summaryrefslogtreecommitdiffstats
path: root/Lib/asyncio/streams.py
blob: b53080ef5c854f66f6c6b71a8c4bb871083f49e3 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
"""Stream-related things."""

__all__ = ['StreamReader', 'StreamWriter', 'StreamReaderProtocol',
           'open_connection', 'start_server',
           ]

import collections

from . import events
from . import futures
from . import protocols
from . import tasks


_DEFAULT_LIMIT = 2**16


@tasks.coroutine
def open_connection(host=None, port=None, *,
                    loop=None, limit=_DEFAULT_LIMIT, **kwds):
    """A wrapper for create_connection() returning a (reader, writer) pair.

    The reader returned is a StreamReader instance; the writer is a
    StreamWriter instance.

    The arguments are all the usual arguments to create_connection()
    except protocol_factory; most common are positional host and port,
    with various optional keyword arguments following.

    Additional optional keyword arguments are loop (to set the event loop
    instance to use) and limit (to set the buffer limit passed to the
    StreamReader).

    (If you want to customize the StreamReader and/or
    StreamReaderProtocol classes, just copy the code -- there's
    really nothing special here except some convenience.)
    """
    if loop is None:
        loop = events.get_event_loop()
    reader = StreamReader(limit=limit, loop=loop)
    protocol = StreamReaderProtocol(reader, loop=loop)
    transport, _ = yield from loop.create_connection(
        lambda: protocol, host, port, **kwds)
    writer = StreamWriter(transport, protocol, reader, loop)
    return reader, writer


@tasks.coroutine
def start_server(client_connected_cb, host=None, port=None, *,
                 loop=None, limit=_DEFAULT_LIMIT, **kwds):
    """Start a socket server, call back for each client connected.

    The first parameter, `client_connected_cb`, takes two parameters:
    client_reader, client_writer.  client_reader is a StreamReader
    object, while client_writer is a StreamWriter object.  This
    parameter can either be a plain callback function or a coroutine;
    if it is a coroutine, it will be automatically converted into a
    Task.

    The rest of the arguments are all the usual arguments to
    loop.create_server() except protocol_factory; most common are
    positional host and port, with various optional keyword arguments
    following.  The return value is the same as loop.create_server().

    Additional optional keyword arguments are loop (to set the event loop
    instance to use) and limit (to set the buffer limit passed to the
    StreamReader).

    The return value is the same as loop.create_server(), i.e. a
    Server object which can be used to stop the service.
    """
    if loop is None:
        loop = events.get_event_loop()

    def factory():
        reader = StreamReader(limit=limit, loop=loop)
        protocol = StreamReaderProtocol(reader, client_connected_cb,
                                        loop=loop)
        return protocol

    return (yield from loop.create_server(factory, host, port, **kwds))


class StreamReaderProtocol(protocols.Protocol):
    """Trivial helper class to adapt between Protocol and StreamReader.

    (This is a helper class instead of making StreamReader itself a
    Protocol subclass, because the StreamReader has other potential
    uses, and to prevent the user of the StreamReader to accidentally
    call inappropriate methods of the protocol.)
    """

    def __init__(self, stream_reader, client_connected_cb=None, loop=None):
        self._stream_reader = stream_reader
        self._stream_writer = None
        self._drain_waiter = None
        self._paused = False
        self._client_connected_cb = client_connected_cb
        self._loop = loop  # May be None; we may never need it.

    def connection_made(self, transport):
        self._stream_reader.set_transport(transport)
        if self._client_connected_cb is not None:
            self._stream_writer = StreamWriter(transport, self,
                                               self._stream_reader,
                                               self._loop)
            res = self._client_connected_cb(self._stream_reader,
                                            self._stream_writer)
            if tasks.iscoroutine(res):
                tasks.Task(res, loop=self._loop)

    def connection_lost(self, exc):
        if exc is None:
            self._stream_reader.feed_eof()
        else:
            self._stream_reader.set_exception(exc)
        # Also wake up the writing side.
        if self._paused:
            waiter = self._drain_waiter
            if waiter is not None:
                self._drain_waiter = None
                if not waiter.done():
                    if exc is None:
                        waiter.set_result(None)
                    else:
                        waiter.set_exception(exc)

    def data_received(self, data):
        self._stream_reader.feed_data(data)

    def eof_received(self):
        self._stream_reader.feed_eof()

    def pause_writing(self):
        assert not self._paused
        self._paused = True

    def resume_writing(self):
        assert self._paused
        self._paused = False
        waiter = self._drain_waiter
        if waiter is not None:
            self._drain_waiter = None
            if not waiter.done():
                waiter.set_result(None)


class StreamWriter:
    """Wraps a Transport.

    This exposes write(), writelines(), [can_]write_eof(),
    get_extra_info() and close().  It adds drain() which returns an
    optional Future on which you can wait for flow control.  It also
    adds a transport property which references the Transport
    directly.
    """

    def __init__(self, transport, protocol, reader, loop):
        self._transport = transport
        self._protocol = protocol
        self._reader = reader
        self._loop = loop

    @property
    def transport(self):
        return self._transport

    def write(self, data):
        self._transport.write(data)

    def writelines(self, data):
        self._transport.writelines(data)

    def write_eof(self):
        return self._transport.write_eof()

    def can_write_eof(self):
        return self._transport.can_write_eof()

    def close(self):
        return self._transport.close()

    def get_extra_info(self, name, default=None):
        return self._transport.get_extra_info(name, default)

    def drain(self):
        """This method has an unusual return value.

        The intended use is to write

          w.write(data)
          yield from w.drain()

        When there's nothing to wait for, drain() returns (), and the
        yield-from continues immediately.  When the transport buffer
        is full (the protocol is paused), drain() creates and returns
        a Future and the yield-from will block until that Future is
        completed, which will happen when the buffer is (partially)
        drained and the protocol is resumed.
        """
        if self._reader._exception is not None:
            raise self._reader._exception
        if self._transport._conn_lost:  # Uses private variable.
            raise ConnectionResetError('Connection lost')
        if not self._protocol._paused:
            return ()
        waiter = self._protocol._drain_waiter
        assert waiter is None or waiter.cancelled()
        waiter = futures.Future(loop=self._loop)
        self._protocol._drain_waiter = waiter
        return waiter


class StreamReader:

    def __init__(self, limit=_DEFAULT_LIMIT, loop=None):
        # The line length limit is  a security feature;
        # it also doubles as half the buffer limit.
        self._limit = limit
        if loop is None:
            loop = events.get_event_loop()
        self._loop = loop
        # TODO: Use a bytearray for a buffer, like the transport.
        self._buffer = collections.deque()  # Deque of bytes objects.
        self._byte_count = 0  # Bytes in buffer.
        self._eof = False  # Whether we're done.
        self._waiter = None  # A future.
        self._exception = None
        self._transport = None
        self._paused = False

    def exception(self):
        return self._exception

    def set_exception(self, exc):
        self._exception = exc

        waiter = self._waiter
        if waiter is not None:
            self._waiter = None
            if not waiter.cancelled():
                waiter.set_exception(exc)

    def set_transport(self, transport):
        assert self._transport is None, 'Transport already set'
        self._transport = transport

    def _maybe_resume_transport(self):
        if self._paused and self._byte_count <= self._limit:
            self._paused = False
            self._transport.resume_reading()

    def feed_eof(self):
        self._eof = True
        waiter = self._waiter
        if waiter is not None:
            self._waiter = None
            if not waiter.cancelled():
                waiter.set_result(True)

    def feed_data(self, data):
        if not data:
            return

        self._buffer.append(data)
        self._byte_count += len(data)

        waiter = self._waiter
        if waiter is not None:
            self._waiter = None
            if not waiter.cancelled():
                waiter.set_result(False)

        if (self._transport is not None and
            not self._paused and
            self._byte_count > 2*self._limit):
            try:
                self._transport.pause_reading()
            except NotImplementedError:
                # The transport can't be paused.
                # We'll just have to buffer all data.
                # Forget the transport so we don't keep trying.
                self._transport = None
            else:
                self._paused = True

    def _create_waiter(self, func_name):
        # StreamReader uses a future to link the protocol feed_data() method
        # to a read coroutine. Running two read coroutines at the same time
        # would have an unexpected behaviour. It would not possible to know
        # which coroutine would get the next data.
        if self._waiter is not None:
            raise RuntimeError('%s() called while another coroutine is '
                               'already waiting for incoming data' % func_name)
        return futures.Future(loop=self._loop)

    @tasks.coroutine
    def readline(self):
        if self._exception is not None:
            raise self._exception

        parts = []
        parts_size = 0
        not_enough = True

        while not_enough:
            while self._buffer and not_enough:
                data = self._buffer.popleft()
                ichar = data.find(b'\n')
                if ichar < 0:
                    parts.append(data)
                    parts_size += len(data)
                else:
                    ichar += 1
                    head, tail = data[:ichar], data[ichar:]
                    if tail:
                        self._buffer.appendleft(tail)
                    not_enough = False
                    parts.append(head)
                    parts_size += len(head)

                if parts_size > self._limit:
                    self._byte_count -= parts_size
                    self._maybe_resume_transport()
                    raise ValueError('Line is too long')

            if self._eof:
                break

            if not_enough:
                self._waiter = self._create_waiter('readline')
                try:
                    yield from self._waiter
                finally:
                    self._waiter = None

        line = b''.join(parts)
        self._byte_count -= parts_size
        self._maybe_resume_transport()

        return line

    @tasks.coroutine
    def read(self, n=-1):
        if self._exception is not None:
            raise self._exception

        if not n:
            return b''

        if n < 0:
            while not self._eof:
                self._waiter = self._create_waiter('read')
                try:
                    yield from self._waiter
                finally:
                    self._waiter = None
        else:
            if not self._byte_count and not self._eof:
                self._waiter = self._create_waiter('read')
                try:
                    yield from self._waiter
                finally:
                    self._waiter = None

        if n < 0 or self._byte_count <= n:
            data = b''.join(self._buffer)
            self._buffer.clear()
            self._byte_count = 0
            self._maybe_resume_transport()
            return data

        parts = []
        parts_bytes = 0
        while self._buffer and parts_bytes < n:
            data = self._buffer.popleft()
            data_bytes = len(data)
            if n < parts_bytes + data_bytes:
                data_bytes = n - parts_bytes
                data, rest = data[:data_bytes], data[data_bytes:]
                self._buffer.appendleft(rest)

            parts.append(data)
            parts_bytes += data_bytes
            self._byte_count -= data_bytes
            self._maybe_resume_transport()

        return b''.join(parts)

    @tasks.coroutine
    def readexactly(self, n):
        if self._exception is not None:
            raise self._exception

        # There used to be "optimized" code here.  It created its own
        # Future and waited until self._buffer had at least the n
        # bytes, then called read(n).  Unfortunately, this could pause
        # the transport if the argument was larger than the pause
        # limit (which is twice self._limit).  So now we just read()
        # into a local buffer.

        blocks = []
        while n > 0:
            block = yield from self.read(n)
            if not block:
                break
            blocks.append(block)
            n -= len(block)

        # TODO: Raise EOFError if we break before n == 0?  (That would
        # be a change in specification, but I've always had to add an
        # explicit size check to the caller.)

        return b''.join(blocks)