blob: f728d1b47053ea894f30592b58ae6215583dad16 [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 asyncore
Raymond Hettingerac093c62004-02-07 03:19:10 +000049from collections import deque
Guido van Rossum0039d7b1999-01-12 20:19:27 +000050
Josiah Carlsond74900e2008-07-07 04:15:08 +000051
Victor Stinnerfd5d1b52014-07-08 00:16:54 +020052class async_chat(asyncore.dispatcher):
Tim Peters146965a2001-01-14 18:09:23 +000053 """This is an abstract class. You must derive from this class, and add
54 the two methods collect_incoming_data() and found_terminator()"""
Guido van Rossum0039d7b1999-01-12 20:19:27 +000055
Tim Peters146965a2001-01-14 18:09:23 +000056 # these are overridable defaults
Guido van Rossum0039d7b1999-01-12 20:19:27 +000057
Victor Stinnerfd5d1b52014-07-08 00:16:54 +020058 ac_in_buffer_size = 65536
59 ac_out_buffer_size = 65536
Guido van Rossum0039d7b1999-01-12 20:19:27 +000060
Josiah Carlsond74900e2008-07-07 04:15:08 +000061 # we don't want to enable the use of encoding by default, because that is a
62 # sign of an application bug that we don't want to pass silently
63
Victor Stinnerfd5d1b52014-07-08 00:16:54 +020064 use_encoding = 0
65 encoding = 'latin-1'
Josiah Carlsond74900e2008-07-07 04:15:08 +000066
Victor Stinnerfd5d1b52014-07-08 00:16:54 +020067 def __init__(self, sock=None, map=None):
Josiah Carlsond74900e2008-07-07 04:15:08 +000068 # for string terminator matching
Guido van Rossum076da092007-07-12 07:58:54 +000069 self.ac_in_buffer = b''
Josiah Carlsond74900e2008-07-07 04:15:08 +000070
Serhiy Storchaka50254c52013-08-29 11:35:43 +030071 # we use a list here rather than io.BytesIO for a few reasons...
72 # del lst[:] is faster than bio.truncate(0)
73 # lst = [] is faster than bio.truncate(0)
Josiah Carlsond74900e2008-07-07 04:15:08 +000074 self.incoming = []
75
76 # we toss the use of the "simple producer" and replace it with
77 # a pure deque, which the original fifo was a wrapping of
78 self.producer_fifo = deque()
Victor Stinnerfd5d1b52014-07-08 00:16:54 +020079 asyncore.dispatcher.__init__(self, sock, map)
Guido van Rossum0039d7b1999-01-12 20:19:27 +000080
Andrew M. Kuchling7dd5f3c2002-03-08 18:27:11 +000081 def collect_incoming_data(self, data):
Collin Winterce36ad82007-08-30 01:19:48 +000082 raise NotImplementedError("must be implemented in subclass")
Tim Peters863ac442002-04-16 01:38:40 +000083
Josiah Carlsond74900e2008-07-07 04:15:08 +000084 def _collect_incoming_data(self, data):
85 self.incoming.append(data)
86
87 def _get_data(self):
88 d = b''.join(self.incoming)
89 del self.incoming[:]
90 return d
91
Andrew M. Kuchling7dd5f3c2002-03-08 18:27:11 +000092 def found_terminator(self):
Collin Winterce36ad82007-08-30 01:19:48 +000093 raise NotImplementedError("must be implemented in subclass")
Tim Peters863ac442002-04-16 01:38:40 +000094
Victor Stinnerfd5d1b52014-07-08 00:16:54 +020095 def set_terminator(self, term):
96 """Set the input delimiter.
97
98 Can be a fixed string of any length, an integer, or None.
99 """
Josiah Carlsond74900e2008-07-07 04:15:08 +0000100 if isinstance(term, str) and self.use_encoding:
101 term = bytes(term, self.encoding)
Victor Stinner630a4f62014-07-08 00:26:36 +0200102 elif isinstance(term, int) and term < 0:
103 raise ValueError('the number of received bytes must be positive')
Tim Peters146965a2001-01-14 18:09:23 +0000104 self.terminator = term
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000105
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200106 def get_terminator(self):
Tim Peters146965a2001-01-14 18:09:23 +0000107 return self.terminator
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000108
Tim Peters146965a2001-01-14 18:09:23 +0000109 # grab some more data from the socket,
110 # throw it to the collector method,
111 # check for the terminator,
112 # if found, transition to the next state.
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000113
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200114 def handle_read(self):
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000115
Tim Peters146965a2001-01-14 18:09:23 +0000116 try:
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200117 data = self.recv(self.ac_in_buffer_size)
Victor Stinner45cff662014-07-24 18:49:36 +0200118 except BlockingIOError:
119 return
Andrew Svetlov0832af62012-12-18 23:10:48 +0200120 except OSError as why:
Tim Peters146965a2001-01-14 18:09:23 +0000121 self.handle_error()
122 return
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000123
Josiah Carlsond74900e2008-07-07 04:15:08 +0000124 if isinstance(data, str) and self.use_encoding:
125 data = bytes(str, self.encoding)
126 self.ac_in_buffer = self.ac_in_buffer + data
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000127
Tim Peters146965a2001-01-14 18:09:23 +0000128 # Continue to search for self.terminator in self.ac_in_buffer,
129 # while calling self.collect_incoming_data. The while loop
130 # is necessary because we might read several data+terminator
Josiah Carlsond74900e2008-07-07 04:15:08 +0000131 # combos with a single recv(4096).
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000132
Tim Peters146965a2001-01-14 18:09:23 +0000133 while self.ac_in_buffer:
134 lb = len(self.ac_in_buffer)
135 terminator = self.get_terminator()
Andrew M. Kuchlingca69f022005-06-09 14:59:45 +0000136 if not terminator:
Tim Peters146965a2001-01-14 18:09:23 +0000137 # no terminator, collect it all
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200138 self.collect_incoming_data(self.ac_in_buffer)
Guido van Rossum806c2462007-08-06 23:33:07 +0000139 self.ac_in_buffer = b''
Josiah Carlsond74900e2008-07-07 04:15:08 +0000140 elif isinstance(terminator, int):
Tim Peters146965a2001-01-14 18:09:23 +0000141 # numeric terminator
142 n = terminator
143 if lb < n:
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200144 self.collect_incoming_data(self.ac_in_buffer)
Guido van Rossum076da092007-07-12 07:58:54 +0000145 self.ac_in_buffer = b''
Tim Peters146965a2001-01-14 18:09:23 +0000146 self.terminator = self.terminator - lb
147 else:
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200148 self.collect_incoming_data(self.ac_in_buffer[:n])
Tim Peters146965a2001-01-14 18:09:23 +0000149 self.ac_in_buffer = self.ac_in_buffer[n:]
150 self.terminator = 0
151 self.found_terminator()
152 else:
153 # 3 cases:
154 # 1) end of buffer matches terminator exactly:
155 # collect data, transition
156 # 2) end of buffer matches some prefix:
157 # collect data to the prefix
158 # 3) end of buffer does not match any prefix:
159 # collect data
160 terminator_len = len(terminator)
Tim Petersb5d13922001-04-05 22:38:32 +0000161 index = self.ac_in_buffer.find(terminator)
Tim Peters146965a2001-01-14 18:09:23 +0000162 if index != -1:
163 # we found the terminator
164 if index > 0:
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200165 # don't bother reporting the empty string
166 # (source of subtle bugs)
167 self.collect_incoming_data(self.ac_in_buffer[:index])
Tim Peters146965a2001-01-14 18:09:23 +0000168 self.ac_in_buffer = self.ac_in_buffer[index+terminator_len:]
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200169 # This does the Right Thing if the terminator
170 # is changed here.
Tim Peters146965a2001-01-14 18:09:23 +0000171 self.found_terminator()
172 else:
173 # check for a prefix of the terminator
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200174 index = find_prefix_at_end(self.ac_in_buffer, terminator)
Tim Peters146965a2001-01-14 18:09:23 +0000175 if index:
176 if index != lb:
177 # we found a prefix, collect up to the prefix
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200178 self.collect_incoming_data(self.ac_in_buffer[:-index])
Tim Peters146965a2001-01-14 18:09:23 +0000179 self.ac_in_buffer = self.ac_in_buffer[-index:]
180 break
181 else:
182 # no prefix, collect it all
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200183 self.collect_incoming_data(self.ac_in_buffer)
Guido van Rossum076da092007-07-12 07:58:54 +0000184 self.ac_in_buffer = b''
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000185
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200186 def handle_write(self):
Josiah Carlsond74900e2008-07-07 04:15:08 +0000187 self.initiate_send()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000188
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200189 def handle_close(self):
Tim Peters146965a2001-01-14 18:09:23 +0000190 self.close()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000191
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200192 def push(self, data):
Victor Stinnerd9e810a2014-07-08 00:00:30 +0200193 if not isinstance(data, (bytes, bytearray, memoryview)):
194 raise TypeError('data argument must be byte-ish (%r)',
195 type(data))
Josiah Carlsond74900e2008-07-07 04:15:08 +0000196 sabs = self.ac_out_buffer_size
197 if len(data) > sabs:
198 for i in range(0, len(data), sabs):
199 self.producer_fifo.append(data[i:i+sabs])
200 else:
201 self.producer_fifo.append(data)
Tim Peters146965a2001-01-14 18:09:23 +0000202 self.initiate_send()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000203
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200204 def push_with_producer(self, producer):
Josiah Carlsond74900e2008-07-07 04:15:08 +0000205 self.producer_fifo.append(producer)
Tim Peters146965a2001-01-14 18:09:23 +0000206 self.initiate_send()
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000207
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200208 def readable(self):
Tim Peters146965a2001-01-14 18:09:23 +0000209 "predicate for inclusion in the readable for select()"
Josiah Carlsond74900e2008-07-07 04:15:08 +0000210 # cannot use the old predicate, it violates the claim of the
211 # set_terminator method.
212
213 # return (len(self.ac_in_buffer) <= self.ac_in_buffer_size)
214 return 1
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000215
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200216 def writable(self):
Tim Peters146965a2001-01-14 18:09:23 +0000217 "predicate for inclusion in the writable for select()"
Josiah Carlsond74900e2008-07-07 04:15:08 +0000218 return self.producer_fifo or (not self.connected)
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000219
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200220 def close_when_done(self):
Tim Peters146965a2001-01-14 18:09:23 +0000221 "automatically close this channel once the outgoing queue is empty"
Josiah Carlsond74900e2008-07-07 04:15:08 +0000222 self.producer_fifo.append(None)
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000223
Josiah Carlsond74900e2008-07-07 04:15:08 +0000224 def initiate_send(self):
225 while self.producer_fifo and self.connected:
226 first = self.producer_fifo[0]
227 # handle empty string/buffer or None entry
228 if not first:
229 del self.producer_fifo[0]
230 if first is None:
Josiah Carlsond74900e2008-07-07 04:15:08 +0000231 self.handle_close()
Tim Peters146965a2001-01-14 18:09:23 +0000232 return
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000233
Josiah Carlsond74900e2008-07-07 04:15:08 +0000234 # handle classic producer behavior
235 obs = self.ac_out_buffer_size
Tim Peters146965a2001-01-14 18:09:23 +0000236 try:
Giampaolo Rodola'd9f38bc2012-08-04 14:38:16 +0200237 data = first[:obs]
Josiah Carlsond74900e2008-07-07 04:15:08 +0000238 except TypeError:
239 data = first.more()
240 if data:
241 self.producer_fifo.appendleft(data)
242 else:
243 del self.producer_fifo[0]
244 continue
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000245
Josiah Carlsond74900e2008-07-07 04:15:08 +0000246 if isinstance(data, str) and self.use_encoding:
247 data = bytes(data, self.encoding)
248
249 # send the data
250 try:
251 num_sent = self.send(data)
Andrew Svetlov0832af62012-12-18 23:10:48 +0200252 except OSError:
Tim Peters146965a2001-01-14 18:09:23 +0000253 self.handle_error()
254 return
255
Josiah Carlsond74900e2008-07-07 04:15:08 +0000256 if num_sent:
257 if num_sent < len(data) or obs < len(first):
258 self.producer_fifo[0] = first[num_sent:]
259 else:
260 del self.producer_fifo[0]
261 # we tried to send some actual data
262 return
263
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200264 def discard_buffers(self):
Tim Peters146965a2001-01-14 18:09:23 +0000265 # Emergencies only!
Guido van Rossum076da092007-07-12 07:58:54 +0000266 self.ac_in_buffer = b''
Josiah Carlsond74900e2008-07-07 04:15:08 +0000267 del self.incoming[:]
268 self.producer_fifo.clear()
Andrew M. Kuchlingda85a272000-09-08 20:30:39 +0000269
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200270
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000271class simple_producer:
Guido van Rossuma8d0f4f1999-06-08 13:20:05 +0000272
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200273 def __init__(self, data, buffer_size=512):
Tim Peters146965a2001-01-14 18:09:23 +0000274 self.data = data
275 self.buffer_size = buffer_size
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000276
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200277 def more(self):
278 if len(self.data) > self.buffer_size:
Tim Peters146965a2001-01-14 18:09:23 +0000279 result = self.data[:self.buffer_size]
280 self.data = self.data[self.buffer_size:]
281 return result
282 else:
283 result = self.data
Guido van Rossum076da092007-07-12 07:58:54 +0000284 self.data = b''
Tim Peters146965a2001-01-14 18:09:23 +0000285 return result
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000286
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200287
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000288class fifo:
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200289 def __init__(self, list=None):
Giampaolo Rodola'bd048762014-06-21 13:58:30 +0200290 import warnings
291 warnings.warn('fifo class will be removed in Python 3.6',
292 DeprecationWarning, stacklevel=2)
Tim Peters146965a2001-01-14 18:09:23 +0000293 if not list:
Raymond Hettingerac093c62004-02-07 03:19:10 +0000294 self.list = deque()
Tim Peters146965a2001-01-14 18:09:23 +0000295 else:
Raymond Hettingerac093c62004-02-07 03:19:10 +0000296 self.list = deque(list)
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000297
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200298 def __len__(self):
Tim Peters146965a2001-01-14 18:09:23 +0000299 return len(self.list)
Guido van Rossuma8d0f4f1999-06-08 13:20:05 +0000300
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200301 def is_empty(self):
Armin Rigob562bc62004-09-27 17:49:00 +0000302 return not self.list
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000303
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200304 def first(self):
Raymond Hettinger0a4977c2004-03-01 23:16:22 +0000305 return self.list[0]
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000306
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200307 def push(self, data):
Raymond Hettingerac093c62004-02-07 03:19:10 +0000308 self.list.append(data)
Tim Peters146965a2001-01-14 18:09:23 +0000309
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200310 def pop(self):
Tim Peters146965a2001-01-14 18:09:23 +0000311 if self.list:
Raymond Hettingerac093c62004-02-07 03:19:10 +0000312 return (1, self.list.popleft())
Tim Peters146965a2001-01-14 18:09:23 +0000313 else:
314 return (0, None)
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000315
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200316
Guido van Rossum0039d7b1999-01-12 20:19:27 +0000317# 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:
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200321# f_p_a_e("qwerty\r", "\r\n") => 1
322# f_p_a_e("qwertydkjf", "\r\n") => 0
323# 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
Victor Stinnerfd5d1b52014-07-08 00:16:54 +0200332def 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