blob: 6f99ba1063c85e48dae1c7f9a15965edcec4db61 [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"""
48
Guido van Rossum0039d7b1999-01-12 20:19:27 +000049import socket
50import asyncore
Raymond Hettingerac093c62004-02-07 03:19:10 +000051from collections import deque
Guido van Rossum0039d7b1999-01-12 20:19:27 +000052
Guido van Rossum0039d7b1999-01-12 20:19:27 +000053class async_chat (asyncore.dispatcher):
Tim Peters146965a2001-01-14 18:09:23 +000054 """This is an abstract class. You must derive from this class, and add
55 the two methods collect_incoming_data() and found_terminator()"""
Guido van Rossum0039d7b1999-01-12 20:19:27 +000056
Tim Peters146965a2001-01-14 18:09:23 +000057 # these are overridable defaults
Guido van Rossum0039d7b1999-01-12 20:19:27 +000058
Tim Peters146965a2001-01-14 18:09:23 +000059 ac_in_buffer_size = 4096
60 ac_out_buffer_size = 4096
Guido van Rossum0039d7b1999-01-12 20:19:27 +000061
Tim Peters146965a2001-01-14 18:09:23 +000062 def __init__ (self, conn=None):
63 self.ac_in_buffer = ''
64 self.ac_out_buffer = ''
65 self.producer_fifo = fifo()
66 asyncore.dispatcher.__init__ (self, conn)
Guido van Rossum0039d7b1999-01-12 20:19:27 +000067
Andrew M. Kuchling7dd5f3c2002-03-08 18:27:11 +000068 def collect_incoming_data(self, data):
69 raise NotImplementedError, "must be implemented in subclass"
Tim Peters863ac442002-04-16 01:38:40 +000070
Andrew M. Kuchling7dd5f3c2002-03-08 18:27:11 +000071 def found_terminator(self):
72 raise NotImplementedError, "must be implemented in subclass"
Tim Peters863ac442002-04-16 01:38:40 +000073
Tim Peters146965a2001-01-14 18:09:23 +000074 def set_terminator (self, term):
75 "Set the input delimiter. Can be a fixed string of any length, an integer, or None"
76 self.terminator = term
Guido van Rossum0039d7b1999-01-12 20:19:27 +000077
Tim Peters146965a2001-01-14 18:09:23 +000078 def get_terminator (self):
79 return self.terminator
Guido van Rossum0039d7b1999-01-12 20:19:27 +000080
Tim Peters146965a2001-01-14 18:09:23 +000081 # grab some more data from the socket,
82 # throw it to the collector method,
83 # check for the terminator,
84 # if found, transition to the next state.
Guido van Rossum0039d7b1999-01-12 20:19:27 +000085
Tim Peters146965a2001-01-14 18:09:23 +000086 def handle_read (self):
Guido van Rossum0039d7b1999-01-12 20:19:27 +000087
Tim Peters146965a2001-01-14 18:09:23 +000088 try:
89 data = self.recv (self.ac_in_buffer_size)
90 except socket.error, why:
91 self.handle_error()
92 return
Guido van Rossum0039d7b1999-01-12 20:19:27 +000093
Tim Peters146965a2001-01-14 18:09:23 +000094 self.ac_in_buffer = self.ac_in_buffer + data
Guido van Rossum0039d7b1999-01-12 20:19:27 +000095
Tim Peters146965a2001-01-14 18:09:23 +000096 # Continue to search for self.terminator in self.ac_in_buffer,
97 # while calling self.collect_incoming_data. The while loop
98 # is necessary because we might read several data+terminator
99 # combos with a single recv(1024).
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000100
Tim Peters146965a2001-01-14 18:09:23 +0000101 while self.ac_in_buffer:
102 lb = len(self.ac_in_buffer)
103 terminator = self.get_terminator()
Andrew M. Kuchlingca69f022005-06-09 14:59:45 +0000104 if not terminator:
Tim Peters146965a2001-01-14 18:09:23 +0000105 # no terminator, collect it all
106 self.collect_incoming_data (self.ac_in_buffer)
107 self.ac_in_buffer = ''
Andrew M. Kuchlingca69f022005-06-09 14:59:45 +0000108 elif isinstance(terminator, int) or isinstance(terminator, long):
Tim Peters146965a2001-01-14 18:09:23 +0000109 # numeric terminator
110 n = terminator
111 if lb < n:
112 self.collect_incoming_data (self.ac_in_buffer)
113 self.ac_in_buffer = ''
114 self.terminator = self.terminator - lb
115 else:
116 self.collect_incoming_data (self.ac_in_buffer[:n])
117 self.ac_in_buffer = self.ac_in_buffer[n:]
118 self.terminator = 0
119 self.found_terminator()
120 else:
121 # 3 cases:
122 # 1) end of buffer matches terminator exactly:
123 # collect data, transition
124 # 2) end of buffer matches some prefix:
125 # collect data to the prefix
126 # 3) end of buffer does not match any prefix:
127 # collect data
128 terminator_len = len(terminator)
Tim Petersb5d13922001-04-05 22:38:32 +0000129 index = self.ac_in_buffer.find(terminator)
Tim Peters146965a2001-01-14 18:09:23 +0000130 if index != -1:
131 # we found the terminator
132 if index > 0:
133 # don't bother reporting the empty string (source of subtle bugs)
134 self.collect_incoming_data (self.ac_in_buffer[:index])
135 self.ac_in_buffer = self.ac_in_buffer[index+terminator_len:]
136 # This does the Right Thing if the terminator is changed here.
137 self.found_terminator()
138 else:
139 # check for a prefix of the terminator
140 index = find_prefix_at_end (self.ac_in_buffer, terminator)
141 if index:
142 if index != lb:
143 # we found a prefix, collect up to the prefix
144 self.collect_incoming_data (self.ac_in_buffer[:-index])
145 self.ac_in_buffer = self.ac_in_buffer[-index:]
146 break
147 else:
148 # no prefix, collect it all
149 self.collect_incoming_data (self.ac_in_buffer)
150 self.ac_in_buffer = ''
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000151
Tim Peters146965a2001-01-14 18:09:23 +0000152 def handle_write (self):
153 self.initiate_send ()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000154
Tim Peters146965a2001-01-14 18:09:23 +0000155 def handle_close (self):
156 self.close()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000157
Tim Peters146965a2001-01-14 18:09:23 +0000158 def push (self, data):
159 self.producer_fifo.push (simple_producer (data))
160 self.initiate_send()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000161
Tim Peters146965a2001-01-14 18:09:23 +0000162 def push_with_producer (self, producer):
163 self.producer_fifo.push (producer)
164 self.initiate_send()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000165
Tim Peters146965a2001-01-14 18:09:23 +0000166 def readable (self):
167 "predicate for inclusion in the readable for select()"
168 return (len(self.ac_in_buffer) <= self.ac_in_buffer_size)
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000169
Tim Peters146965a2001-01-14 18:09:23 +0000170 def writable (self):
171 "predicate for inclusion in the writable for select()"
172 # return len(self.ac_out_buffer) or len(self.producer_fifo) or (not self.connected)
173 # this is about twice as fast, though not as clear.
174 return not (
Tim Peters6fd71202001-04-08 07:23:44 +0000175 (self.ac_out_buffer == '') and
Tim Peters146965a2001-01-14 18:09:23 +0000176 self.producer_fifo.is_empty() and
177 self.connected
178 )
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000179
Tim Peters146965a2001-01-14 18:09:23 +0000180 def close_when_done (self):
181 "automatically close this channel once the outgoing queue is empty"
182 self.producer_fifo.push (None)
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000183
Tim Peters146965a2001-01-14 18:09:23 +0000184 # refill the outgoing buffer by calling the more() method
185 # of the first producer in the queue
186 def refill_buffer (self):
Tim Peters146965a2001-01-14 18:09:23 +0000187 while 1:
188 if len(self.producer_fifo):
189 p = self.producer_fifo.first()
190 # a 'None' in the producer fifo is a sentinel,
191 # telling us to close the channel.
192 if p is None:
193 if not self.ac_out_buffer:
194 self.producer_fifo.pop()
195 self.close()
196 return
Andrew M. Kuchlingfaef74a2003-03-10 15:12:00 +0000197 elif isinstance(p, str):
Tim Peters146965a2001-01-14 18:09:23 +0000198 self.producer_fifo.pop()
199 self.ac_out_buffer = self.ac_out_buffer + p
200 return
201 data = p.more()
202 if data:
203 self.ac_out_buffer = self.ac_out_buffer + data
204 return
205 else:
206 self.producer_fifo.pop()
207 else:
208 return
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000209
Tim Peters146965a2001-01-14 18:09:23 +0000210 def initiate_send (self):
211 obs = self.ac_out_buffer_size
212 # try to refill the buffer
213 if (len (self.ac_out_buffer) < obs):
214 self.refill_buffer()
Guido van Rossuma8d0f4f1999-06-08 13:20:05 +0000215
Tim Peters146965a2001-01-14 18:09:23 +0000216 if self.ac_out_buffer and self.connected:
217 # try to send the buffer
218 try:
219 num_sent = self.send (self.ac_out_buffer[:obs])
220 if num_sent:
221 self.ac_out_buffer = self.ac_out_buffer[num_sent:]
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000222
Tim Peters146965a2001-01-14 18:09:23 +0000223 except socket.error, why:
224 self.handle_error()
225 return
226
227 def discard_buffers (self):
228 # Emergencies only!
229 self.ac_in_buffer = ''
230 self.ac_out_buffer = ''
231 while self.producer_fifo:
232 self.producer_fifo.pop()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000233
Andrew M. Kuchlingda85a272000-09-08 20:30:39 +0000234
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000235class simple_producer:
Guido van Rossuma8d0f4f1999-06-08 13:20:05 +0000236
Tim Peters146965a2001-01-14 18:09:23 +0000237 def __init__ (self, data, buffer_size=512):
238 self.data = data
239 self.buffer_size = buffer_size
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000240
Tim Peters146965a2001-01-14 18:09:23 +0000241 def more (self):
242 if len (self.data) > self.buffer_size:
243 result = self.data[:self.buffer_size]
244 self.data = self.data[self.buffer_size:]
245 return result
246 else:
247 result = self.data
248 self.data = ''
249 return result
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000250
251class fifo:
Tim Peters146965a2001-01-14 18:09:23 +0000252 def __init__ (self, list=None):
253 if not list:
Raymond Hettingerac093c62004-02-07 03:19:10 +0000254 self.list = deque()
Tim Peters146965a2001-01-14 18:09:23 +0000255 else:
Raymond Hettingerac093c62004-02-07 03:19:10 +0000256 self.list = deque(list)
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000257
Tim Peters146965a2001-01-14 18:09:23 +0000258 def __len__ (self):
259 return len(self.list)
Guido van Rossuma8d0f4f1999-06-08 13:20:05 +0000260
Tim Peters146965a2001-01-14 18:09:23 +0000261 def is_empty (self):
Armin Rigob562bc62004-09-27 17:49:00 +0000262 return not self.list
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000263
Tim Peters146965a2001-01-14 18:09:23 +0000264 def first (self):
Raymond Hettinger0a4977c2004-03-01 23:16:22 +0000265 return self.list[0]
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000266
Tim Peters146965a2001-01-14 18:09:23 +0000267 def push (self, data):
Raymond Hettingerac093c62004-02-07 03:19:10 +0000268 self.list.append(data)
Tim Peters146965a2001-01-14 18:09:23 +0000269
270 def pop (self):
271 if self.list:
Raymond Hettingerac093c62004-02-07 03:19:10 +0000272 return (1, self.list.popleft())
Tim Peters146965a2001-01-14 18:09:23 +0000273 else:
274 return (0, None)
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000275
276# Given 'haystack', see if any prefix of 'needle' is at its end. This
277# assumes an exact match has already been checked. Return the number of
278# characters matched.
279# for example:
280# f_p_a_e ("qwerty\r", "\r\n") => 1
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000281# f_p_a_e ("qwertydkjf", "\r\n") => 0
Andrew M. Kuchlingc63a3962002-03-20 02:22:58 +0000282# f_p_a_e ("qwerty\r\n", "\r\n") => <undefined>
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000283
284# this could maybe be made faster with a computed regex?
Andrew M. Kuchlingd305f512001-01-24 21:10:55 +0000285# [answer: no; circa Python-2.0, Jan 2001]
Andrew M. Kuchlingc63a3962002-03-20 02:22:58 +0000286# new python: 28961/s
287# old python: 18307/s
Andrew M. Kuchlingd305f512001-01-24 21:10:55 +0000288# re: 12820/s
289# regex: 14035/s
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000290
291def find_prefix_at_end (haystack, needle):
Tim Peters863ac442002-04-16 01:38:40 +0000292 l = len(needle) - 1
293 while l and not haystack.endswith(needle[:l]):
294 l -= 1
295 return l