This file is indexed.

/usr/lib/python2.7/dist-packages/dockerpty/io.py is in python-dockerpty 0.4.1-1.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

  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
# dockerpty: io.py
#
# Copyright 2014 Chris Corbyn <chris@w3style.co.uk>
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import os
import fcntl
import errno
import struct
import select as builtin_select
import six


def set_blocking(fd, blocking=True):
    """
    Set the given file-descriptor blocking or non-blocking.

    Returns the original blocking status.
    """

    old_flag = fcntl.fcntl(fd, fcntl.F_GETFL)

    if blocking:
        new_flag = old_flag & ~ os.O_NONBLOCK
    else:
        new_flag = old_flag | os.O_NONBLOCK

    fcntl.fcntl(fd, fcntl.F_SETFL, new_flag)

    return not bool(old_flag & os.O_NONBLOCK)


def select(read_streams, write_streams, timeout=0):
    """
    Select the streams from `read_streams` that are ready for reading, and
    streams from `write_streams` ready for writing.

    Uses `select.select()` internally but only returns two lists of ready streams.
    """

    exception_streams = []

    try:
        return builtin_select.select(
            read_streams,
            write_streams,
            exception_streams,
            timeout,
        )[0:2]
    except builtin_select.error as e:
        # POSIX signals interrupt select()
        no = e.errno if six.PY3 else e[0]
        if no == errno.EINTR:
            return ([], [])
        else:
            raise e


class Stream(object):
    """
    Generic Stream class.

    This is a file-like abstraction on top of os.read() and os.write(), which
    add consistency to the reading of sockets and files alike.
    """

    """
    Recoverable IO/OS Errors.
    """
    ERRNO_RECOVERABLE = [
        errno.EINTR,
        errno.EDEADLK,
        errno.EWOULDBLOCK,
    ]

    def __init__(self, fd):
        """
        Initialize the Stream for the file descriptor `fd`.

        The `fd` object must have a `fileno()` method.
        """
        self.fd = fd
        self.buffer = b''
        self.close_requested = False
        self.closed = False

    def fileno(self):
        """
        Return the fileno() of the file descriptor.
        """

        return self.fd.fileno()

    def set_blocking(self, value):
        if hasattr(self.fd, 'setblocking'):
            self.fd.setblocking(value)
            return True
        else:
            return set_blocking(self.fd, value)

    def read(self, n=4096):
        """
        Return `n` bytes of data from the Stream, or None at end of stream.
        """

        while True:
            try:
                if hasattr(self.fd, 'recv'):
                    return self.fd.recv(n)
                return os.read(self.fd.fileno(), n)
            except EnvironmentError as e:
                if e.errno not in Stream.ERRNO_RECOVERABLE:
                    raise e


    def write(self, data):
        """
        Write `data` to the Stream. Not all data may be written right away.
        Use select to find when the stream is writeable, and call do_write()
        to flush the internal buffer.
        """

        if not data:
            return None

        self.buffer += data
        self.do_write()

        return len(data)

    def do_write(self):
        """
        Flushes as much pending data from the internal write buffer as possible.
        """
        while True:
            try:
                written = 0

                if hasattr(self.fd, 'send'):
                    written = self.fd.send(self.buffer)
                else:
                    written = os.write(self.fd.fileno(), self.buffer)

                self.buffer = self.buffer[written:]

                # try to close after writes if a close was requested
                if self.close_requested and len(self.buffer) == 0:
                    self.close()

                return written
            except EnvironmentError as e:
                if e.errno not in Stream.ERRNO_RECOVERABLE:
                    raise e

    def needs_write(self):
        """
        Returns True if the stream has data waiting to be written.
        """
        return len(self.buffer) > 0

    def close(self):
        self.close_requested = True

        # We don't close the fd immediately, as there may still be data pending
        # to write.
        if not self.closed and len(self.buffer) == 0:
            self.closed = True
            if hasattr(self.fd, 'close'):
                self.fd.close()
            else:
                os.close(self.fd.fileno())

    def __repr__(self):
        return "{cls}({fd})".format(cls=type(self).__name__, fd=self.fd)


class Demuxer(object):
    """
    Wraps a multiplexed Stream to read in data demultiplexed.

    Docker multiplexes streams together when there is no PTY attached, by
    sending an 8-byte header, followed by a chunk of data.

    The first 4 bytes of the header denote the stream from which the data came
    (i.e. 0x01 = stdout, 0x02 = stderr). Only the first byte of these initial 4
    bytes is used.

    The next 4 bytes indicate the length of the following chunk of data as an
    integer in big endian format. This much data must be consumed before the
    next 8-byte header is read.
    """

    def __init__(self, stream):
        """
        Initialize a new Demuxer reading from `stream`.
        """

        self.stream = stream
        self.remain = 0

    def fileno(self):
        """
        Returns the fileno() of the underlying Stream.

        This is useful for select() to work.
        """

        return self.stream.fileno()

    def set_blocking(self, value):
        return self.stream.set_blocking(value)

    def read(self, n=4096):
        """
        Read up to `n` bytes of data from the Stream, after demuxing.

        Less than `n` bytes of data may be returned depending on the available
        payload, but the number of bytes returned will never exceed `n`.

        Because demuxing involves scanning 8-byte headers, the actual amount of
        data read from the underlying stream may be greater than `n`.
        """

        size = self._next_packet_size(n)

        if size <= 0:
            return
        else:
            data = six.binary_type()
            while len(data) < size:
                nxt = self.stream.read(size - len(data))
                if not nxt:
                    # the stream has closed, return what data we got
                    return data
                data = data + nxt
            return data

    def write(self, data):
        """
        Delegates the the underlying Stream.
        """

        return self.stream.write(data)

    def needs_write(self):
        """
        Delegates to underlying Stream.
        """

        if hasattr(self.stream, 'needs_write'):
            return self.stream.needs_write()

        return False

    def do_write(self):
        """
        Delegates to underlying Stream.
        """

        if hasattr(self.stream, 'do_write'):
            return self.stream.do_write()

        return False

    def close(self):
        """
        Delegates to underlying Stream.
        """

        return self.stream.close()

    def _next_packet_size(self, n=0):
        size = 0

        if self.remain > 0:
            size = min(n, self.remain)
            self.remain -= size
        else:
            data = six.binary_type()
            while len(data) < 8:
                nxt = self.stream.read(8 - len(data))
                if not nxt:
                    # The stream has closed, there's nothing more to read
                    return 0
                data = data + nxt

            if data is None:
                return 0
            if len(data) == 8:
                __, actual = struct.unpack('>BxxxL', data)
                size = min(n, actual)
                self.remain = actual - size

        return size

    def __repr__(self):
        return "{cls}({stream})".format(cls=type(self).__name__,
                                        stream=self.stream)


class Pump(object):
    """
    Stream pump class.

    A Pump wraps two Streams, reading from one and and writing its data into
    the other, much like a pipe but manually managed.

    This abstraction is used to facilitate piping data between the file
    descriptors associated with the tty and those associated with a container's
    allocated pty.

    Pumps are selectable based on the 'read' end of the pipe.
    """

    def __init__(self,
                 from_stream,
                 to_stream,
                 wait_for_output=True,
                 propagate_close=True):
        """
        Initialize a Pump with a Stream to read from and another to write to.

        `wait_for_output` is a flag that says that we need to wait for EOF
        on the from_stream in order to consider this pump as "done".
        """

        self.from_stream = from_stream
        self.to_stream = to_stream
        self.eof = False
        self.wait_for_output = wait_for_output
        self.propagate_close = propagate_close

    def fileno(self):
        """
        Returns the `fileno()` of the reader end of the Pump.

        This is useful to allow Pumps to function with `select()`.
        """

        return self.from_stream.fileno()

    def set_blocking(self, value):
        return self.from_stream.set_blocking(value)

    def flush(self, n=4096):
        """
        Flush `n` bytes of data from the reader Stream to the writer Stream.

        Returns the number of bytes that were actually flushed. A return value
        of zero is not an error.

        If EOF has been reached, `None` is returned.
        """

        try:
            read = self.from_stream.read(n)

            if read is None or len(read) == 0:
                self.eof = True
                if self.propagate_close:
                    self.to_stream.close()
                return None

            return self.to_stream.write(read)
        except OSError as e:
            if e.errno != errno.EPIPE:
                raise e

    def is_done(self):
        """
        Returns True if the read stream is done (either it's returned EOF or
        the pump doesn't have wait_for_output set), and the write
        side does not have pending bytes to send.
        """

        return (not self.wait_for_output or self.eof) and \
                not (hasattr(self.to_stream, 'needs_write') and self.to_stream.needs_write())

    def __repr__(self):
        return "{cls}(from={from_stream}, to={to_stream})".format(
            cls=type(self).__name__,
            from_stream=self.from_stream,
            to_stream=self.to_stream)