_in_buffer[n:]
159 self.terminator = 0
160 self.found_terminator()
161 else:
162 # 3 cases:
163 # 1) end of buffer matches terminator exactly:
164 # collect data, transition
165 # 2) end of buffer matches some prefix:
166 # collect data to the prefix
167 # 3) end of buffer does not match any prefix:
168 # collect data
169 terminator_len = len(terminator)
Tim Petersb5d13922001-04-05 22:38:32 +0000170 index = self.ac_in_buffer.find(terminator)
Tim Peters146965a2001-01-14 18:09:23 +0000171 if index != -1:
172 # we found the terminator
173 if index > 0:
174 # don't bother reporting the empty string (source of subtle bugs)
175 self.collect_incoming_data (self.ac_in_buffer[:index])
176 self.ac_in_buffer = self.ac_in_buffer[index+terminator_len:]
177 # This does the Right Thing if the terminator is changed here.
178 self.found_terminator()
179 else:
180 # check for a prefix of the terminator
181 index = find_prefix_at_end (self.ac_in_buffer, terminator)
182 if index:
183 if index != lb:
184 # we found a prefix, collect up to the prefix
185 self.collect_incoming_data (self.ac_in_buffer[:-index])
186 self.ac_in_buffer = self.ac_in_buffer[-index:]
187 break
188 else:
189 # no prefix, collect it all
190 self.collect_incoming_data (self.ac_in_buffer)
Guido van Rossum076da092007-07-12 07:58:54 +0000191 self.ac_in_buffer = b''
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000192
Tim Peters146965a2001-01-14 18:09:23 +0000193 def handle_write (self):
Josiah Carlsond74900e2008-07-07 04:15:08 +0000194 self.initiate_send()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000195
Tim Peters146965a2001-01-14 18:09:23 +0000196 def handle_close (self):
197 self.close()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000198
Tim Peters146965a2001-01-14 18:09:23 +0000199 def push (self, data):
Josiah Carlsond74900e2008-07-07 04:15:08 +0000200 sabs = self.ac_out_buffer_size
201 if len(data) > sabs:
202 for i in range(0, len(data), sabs):
203 self.producer_fifo.append(data[i:i+sabs])
204 else:
205 self.producer_fifo.append(data)
Tim Peters146965a2001-01-14 18:09:23 +0000206 self.initiate_send()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000207
Tim Peters146965a2001-01-14 18:09:23 +0000208 def push_with_producer (self, producer):
Josiah Carlsond74900e2008-07-07 04:15:08 +0000209 self.producer_fifo.append(producer)
Tim Peters146965a2001-01-14 18:09:23 +0000210 self.initiate_send()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000211
Tim Peters146965a2001-01-14 18:09:23 +0000212 def readable (self):
213 "predicate for inclusion in the readable for select()"
Josiah Carlsond74900e2008-07-07 04:15:08 +0000214 # cannot use the old predicate, it violates the claim of the
215 # set_terminator method.
216
217 # return (len(self.ac_in_buffer) <= self.ac_in_buffer_size)
218 return 1
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000219
Tim Peters146965a2001-01-14 18:09:23 +0000220 def writable (self):
221 "predicate for inclusion in the writable for select()"
Josiah Carlsond74900e2008-07-07 04:15:08 +0000222 return self.producer_fifo or (not self.connected)
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000223
Tim Peters146965a2001-01-14 18:09:23 +0000224 def close_when_done (self):
225 "automatically close this channel once the outgoing queue is empty"
Josiah Carlsond74900e2008-07-07 04:15:08 +0000226 self.producer_fifo.append(None)
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000227
Josiah Carlsond74900e2008-07-07 04:15:08 +0000228 def initiate_send(self):
229 while self.producer_fifo and self.connected:
230 first = self.producer_fifo[0]
231 # handle empty string/buffer or None entry
232 if not first:
233 del self.producer_fifo[0]
234 if first is None:
235 ## print("first is None")
236 self.handle_close()
Tim Peters146965a2001-01-14 18:09:23 +0000237 return
Josiah Carlsond74900e2008-07-07 04:15:08 +0000238 ## print("first is not None")
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000239
Josiah Carlsond74900e2008-07-07 04:15:08 +0000240 # handle classic producer behavior
241 obs = self.ac_out_buffer_size
Tim Peters146965a2001-01-14 18:09:23 +0000242 try:
Josiah Carlsond74900e2008-07-07 04:15:08 +0000243 data = buffer(first, 0, obs)
244 except TypeError:
245 data = first.more()
246 if data:
247 self.producer_fifo.appendleft(data)
248 else:
249 del self.producer_fifo[0]
250 continue
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000251
Josiah Carlsond74900e2008-07-07 04:15:08 +0000252 if isinstance(data, str) and self.use_encoding:
253 data = bytes(data, self.encoding)
254
255 # send the data
256 try:
257 num_sent = self.send(data)
258 except socket.error:
Tim Peters146965a2001-01-14 18:09:23 +0000259 self.handle_error()
260 return
261
Josiah Carlsond74900e2008-07-07 04:15:08 +0000262 if num_sent:
263 if num_sent < len(data) or obs < len(first):
264 self.producer_fifo[0] = first[num_sent:]
265 else:
266 del self.producer_fifo[0]
267 # we tried to send some actual data
268 return
269
Tim Peters146965a2001-01-14 18:09:23 +0000270 def discard_buffers (self):
271 # Emergencies only!
Guido van Rossum076da092007-07-12 07:58:54 +0000272 self.ac_in_buffer = b''
Josiah Carlsond74900e2008-07-07 04:15:08 +0000273 del self.incoming[:]
274 self.producer_fifo.clear()
Andrew M. Kuchlingda85a272000-09-08 20:30:39 +0000275
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000276class simple_producer:
Guido van Rossuma8d0f4f1999-06-08 13:20:05 +0000277
Tim Peters146965a2001-01-14 18:09:23 +0000278 def __init__ (self, data, buffer_size=512):
279 self.data = data
280 self.buffer_size = buffer_size
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000281
Tim Peters146965a2001-01-14 18:09:23 +0000282 def more (self):
283 if len (self.data) > self.buffer_size:
284 result = self.data[:self.buffer_size]
285 self.data = self.data[self.buffer_size:]
286 return result
287 else:
288 result = self.data
Guido van Rossum076da092007-07-12 07:58:54 +0000289 self.data = b''
Tim Peters146965a2001-01-14 18:09:23 +0000290 return result
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000291
292class fifo:
Tim Peters146965a2001-01-14 18:09:23 +0000293 def __init__ (self, list=None):
294 if not list:
Raymond Hettingerac093c62004-02-07 03:19:10 +0000295 self.list = deque()
Tim Peters146965a2001-01-14 18:09:23 +0000296 else:
Raymond Hettingerac093c62004-02-07 03:19:10 +0000297 self.list = deque(list)
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000298
Tim Peters146965a2001-01-14 18:09:23 +0000299 def __len__ (self):
300 return len(self.list)
Guido van Rossuma8d0f4f1999-06-08 13:20:05 +0000301
Tim Peters146965a2001-01-14 18:09:23 +0000302 def is_empty (self):
Armin Rigob562bc62004-09-27 17:49:00 +0000303 return not self.list
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000304
Tim Peters146965a2001-01-14 18:09:23 +0000305 def first (self):
Raymond Hettinger0a4977c2004-03-01 23:16:22 +0000306 return self.list[0]
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000307
Tim Peters146965a2001-01-14 18:09:23 +0000308 def push (self, data):
Raymond Hettingerac093c62004-02-07 03:19:10 +0000309 self.list.append(data)
Tim Peters146965a2001-01-14 18:09:23 +0000310
311 def pop (self):
312 if self.list:
Raymond Hettingerac093c62004-02-07 03:19:10 +0000313 return (1, self.list.popleft())
Tim Peters146965a2001-01-14 18:09:23 +0000314 else:
315 return (0, None)
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000316
317# Given 'haystack', see if any prefix of 'needle' is at its end. This
318# assumes an exact match has already been checked. Return the number of
319# characters matched.
320# for example:
321# f_p_a_e ("qwerty\r", "\r\n") => 1
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000322# f_p_a_e ("qwertydkjf", "\r\n") => 0
Andrew M. Kuchlingc63a3962002-03-20 02:22:58 +0000323# f_p_a_e ("qwerty\r\n", "\r\n") => <undefined>
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000324
325# this could maybe be made faster with a computed regex?
Andrew M. Kuchlingd305f512001-01-24 21:10:55 +0000326# [answer: no; circa Python-2.0, Jan 2001]
Andrew M. Kuchlingc63a3962002-03-20 02:22:58 +0000327# new python: 28961/s
328# old python: 18307/s
Andrew M. Kuchlingd305f512001-01-24 21:10:55 +0000329# re: 12820/s
330# regex: 14035/s
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000331
332def find_prefix_at_end (haystack, needle):
Tim Peters863ac442002-04-16 01:38:40 +0000333 l = len(needle) - 1
334 while l and not haystack.endswith(needle[:l]):
335 l -= 1
336 return l