/Doc/library/socket.rst
ReStructuredText | 1430 lines | 1004 code | 426 blank | 0 comment | 0 complexity | 97fb9dc4f9bc695ad7e38014fc9d7271 MD5 | raw file
Possible License(s): BSD-3-Clause, 0BSD
1:mod:`socket` --- Low-level networking interface 2================================================ 3 4.. module:: socket 5 :synopsis: Low-level networking interface. 6 7 8This module provides access to the BSD *socket* interface. It is available on 9all modern Unix systems, Windows, MacOS, OS/2, and probably additional 10platforms. 11 12.. note:: 13 14 Some behavior may be platform dependent, since calls are made to the operating 15 system socket APIs. 16 17.. index:: object: socket 18 19The Python interface is a straightforward transliteration of the Unix system 20call and library interface for sockets to Python's object-oriented style: the 21:func:`socket` function returns a :dfn:`socket object` whose methods implement 22the various socket system calls. Parameter types are somewhat higher-level than 23in the C interface: as with :meth:`read` and :meth:`write` operations on Python 24files, buffer allocation on receive operations is automatic, and buffer length 25is implicit on send operations. 26 27 28.. seealso:: 29 30 Module :mod:`socketserver` 31 Classes that simplify writing network servers. 32 33 Module :mod:`ssl` 34 A TLS/SSL wrapper for socket objects. 35 36 37Socket families 38--------------- 39 40Depending on the system and the build options, various socket families 41are supported by this module. 42 43The address format required by a particular socket object is automatically 44selected based on the address family specified when the socket object was 45created. Socket addresses are represented as follows: 46 47- The address of an :const:`AF_UNIX` socket bound to a file system node 48 is represented as a string, using the file system encoding and the 49 ``'surrogateescape'`` error handler (see :pep:`383`). An address in 50 Linux's abstract namespace is returned as a :class:`bytes` object with 51 an initial null byte; note that sockets in this namespace can 52 communicate with normal file system sockets, so programs intended to 53 run on Linux may need to deal with both types of address. A string or 54 :class:`bytes` object can be used for either type of address when 55 passing it as an argument. 56 57 .. versionchanged:: 3.3 58 Previously, :const:`AF_UNIX` socket paths were assumed to use UTF-8 59 encoding. 60 61- A pair ``(host, port)`` is used for the :const:`AF_INET` address family, 62 where *host* is a string representing either a hostname in Internet domain 63 notation like ``'daring.cwi.nl'`` or an IPv4 address like ``'100.50.200.5'``, 64 and *port* is an integer. 65 66- For :const:`AF_INET6` address family, a four-tuple ``(host, port, flowinfo, 67 scopeid)`` is used, where *flowinfo* and *scopeid* represent the ``sin6_flowinfo`` 68 and ``sin6_scope_id`` members in :const:`struct sockaddr_in6` in C. For 69 :mod:`socket` module methods, *flowinfo* and *scopeid* can be omitted just for 70 backward compatibility. Note, however, omission of *scopeid* can cause problems 71 in manipulating scoped IPv6 addresses. 72 73- :const:`AF_NETLINK` sockets are represented as pairs ``(pid, groups)``. 74 75- Linux-only support for TIPC is available using the :const:`AF_TIPC` 76 address family. TIPC is an open, non-IP based networked protocol designed 77 for use in clustered computer environments. Addresses are represented by a 78 tuple, and the fields depend on the address type. The general tuple form is 79 ``(addr_type, v1, v2, v3 [, scope])``, where: 80 81 - *addr_type* is one of :const:`TIPC_ADDR_NAMESEQ`, :const:`TIPC_ADDR_NAME`, 82 or :const:`TIPC_ADDR_ID`. 83 - *scope* is one of :const:`TIPC_ZONE_SCOPE`, :const:`TIPC_CLUSTER_SCOPE`, and 84 :const:`TIPC_NODE_SCOPE`. 85 - If *addr_type* is :const:`TIPC_ADDR_NAME`, then *v1* is the server type, *v2* is 86 the port identifier, and *v3* should be 0. 87 88 If *addr_type* is :const:`TIPC_ADDR_NAMESEQ`, then *v1* is the server type, *v2* 89 is the lower port number, and *v3* is the upper port number. 90 91 If *addr_type* is :const:`TIPC_ADDR_ID`, then *v1* is the node, *v2* is the 92 reference, and *v3* should be set to 0. 93 94 If *addr_type* is :const:`TIPC_ADDR_ID`, then *v1* is the node, *v2* is the 95 reference, and *v3* should be set to 0. 96 97- A tuple ``(interface, )`` is used for the :const:`AF_CAN` address family, 98 where *interface* is a string representing a network interface name like 99 ``'can0'``. The network interface name ``''`` can be used to receive packets 100 from all network interfaces of this family. 101 102- A string or a tuple ``(id, unit)`` is used for the :const:`SYSPROTO_CONTROL` 103 protocol of the :const:`PF_SYSTEM` family. The string is the name of a 104 kernel control using a dynamically-assigned ID. The tuple can be used if ID 105 and unit number of the kernel control are known or if a registered ID is 106 used. 107 108 .. versionadded:: 3.3 109 110- Certain other address families (:const:`AF_BLUETOOTH`, :const:`AF_PACKET`, 111 :const:`AF_CAN`) support specific representations. 112 113 .. XXX document them! 114 115For IPv4 addresses, two special forms are accepted instead of a host address: 116the empty string represents :const:`INADDR_ANY`, and the string 117``'<broadcast>'`` represents :const:`INADDR_BROADCAST`. This behavior is not 118compatible with IPv6, therefore, you may want to avoid these if you intend 119to support IPv6 with your Python programs. 120 121If you use a hostname in the *host* portion of IPv4/v6 socket address, the 122program may show a nondeterministic behavior, as Python uses the first address 123returned from the DNS resolution. The socket address will be resolved 124differently into an actual IPv4/v6 address, depending on the results from DNS 125resolution and/or the host configuration. For deterministic behavior use a 126numeric address in *host* portion. 127 128All errors raise exceptions. The normal exceptions for invalid argument types 129and out-of-memory conditions can be raised; starting from Python 3.3, errors 130related to socket or address semantics raise :exc:`OSError` or one of its 131subclasses (they used to raise :exc:`socket.error`). 132 133Non-blocking mode is supported through :meth:`~socket.setblocking`. A 134generalization of this based on timeouts is supported through 135:meth:`~socket.settimeout`. 136 137 138Module contents 139--------------- 140 141The module :mod:`socket` exports the following constants and functions: 142 143 144.. exception:: error 145 146 A deprecated alias of :exc:`OSError`. 147 148 .. versionchanged:: 3.3 149 Following :pep:`3151`, this class was made an alias of :exc:`OSError`. 150 151 152.. exception:: herror 153 154 A subclass of :exc:`OSError`, this exception is raised for 155 address-related errors, i.e. for functions that use *h_errno* in the POSIX 156 C API, including :func:`gethostbyname_ex` and :func:`gethostbyaddr`. 157 The accompanying value is a pair ``(h_errno, string)`` representing an 158 error returned by a library call. *h_errno* is a numeric value, while 159 *string* represents the description of *h_errno*, as returned by the 160 :c:func:`hstrerror` C function. 161 162 .. versionchanged:: 3.3 163 This class was made a subclass of :exc:`OSError`. 164 165.. exception:: gaierror 166 167 A subclass of :exc:`OSError`, this exception is raised for 168 address-related errors by :func:`getaddrinfo` and :func:`getnameinfo`. 169 The accompanying value is a pair ``(error, string)`` representing an error 170 returned by a library call. *string* represents the description of 171 *error*, as returned by the :c:func:`gai_strerror` C function. The 172 numeric *error* value will match one of the :const:`EAI_\*` constants 173 defined in this module. 174 175 .. versionchanged:: 3.3 176 This class was made a subclass of :exc:`OSError`. 177 178.. exception:: timeout 179 180 A subclass of :exc:`OSError`, this exception is raised when a timeout 181 occurs on a socket which has had timeouts enabled via a prior call to 182 :meth:`~socket.settimeout` (or implicitly through 183 :func:`~socket.setdefaulttimeout`). The accompanying value is a string 184 whose value is currently always "timed out". 185 186 .. versionchanged:: 3.3 187 This class was made a subclass of :exc:`OSError`. 188 189.. data:: AF_UNIX 190 AF_INET 191 AF_INET6 192 193 These constants represent the address (and protocol) families, used for the 194 first argument to :func:`socket`. If the :const:`AF_UNIX` constant is not 195 defined then this protocol is unsupported. More constants may be available 196 depending on the system. 197 198 199.. data:: SOCK_STREAM 200 SOCK_DGRAM 201 SOCK_RAW 202 SOCK_RDM 203 SOCK_SEQPACKET 204 205 These constants represent the socket types, used for the second argument to 206 :func:`socket`. More constants may be available depending on the system. 207 (Only :const:`SOCK_STREAM` and :const:`SOCK_DGRAM` appear to be generally 208 useful.) 209 210.. data:: SOCK_CLOEXEC 211 SOCK_NONBLOCK 212 213 These two constants, if defined, can be combined with the socket types and 214 allow you to set some flags atomically (thus avoiding possible race 215 conditions and the need for separate calls). 216 217 .. seealso:: 218 219 `Secure File Descriptor Handling <http://udrepper.livejournal.com/20407.html>`_ 220 for a more thorough explanation. 221 222 Availability: Linux >= 2.6.27. 223 224 .. versionadded:: 3.2 225 226.. data:: SO_* 227 SOMAXCONN 228 MSG_* 229 SOL_* 230 SCM_* 231 IPPROTO_* 232 IPPORT_* 233 INADDR_* 234 IP_* 235 IPV6_* 236 EAI_* 237 AI_* 238 NI_* 239 TCP_* 240 241 Many constants of these forms, documented in the Unix documentation on sockets 242 and/or the IP protocol, are also defined in the socket module. They are 243 generally used in arguments to the :meth:`setsockopt` and :meth:`getsockopt` 244 methods of socket objects. In most cases, only those symbols that are defined 245 in the Unix header files are defined; for a few symbols, default values are 246 provided. 247 248.. data:: AF_CAN 249 PF_CAN 250 SOL_CAN_* 251 CAN_* 252 253 Many constants of these forms, documented in the Linux documentation, are 254 also defined in the socket module. 255 256 Availability: Linux >= 2.6.25. 257 258 .. versionadded:: 3.3 259 260.. data:: CAN_BCM 261 CAN_BCM_* 262 263 CAN_BCM, in the CAN protocol family, is the broadcast manager (BCM) protocol. 264 Broadcast manager constants, documented in the Linux documentation, are also 265 defined in the socket module. 266 267 Availability: Linux >= 2.6.25. 268 269 .. versionadded:: 3.4 270 271.. data:: AF_RDS 272 PF_RDS 273 SOL_RDS 274 RDS_* 275 276 Many constants of these forms, documented in the Linux documentation, are 277 also defined in the socket module. 278 279 Availability: Linux >= 2.6.30. 280 281 .. versionadded:: 3.3 282 283 284.. data:: SIO_* 285 RCVALL_* 286 287 Constants for Windows' WSAIoctl(). The constants are used as arguments to the 288 :meth:`ioctl` method of socket objects. 289 290 291.. data:: TIPC_* 292 293 TIPC related constants, matching the ones exported by the C socket API. See 294 the TIPC documentation for more information. 295 296 297.. data:: has_ipv6 298 299 This constant contains a boolean value which indicates if IPv6 is supported on 300 this platform. 301 302 303.. function:: create_connection(address[, timeout[, source_address]]) 304 305 Connect to a TCP service listening on the Internet *address* (a 2-tuple 306 ``(host, port)``), and return the socket object. This is a higher-level 307 function than :meth:`socket.connect`: if *host* is a non-numeric hostname, 308 it will try to resolve it for both :data:`AF_INET` and :data:`AF_INET6`, 309 and then try to connect to all possible addresses in turn until a 310 connection succeeds. This makes it easy to write clients that are 311 compatible to both IPv4 and IPv6. 312 313 Passing the optional *timeout* parameter will set the timeout on the 314 socket instance before attempting to connect. If no *timeout* is 315 supplied, the global default timeout setting returned by 316 :func:`getdefaulttimeout` is used. 317 318 If supplied, *source_address* must be a 2-tuple ``(host, port)`` for the 319 socket to bind to as its source address before connecting. If host or port 320 are '' or 0 respectively the OS default behavior will be used. 321 322 .. versionchanged:: 3.2 323 *source_address* was added. 324 325 .. versionchanged:: 3.2 326 support for the :keyword:`with` statement was added. 327 328 329.. function:: getaddrinfo(host, port, family=0, type=0, proto=0, flags=0) 330 331 Translate the *host*/*port* argument into a sequence of 5-tuples that contain 332 all the necessary arguments for creating a socket connected to that service. 333 *host* is a domain name, a string representation of an IPv4/v6 address 334 or ``None``. *port* is a string service name such as ``'http'``, a numeric 335 port number or ``None``. By passing ``None`` as the value of *host* 336 and *port*, you can pass ``NULL`` to the underlying C API. 337 338 The *family*, *type* and *proto* arguments can be optionally specified 339 in order to narrow the list of addresses returned. Passing zero as a 340 value for each of these arguments selects the full range of results. 341 The *flags* argument can be one or several of the ``AI_*`` constants, 342 and will influence how results are computed and returned. 343 For example, :const:`AI_NUMERICHOST` will disable domain name resolution 344 and will raise an error if *host* is a domain name. 345 346 The function returns a list of 5-tuples with the following structure: 347 348 ``(family, type, proto, canonname, sockaddr)`` 349 350 In these tuples, *family*, *type*, *proto* are all integers and are 351 meant to be passed to the :func:`socket` function. *canonname* will be 352 a string representing the canonical name of the *host* if 353 :const:`AI_CANONNAME` is part of the *flags* argument; else *canonname* 354 will be empty. *sockaddr* is a tuple describing a socket address, whose 355 format depends on the returned *family* (a ``(address, port)`` 2-tuple for 356 :const:`AF_INET`, a ``(address, port, flow info, scope id)`` 4-tuple for 357 :const:`AF_INET6`), and is meant to be passed to the :meth:`socket.connect` 358 method. 359 360 The following example fetches address information for a hypothetical TCP 361 connection to ``www.python.org`` on port 80 (results may differ on your 362 system if IPv6 isn't enabled):: 363 364 >>> socket.getaddrinfo("www.python.org", 80, proto=socket.SOL_TCP) 365 [(2, 1, 6, '', ('82.94.164.162', 80)), 366 (10, 1, 6, '', ('2001:888:2000:d::a2', 80, 0, 0))] 367 368 .. versionchanged:: 3.2 369 parameters can now be passed as single keyword arguments. 370 371.. function:: getfqdn([name]) 372 373 Return a fully qualified domain name for *name*. If *name* is omitted or empty, 374 it is interpreted as the local host. To find the fully qualified name, the 375 hostname returned by :func:`gethostbyaddr` is checked, followed by aliases for the 376 host, if available. The first name which includes a period is selected. In 377 case no fully qualified domain name is available, the hostname as returned by 378 :func:`gethostname` is returned. 379 380 381.. function:: gethostbyname(hostname) 382 383 Translate a host name to IPv4 address format. The IPv4 address is returned as a 384 string, such as ``'100.50.200.5'``. If the host name is an IPv4 address itself 385 it is returned unchanged. See :func:`gethostbyname_ex` for a more complete 386 interface. :func:`gethostbyname` does not support IPv6 name resolution, and 387 :func:`getaddrinfo` should be used instead for IPv4/v6 dual stack support. 388 389 390.. function:: gethostbyname_ex(hostname) 391 392 Translate a host name to IPv4 address format, extended interface. Return a 393 triple ``(hostname, aliaslist, ipaddrlist)`` where *hostname* is the primary 394 host name responding to the given *ip_address*, *aliaslist* is a (possibly 395 empty) list of alternative host names for the same address, and *ipaddrlist* is 396 a list of IPv4 addresses for the same interface on the same host (often but not 397 always a single address). :func:`gethostbyname_ex` does not support IPv6 name 398 resolution, and :func:`getaddrinfo` should be used instead for IPv4/v6 dual 399 stack support. 400 401 402.. function:: gethostname() 403 404 Return a string containing the hostname of the machine where the Python 405 interpreter is currently executing. 406 407 If you want to know the current machine's IP address, you may want to use 408 ``gethostbyname(gethostname())``. This operation assumes that there is a 409 valid address-to-host mapping for the host, and the assumption does not 410 always hold. 411 412 Note: :func:`gethostname` doesn't always return the fully qualified domain 413 name; use ``getfqdn()`` (see above). 414 415 416.. function:: gethostbyaddr(ip_address) 417 418 Return a triple ``(hostname, aliaslist, ipaddrlist)`` where *hostname* is the 419 primary host name responding to the given *ip_address*, *aliaslist* is a 420 (possibly empty) list of alternative host names for the same address, and 421 *ipaddrlist* is a list of IPv4/v6 addresses for the same interface on the same 422 host (most likely containing only a single address). To find the fully qualified 423 domain name, use the function :func:`getfqdn`. :func:`gethostbyaddr` supports 424 both IPv4 and IPv6. 425 426 427.. function:: getnameinfo(sockaddr, flags) 428 429 Translate a socket address *sockaddr* into a 2-tuple ``(host, port)``. Depending 430 on the settings of *flags*, the result can contain a fully-qualified domain name 431 or numeric address representation in *host*. Similarly, *port* can contain a 432 string port name or a numeric port number. 433 434 435.. function:: getprotobyname(protocolname) 436 437 Translate an Internet protocol name (for example, ``'icmp'``) to a constant 438 suitable for passing as the (optional) third argument to the :func:`socket` 439 function. This is usually only needed for sockets opened in "raw" mode 440 (:const:`SOCK_RAW`); for the normal socket modes, the correct protocol is chosen 441 automatically if the protocol is omitted or zero. 442 443 444.. function:: getservbyname(servicename[, protocolname]) 445 446 Translate an Internet service name and protocol name to a port number for that 447 service. The optional protocol name, if given, should be ``'tcp'`` or 448 ``'udp'``, otherwise any protocol will match. 449 450 451.. function:: getservbyport(port[, protocolname]) 452 453 Translate an Internet port number and protocol name to a service name for that 454 service. The optional protocol name, if given, should be ``'tcp'`` or 455 ``'udp'``, otherwise any protocol will match. 456 457 458.. function:: socket([family[, type[, proto]]]) 459 460 Create a new socket using the given address family, socket type and protocol 461 number. The address family should be :const:`AF_INET` (the default), 462 :const:`AF_INET6`, :const:`AF_UNIX`, :const:`AF_CAN` or :const:`AF_RDS`. The 463 socket type should be :const:`SOCK_STREAM` (the default), 464 :const:`SOCK_DGRAM`, :const:`SOCK_RAW` or perhaps one of the other ``SOCK_`` 465 constants. The protocol number is usually zero and may be omitted or in the 466 case where the address family is :const:`AF_CAN` the protocol should be one 467 of :const:`CAN_RAW` or :const:`CAN_BCM`. 468 469 .. versionchanged:: 3.3 470 The AF_CAN family was added. 471 The AF_RDS family was added. 472 473 .. versionchanged:: 3.4 474 The CAN_BCM protocol was added. 475 476.. function:: socketpair([family[, type[, proto]]]) 477 478 Build a pair of connected socket objects using the given address family, socket 479 type, and protocol number. Address family, socket type, and protocol number are 480 as for the :func:`socket` function above. The default family is :const:`AF_UNIX` 481 if defined on the platform; otherwise, the default is :const:`AF_INET`. 482 Availability: Unix. 483 484 .. versionchanged:: 3.2 485 The returned socket objects now support the whole socket API, rather 486 than a subset. 487 488 489.. function:: fromfd(fd, family, type[, proto]) 490 491 Duplicate the file descriptor *fd* (an integer as returned by a file object's 492 :meth:`fileno` method) and build a socket object from the result. Address 493 family, socket type and protocol number are as for the :func:`socket` function 494 above. The file descriptor should refer to a socket, but this is not checked --- 495 subsequent operations on the object may fail if the file descriptor is invalid. 496 This function is rarely needed, but can be used to get or set socket options on 497 a socket passed to a program as standard input or output (such as a server 498 started by the Unix inet daemon). The socket is assumed to be in blocking mode. 499 500 501.. function:: ntohl(x) 502 503 Convert 32-bit positive integers from network to host byte order. On machines 504 where the host byte order is the same as network byte order, this is a no-op; 505 otherwise, it performs a 4-byte swap operation. 506 507 508.. function:: ntohs(x) 509 510 Convert 16-bit positive integers from network to host byte order. On machines 511 where the host byte order is the same as network byte order, this is a no-op; 512 otherwise, it performs a 2-byte swap operation. 513 514 515.. function:: htonl(x) 516 517 Convert 32-bit positive integers from host to network byte order. On machines 518 where the host byte order is the same as network byte order, this is a no-op; 519 otherwise, it performs a 4-byte swap operation. 520 521 522.. function:: htons(x) 523 524 Convert 16-bit positive integers from host to network byte order. On machines 525 where the host byte order is the same as network byte order, this is a no-op; 526 otherwise, it performs a 2-byte swap operation. 527 528 529.. function:: inet_aton(ip_string) 530 531 Convert an IPv4 address from dotted-quad string format (for example, 532 '123.45.67.89') to 32-bit packed binary format, as a bytes object four characters in 533 length. This is useful when conversing with a program that uses the standard C 534 library and needs objects of type :c:type:`struct in_addr`, which is the C type 535 for the 32-bit packed binary this function returns. 536 537 :func:`inet_aton` also accepts strings with less than three dots; see the 538 Unix manual page :manpage:`inet(3)` for details. 539 540 If the IPv4 address string passed to this function is invalid, 541 :exc:`OSError` will be raised. Note that exactly what is valid depends on 542 the underlying C implementation of :c:func:`inet_aton`. 543 544 :func:`inet_aton` does not support IPv6, and :func:`inet_pton` should be used 545 instead for IPv4/v6 dual stack support. 546 547 548.. function:: inet_ntoa(packed_ip) 549 550 Convert a 32-bit packed IPv4 address (a bytes object four characters in 551 length) to its standard dotted-quad string representation (for example, 552 '123.45.67.89'). This is useful when conversing with a program that uses the 553 standard C library and needs objects of type :c:type:`struct in_addr`, which 554 is the C type for the 32-bit packed binary data this function takes as an 555 argument. 556 557 If the byte sequence passed to this function is not exactly 4 bytes in 558 length, :exc:`OSError` will be raised. :func:`inet_ntoa` does not 559 support IPv6, and :func:`inet_ntop` should be used instead for IPv4/v6 dual 560 stack support. 561 562 563.. function:: inet_pton(address_family, ip_string) 564 565 Convert an IP address from its family-specific string format to a packed, 566 binary format. :func:`inet_pton` is useful when a library or network protocol 567 calls for an object of type :c:type:`struct in_addr` (similar to 568 :func:`inet_aton`) or :c:type:`struct in6_addr`. 569 570 Supported values for *address_family* are currently :const:`AF_INET` and 571 :const:`AF_INET6`. If the IP address string *ip_string* is invalid, 572 :exc:`OSError` will be raised. Note that exactly what is valid depends on 573 both the value of *address_family* and the underlying implementation of 574 :c:func:`inet_pton`. 575 576 Availability: Unix (maybe not all platforms). 577 578 579.. function:: inet_ntop(address_family, packed_ip) 580 581 Convert a packed IP address (a bytes object of some number of characters) to its 582 standard, family-specific string representation (for example, ``'7.10.0.5'`` or 583 ``'5aef:2b::8'``). :func:`inet_ntop` is useful when a library or network protocol 584 returns an object of type :c:type:`struct in_addr` (similar to :func:`inet_ntoa`) 585 or :c:type:`struct in6_addr`. 586 587 Supported values for *address_family* are currently :const:`AF_INET` and 588 :const:`AF_INET6`. If the string *packed_ip* is not the correct length for the 589 specified address family, :exc:`ValueError` will be raised. A 590 :exc:`OSError` is raised for errors from the call to :func:`inet_ntop`. 591 592 Availability: Unix (maybe not all platforms). 593 594 595.. 596 XXX: Are sendmsg(), recvmsg() and CMSG_*() available on any 597 non-Unix platforms? The old (obsolete?) 4.2BSD form of the 598 interface, in which struct msghdr has no msg_control or 599 msg_controllen members, is not currently supported. 600 601.. function:: CMSG_LEN(length) 602 603 Return the total length, without trailing padding, of an ancillary 604 data item with associated data of the given *length*. This value 605 can often be used as the buffer size for :meth:`~socket.recvmsg` to 606 receive a single item of ancillary data, but :rfc:`3542` requires 607 portable applications to use :func:`CMSG_SPACE` and thus include 608 space for padding, even when the item will be the last in the 609 buffer. Raises :exc:`OverflowError` if *length* is outside the 610 permissible range of values. 611 612 Availability: most Unix platforms, possibly others. 613 614 .. versionadded:: 3.3 615 616 617.. function:: CMSG_SPACE(length) 618 619 Return the buffer size needed for :meth:`~socket.recvmsg` to 620 receive an ancillary data item with associated data of the given 621 *length*, along with any trailing padding. The buffer space needed 622 to receive multiple items is the sum of the :func:`CMSG_SPACE` 623 values for their associated data lengths. Raises 624 :exc:`OverflowError` if *length* is outside the permissible range 625 of values. 626 627 Note that some systems might support ancillary data without 628 providing this function. Also note that setting the buffer size 629 using the results of this function may not precisely limit the 630 amount of ancillary data that can be received, since additional 631 data may be able to fit into the padding area. 632 633 Availability: most Unix platforms, possibly others. 634 635 .. versionadded:: 3.3 636 637 638.. function:: getdefaulttimeout() 639 640 Return the default timeout in seconds (float) for new socket objects. A value 641 of ``None`` indicates that new socket objects have no timeout. When the socket 642 module is first imported, the default is ``None``. 643 644 645.. function:: setdefaulttimeout(timeout) 646 647 Set the default timeout in seconds (float) for new socket objects. When 648 the socket module is first imported, the default is ``None``. See 649 :meth:`~socket.settimeout` for possible values and their respective 650 meanings. 651 652 653.. function:: sethostname(name) 654 655 Set the machine's hostname to *name*. This will raise a 656 :exc:`OSError` if you don't have enough rights. 657 658 Availability: Unix. 659 660 .. versionadded:: 3.3 661 662 663.. function:: if_nameindex() 664 665 Return a list of network interface information 666 (index int, name string) tuples. 667 :exc:`OSError` if the system call fails. 668 669 Availability: Unix. 670 671 .. versionadded:: 3.3 672 673 674.. function:: if_nametoindex(if_name) 675 676 Return a network interface index number corresponding to an 677 interface name. 678 :exc:`OSError` if no interface with the given name exists. 679 680 Availability: Unix. 681 682 .. versionadded:: 3.3 683 684 685.. function:: if_indextoname(if_index) 686 687 Return a network interface name corresponding to a 688 interface index number. 689 :exc:`OSError` if no interface with the given index exists. 690 691 Availability: Unix. 692 693 .. versionadded:: 3.3 694 695 696.. function:: fromshare(data) 697 698 Instantiate a socket from data obtained from :meth:`~socket.share`. 699 The socket is assumed to be in blocking mode. 700 701 Availability: Windows. 702 703 .. versionadded:: 3.3 704 705 706.. data:: SocketType 707 708 This is a Python type object that represents the socket object type. It is the 709 same as ``type(socket(...))``. 710 711 712.. _socket-objects: 713 714Socket Objects 715-------------- 716 717Socket objects have the following methods. Except for :meth:`makefile` these 718correspond to Unix system calls applicable to sockets. 719 720 721.. method:: socket.accept() 722 723 Accept a connection. The socket must be bound to an address and listening for 724 connections. The return value is a pair ``(conn, address)`` where *conn* is a 725 *new* socket object usable to send and receive data on the connection, and 726 *address* is the address bound to the socket on the other end of the connection. 727 728 729.. method:: socket.bind(address) 730 731 Bind the socket to *address*. The socket must not already be bound. (The format 732 of *address* depends on the address family --- see above.) 733 734 735.. method:: socket.close() 736 737 Close the socket. All future operations on the socket object will fail. The 738 remote end will receive no more data (after queued data is flushed). Sockets are 739 automatically closed when they are garbage-collected. 740 741 .. note:: 742 :meth:`close()` releases the resource associated with a connection but 743 does not necessarily close the connection immediately. If you want 744 to close the connection in a timely fashion, call :meth:`shutdown()` 745 before :meth:`close()`. 746 747 748.. method:: socket.connect(address) 749 750 Connect to a remote socket at *address*. (The format of *address* depends on the 751 address family --- see above.) 752 753 754.. method:: socket.connect_ex(address) 755 756 Like ``connect(address)``, but return an error indicator instead of raising an 757 exception for errors returned by the C-level :c:func:`connect` call (other 758 problems, such as "host not found," can still raise exceptions). The error 759 indicator is ``0`` if the operation succeeded, otherwise the value of the 760 :c:data:`errno` variable. This is useful to support, for example, asynchronous 761 connects. 762 763 764.. method:: socket.detach() 765 766 Put the socket object into closed state without actually closing the 767 underlying file descriptor. The file descriptor is returned, and can 768 be reused for other purposes. 769 770 .. versionadded:: 3.2 771 772 773.. method:: socket.fileno() 774 775 Return the socket's file descriptor (a small integer). This is useful with 776 :func:`select.select`. 777 778 Under Windows the small integer returned by this method cannot be used where a 779 file descriptor can be used (such as :func:`os.fdopen`). Unix does not have 780 this limitation. 781 782 783.. method:: socket.getpeername() 784 785 Return the remote address to which the socket is connected. This is useful to 786 find out the port number of a remote IPv4/v6 socket, for instance. (The format 787 of the address returned depends on the address family --- see above.) On some 788 systems this function is not supported. 789 790 791.. method:: socket.getsockname() 792 793 Return the socket's own address. This is useful to find out the port number of 794 an IPv4/v6 socket, for instance. (The format of the address returned depends on 795 the address family --- see above.) 796 797 798.. method:: socket.getsockopt(level, optname[, buflen]) 799 800 Return the value of the given socket option (see the Unix man page 801 :manpage:`getsockopt(2)`). The needed symbolic constants (:const:`SO_\*` etc.) 802 are defined in this module. If *buflen* is absent, an integer option is assumed 803 and its integer value is returned by the function. If *buflen* is present, it 804 specifies the maximum length of the buffer used to receive the option in, and 805 this buffer is returned as a bytes object. It is up to the caller to decode the 806 contents of the buffer (see the optional built-in module :mod:`struct` for a way 807 to decode C structures encoded as byte strings). 808 809 810.. method:: socket.gettimeout() 811 812 Return the timeout in seconds (float) associated with socket operations, 813 or ``None`` if no timeout is set. This reflects the last call to 814 :meth:`setblocking` or :meth:`settimeout`. 815 816 817.. method:: socket.ioctl(control, option) 818 819 :platform: Windows 820 821 The :meth:`ioctl` method is a limited interface to the WSAIoctl system 822 interface. Please refer to the `Win32 documentation 823 <http://msdn.microsoft.com/en-us/library/ms741621%28VS.85%29.aspx>`_ for more 824 information. 825 826 On other platforms, the generic :func:`fcntl.fcntl` and :func:`fcntl.ioctl` 827 functions may be used; they accept a socket object as their first argument. 828 829.. method:: socket.listen(backlog) 830 831 Listen for connections made to the socket. The *backlog* argument specifies the 832 maximum number of queued connections and should be at least 0; the maximum value 833 is system-dependent (usually 5), the minimum value is forced to 0. 834 835 836.. method:: socket.makefile(mode='r', buffering=None, *, encoding=None, \ 837 errors=None, newline=None) 838 839 .. index:: single: I/O control; buffering 840 841 Return a :term:`file object` associated with the socket. The exact returned 842 type depends on the arguments given to :meth:`makefile`. These arguments are 843 interpreted the same way as by the built-in :func:`open` function. 844 845 Closing the file object won't close the socket unless there are no remaining 846 references to the socket. The socket must be in blocking mode; it can have 847 a timeout, but the file object's internal buffer may end up in a inconsistent 848 state if a timeout occurs. 849 850 .. note:: 851 852 On Windows, the file-like object created by :meth:`makefile` cannot be 853 used where a file object with a file descriptor is expected, such as the 854 stream arguments of :meth:`subprocess.Popen`. 855 856 857.. method:: socket.recv(bufsize[, flags]) 858 859 Receive data from the socket. The return value is a bytes object representing the 860 data received. The maximum amount of data to be received at once is specified 861 by *bufsize*. See the Unix manual page :manpage:`recv(2)` for the meaning of 862 the optional argument *flags*; it defaults to zero. 863 864 .. note:: 865 866 For best match with hardware and network realities, the value of *bufsize* 867 should be a relatively small power of 2, for example, 4096. 868 869 870.. method:: socket.recvfrom(bufsize[, flags]) 871 872 Receive data from the socket. The return value is a pair ``(bytes, address)`` 873 where *bytes* is a bytes object representing the data received and *address* is the 874 address of the socket sending the data. See the Unix manual page 875 :manpage:`recv(2)` for the meaning of the optional argument *flags*; it defaults 876 to zero. (The format of *address* depends on the address family --- see above.) 877 878 879.. method:: socket.recvmsg(bufsize[, ancbufsize[, flags]]) 880 881 Receive normal data (up to *bufsize* bytes) and ancillary data from 882 the socket. The *ancbufsize* argument sets the size in bytes of 883 the internal buffer used to receive the ancillary data; it defaults 884 to 0, meaning that no ancillary data will be received. Appropriate 885 buffer sizes for ancillary data can be calculated using 886 :func:`CMSG_SPACE` or :func:`CMSG_LEN`, and items which do not fit 887 into the buffer might be truncated or discarded. The *flags* 888 argument defaults to 0 and has the same meaning as for 889 :meth:`recv`. 890 891 The return value is a 4-tuple: ``(data, ancdata, msg_flags, 892 address)``. The *data* item is a :class:`bytes` object holding the 893 non-ancillary data received. The *ancdata* item is a list of zero 894 or more tuples ``(cmsg_level, cmsg_type, cmsg_data)`` representing 895 the ancillary data (control messages) received: *cmsg_level* and 896 *cmsg_type* are integers specifying the protocol level and 897 protocol-specific type respectively, and *cmsg_data* is a 898 :class:`bytes` object holding the associated data. The *msg_flags* 899 item is the bitwise OR of various flags indicating conditions on 900 the received message; see your system documentation for details. 901 If the receiving socket is unconnected, *address* is the address of 902 the sending socket, if available; otherwise, its value is 903 unspecified. 904 905 On some systems, :meth:`sendmsg` and :meth:`recvmsg` can be used to 906 pass file descriptors between processes over an :const:`AF_UNIX` 907 socket. When this facility is used (it is often restricted to 908 :const:`SOCK_STREAM` sockets), :meth:`recvmsg` will return, in its 909 ancillary data, items of the form ``(socket.SOL_SOCKET, 910 socket.SCM_RIGHTS, fds)``, where *fds* is a :class:`bytes` object 911 representing the new file descriptors as a binary array of the 912 native C :c:type:`int` type. If :meth:`recvmsg` raises an 913 exception after the system call returns, it will first attempt to 914 close any file descriptors received via this mechanism. 915 916 Some systems do not indicate the truncated length of ancillary data 917 items which have been only partially received. If an item appears 918 to extend beyond the end of the buffer, :meth:`recvmsg` will issue 919 a :exc:`RuntimeWarning`, and will return the part of it which is 920 inside the buffer provided it has not been truncated before the 921 start of its associated data. 922 923 On systems which support the :const:`SCM_RIGHTS` mechanism, the 924 following function will receive up to *maxfds* file descriptors, 925 returning the message data and a list containing the descriptors 926 (while ignoring unexpected conditions such as unrelated control 927 messages being received). See also :meth:`sendmsg`. :: 928 929 import socket, array 930 931 def recv_fds(sock, msglen, maxfds): 932 fds = array.array("i") # Array of ints 933 msg, ancdata, flags, addr = sock.recvmsg(msglen, socket.CMSG_LEN(maxfds * fds.itemsize)) 934 for cmsg_level, cmsg_type, cmsg_data in ancdata: 935 if (cmsg_level == socket.SOL_SOCKET and cmsg_type == socket.SCM_RIGHTS): 936 # Append data, ignoring any truncated integers at the end. 937 fds.fromstring(cmsg_data[:len(cmsg_data) - (len(cmsg_data) % fds.itemsize)]) 938 return msg, list(fds) 939 940 Availability: most Unix platforms, possibly others. 941 942 .. versionadded:: 3.3 943 944 945.. method:: socket.recvmsg_into(buffers[, ancbufsize[, flags]]) 946 947 Receive normal data and ancillary data from the socket, behaving as 948 :meth:`recvmsg` would, but scatter the non-ancillary data into a 949 series of buffers instead of returning a new bytes object. The 950 *buffers* argument must be an iterable of objects that export 951 writable buffers (e.g. :class:`bytearray` objects); these will be 952 filled with successive chunks of the non-ancillary data until it 953 has all been written or there are no more buffers. The operating 954 system may set a limit (:func:`~os.sysconf` value ``SC_IOV_MAX``) 955 on the number of buffers that can be used. The *ancbufsize* and 956 *flags* arguments have the same meaning as for :meth:`recvmsg`. 957 958 The return value is a 4-tuple: ``(nbytes, ancdata, msg_flags, 959 address)``, where *nbytes* is the total number of bytes of 960 non-ancillary data written into the buffers, and *ancdata*, 961 *msg_flags* and *address* are the same as for :meth:`recvmsg`. 962 963 Example:: 964 965 >>> import socket 966 >>> s1, s2 = socket.socketpair() 967 >>> b1 = bytearray(b'----') 968 >>> b2 = bytearray(b'0123456789') 969 >>> b3 = bytearray(b'--------------') 970 >>> s1.send(b'Mary had a little lamb') 971 22 972 >>> s2.recvmsg_into([b1, memoryview(b2)[2:9], b3]) 973 (22, [], 0, None) 974 >>> [b1, b2, b3] 975 [bytearray(b'Mary'), bytearray(b'01 had a 9'), bytearray(b'little lamb---')] 976 977 Availability: most Unix platforms, possibly others. 978 979 .. versionadded:: 3.3 980 981 982.. method:: socket.recvfrom_into(buffer[, nbytes[, flags]]) 983 984 Receive data from the socket, writing it into *buffer* instead of creating a 985 new bytestring. The return value is a pair ``(nbytes, address)`` where *nbytes* is 986 the number of bytes received and *address* is the address of the socket sending 987 the data. See the Unix manual page :manpage:`recv(2)` for the meaning of the 988 optional argument *flags*; it defaults to zero. (The format of *address* 989 depends on the address family --- see above.) 990 991 992.. method:: socket.recv_into(buffer[, nbytes[, flags]]) 993 994 Receive up to *nbytes* bytes from the socket, storing the data into a buffer 995 rather than creating a new bytestring. If *nbytes* is not specified (or 0), 996 receive up to the size available in the given buffer. Returns the number of 997 bytes received. See the Unix manual page :manpage:`recv(2)` for the meaning 998 of the optional argument *flags*; it defaults to zero. 999 1000 1001.. method:: socket.send(bytes[, flags]) 1002 1003 Send data to the socket. The socket must be connected to a remote socket. The 1004 optional *flags* argument has the same meaning as for :meth:`recv` above. 1005 Returns the number of bytes sent. Applications are responsible for checking that 1006 all data has been sent; if only some of the data was transmitted, the 1007 application needs to attempt delivery of the remaining data. For further 1008 information on this topic, consult the :ref:`socket-howto`. 1009 1010 1011.. method:: socket.sendall(bytes[, flags]) 1012 1013 Send data to the socket. The socket must be connected to a remote socket. The 1014 optional *flags* argument has the same meaning as for :meth:`recv` above. 1015 Unlike :meth:`send`, this method continues to send data from *bytes* until 1016 either all data has been sent or an error occurs. ``None`` is returned on 1017 success. On error, an exception is raised, and there is no way to determine how 1018 much data, if any, was successfully sent. 1019 1020 1021.. method:: socket.sendto(bytes, address) 1022 socket.sendto(bytes, flags, address) 1023 1024 Send data to the socket. The socket should not be connected to a remote socket, 1025 since the destination socket is specified by *address*. The optional *flags* 1026 argument has the same meaning as for :meth:`recv` above. Return the number of 1027 bytes sent. (The format of *address* depends on the address family --- see 1028 above.) 1029 1030 1031.. method:: socket.sendmsg(buffers[, ancdata[, flags[, address]]]) 1032 1033 Send normal and ancillary data to the socket, gathering the 1034 non-ancillary data from a series of buffers and concatenating it 1035 into a single message. The *buffers* argument specifies the 1036 non-ancillary data as an iterable of buffer-compatible objects 1037 (e.g. :class:`bytes` objects); the operating system may set a limit 1038 (:func:`~os.sysconf` value ``SC_IOV_MAX``) on the number of buffers 1039 that can be used. The *ancdata* argument specifies the ancillary 1040 data (control messages) as an iterable of zero or more tuples 1041 ``(cmsg_level, cmsg_type, cmsg_data)``, where *cmsg_level* and 1042 *cmsg_type* are integers specifying the protocol level and 1043 protocol-specific type respectively, and *cmsg_data* is a 1044 buffer-compatible object holding the associated data. Note that 1045 some systems (in particular, systems without :func:`CMSG_SPACE`) 1046 might support sending only one control message per call. The 1047 *flags* argument defaults to 0 and has the same meaning as for 1048 :meth:`send`. If *address* is supplied and not ``None``, it sets a 1049 destination address for the message. The return value is the 1050 number of bytes of non-ancillary data sent. 1051 1052 The following function sends the list of file descriptors *fds* 1053 over an :const:`AF_UNIX` socket, on systems which support the 1054 :const:`SCM_RIGHTS` mechanism. See also :meth:`recvmsg`. :: 1055 1056 import socket, array 1057 1058 def send_fds(sock, msg, fds): 1059 return sock.sendmsg([msg], [(socket.SOL_SOCKET, socket.SCM_RIGHTS, array.array("i", fds))]) 1060 1061 Availability: most Unix platforms, possibly others. 1062 1063 .. versionadded:: 3.3 1064 1065 1066.. method:: socket.setblocking(flag) 1067 1068 Set blocking or non-blocking mode of the socket: if *flag* is false, the 1069 socket is set to non-blocking, else to blocking mode. 1070 1071 This method is a shorthand for certain :meth:`~socket.settimeout` calls: 1072 1073 * ``sock.setblocking(True)`` is equivalent to ``sock.settimeout(None)`` 1074 1075 * ``sock.setblocking(False)`` is equivalent to ``sock.settimeout(0.0)`` 1076 1077 1078.. method:: socket.settimeout(value) 1079 1080 Set a timeout on blocking socket operations. The *value* argument can be a 1081 nonnegative floating point number expressing seconds, or ``None``. 1082 If a non-zero value is given, subsequent socket operations will raise a 1083 :exc:`timeout` exception if the timeout period *value* has elapsed before 1084 the operation has completed. If zero is given, the socket is put in 1085 non-blocking mode. If ``None`` is given, the socket is put in blocking mode. 1086 1087 For further information, please consult the :ref:`notes on socket timeouts <socket-timeouts>`. 1088 1089 1090.. method:: socket.setsockopt(level, optname, value) 1091 1092 .. index:: module: struct 1093 1094 Set the value of the given socket option (see the Unix manual page 1095 :manpage:`setsockopt(2)`). The needed symbolic constants are defined in the 1096 :mod:`socket` module (:const:`SO_\*` etc.). The value can be an integer or a 1097 bytes object representing a buffer. In the latter case it is up to the caller to 1098 ensure that the bytestring contains the proper bits (see the optional built-in 1099 module :mod:`struct` for a way to encode C structures as bytestrings). 1100 1101 1102.. method:: socket.shutdown(how) 1103 1104 Shut down one or both halves of the connection. If *how* is :const:`SHUT_RD`, 1105 further receives are disallowed. If *how* is :const:`SHUT_WR`, further sends 1106 are disallowed. If *how* is :const:`SHUT_RDWR`, further sends and receives are 1107 disallowed. 1108 1109 1110.. method:: socket.share(process_id) 1111 1112 :platform: Windows 1113 1114 Duplacet a socket and prepare it for sharing with a target process. The 1115 target process must be provided with *process_id*. The resulting bytes object 1116 can then be passed to the target process using some form of interprocess 1117 communication and the socket can be recreated there using :func:`fromshare`. 1118 Once this method has been called, it is safe to close the socket since 1119 the operating system has already duplicated it for the target process. 1120 1121 .. versionadded:: 3.3 1122 1123 1124Note that there are no methods :meth:`read` or :meth:`write`; use 1125:meth:`~socket.recv` and :meth:`~socket.send` without *flags* argument instead. 1126 1127Socket objects also have these (read-only) attributes that correspond to the 1128values given to the :class:`socket` constructor. 1129 1130 1131.. attribute:: socket.family 1132 1133 The socket family. 1134 1135 1136.. attribute:: socket.type 1137 1138 The socket type. 1139 1140 1141.. attribute:: socket.proto 1142 1143 The socket protocol. 1144 1145 1146 1147.. _socket-timeouts: 1148 1149Notes on socket timeouts 1150------------------------ 1151 1152A socket object can be in one of three modes: blocking, non-blocking, or 1153timeout. Sockets are by default always created in blocking mode, but this 1154can be changed by calling :func:`setdefaulttimeout`. 1155 1156* In *blocking mode*, operations block until complete or the system returns 1157 an error (such as connection timed out). 1158 1159* In *non-blocking mode*, operations fail (with an error that is unfortunately 1160 system-dependent) if they cannot be completed immediately: functions from the 1161 :mod:`select` can be used to know when and whether a socket is available for 1162 reading or writing. 1163 1164* In *timeout mode*, operations fail if they cannot be completed within the 1165 timeout specified for the socket (they raise a :exc:`timeout` exception) 1166 or if the system returns an error. 1167 1168.. note:: 1169 At the operating system level, sockets in *timeout mode* are internally set 1170 in non-blocking mode. Also, the blocking and timeout modes are shared between 1171 file descriptors and socket objects that refer to the same network endpoint. 1172 This implementation detail can have visible consequences if e.g. you decide 1173 to use the :meth:`~socket.fileno()` of a socket. 1174 1175Timeouts and the ``connect`` method 1176^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 1177 1178The :meth:`~socket.connect` operation is also subject to the timeout 1179setting, and in general it is recommended to call :meth:`~socket.settimeout` 1180before calling :meth:`~socket.connect` or pass a timeout parameter to 1181:meth:`create_connection`. However, the system network stack may also 1182return a connection timeout error of its own regardless of any Python socket 1183timeout setting. 1184 1185Timeouts and the ``accept`` method 1186^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 1187 1188If :func:`getdefaulttimeout` is not :const:`None`, sockets returned by 1189the :meth:`~socket.accept` method inherit that timeout. Otherwise, the 1190behaviour depends on settings of the listening socket: 1191 1192* if the listening socket is in *blocking mode* or in *timeout mode*, 1193 the socket returned by :meth:`~socket.accept` is in *blocking mode*; 1194 1195* if the listening socket is in *non-blocking mode*, whether the socket 1196 returned by :meth:`~socket.accept` is in blocking or non-blocking mode 1197 is operating system-dependent. If you want to ensure cross-platform 1198 behaviour, it is recommended you manually override this setting. 1199 1200 1201.. _socket-example: 1202 1203Example 1204------- 1205 1206Here are four minimal example programs using the TCP/IP protocol: a server that 1207echoes all data that it receives back (servicing only one client), and a client 1208using it. Note that a server must perform the sequence :func:`socket`, 1209:meth:`~socket.bind`, :meth:`~socket.listen`, :meth:`~socket.accept` (possibly 1210repeating the :meth:`~socket.accept` to service more than one client), while a 1211client only needs the sequence :func:`socket`, :meth:`~socket.connect`. Also 1212note that the server does not :meth:`~socket.sendall`/:meth:`~socket.recv` on 1213the socket it is listening on but on the new socket returned by 1214:meth:`~socket.accept`. 1215 1216The first two examples support IPv4 only. :: 1217 1218 # Echo server program 1219 import socket 1220 1221 HOST = '' # Symbolic name meaning all available interfaces 1222 PORT = 50007 # Arbitrary non-privileged port 1223 s = socket.socket(socket.AF_INET, socket.SOCK_STREAM) 1224 s.bind((HOST, PORT)) 1225 s.listen(1) 1226 conn, addr = s.accept() 1227 print('Connected by', addr) 1228 while True: 1229 data = conn.recv(1024) 1230 if not data: break 1231 conn.sendall(data) 1232 conn.close() 1233 1234:: 1235 1236 # Echo client program 1237 import socket 1238 1239 HOST = 'daring.cwi.nl' # The remote host 1240 PORT = 50007 # The same port as used by the server 1241 s = socket.socket(socket.AF_INET, socket.SOCK_STREAM) 1242 s.connect((HOST, PORT)) 1243 s.sendall(b'Hello, world') 1244 data = s.recv(1024) 1245 s.close() 1246 print('Received', repr(data)) 1247 1248The next two examples are identical to the above two, but support both IPv4 and 1249IPv6. The server side will listen to the first address family available (it 1250should listen to both instead). On most of IPv6-ready systems, IPv6 will take 1251precedence and the server may not accept IPv4 traffic. The client side will try 1252to connect to the all addresses returned as a result of the name resolution, and 1253sends traffic to the first one connected successfully. :: 1254 1255 # Echo server program 1256 import socket 1257 import sys 1258 1259 HOST = None # Symbolic name meaning all available interfaces 1260 PORT = 50007 # Arbitrary non-privileged port 1261 s = None 1262 for res in socket.getaddrinfo(HOST, PORT, socket.AF_UNSPEC, 1263 socket.SOCK_STREAM, 0, socket.AI_PASSIVE): 1264 af, socktype, proto, canonname, sa = res 1265 try: 1266 s = socket.socket(af, socktype, proto) 1267 except OSError as msg: 1268 s = None 1269 continue 1270 try: 1271 s.bind(sa) 1272 s.listen(1) 1273 except OSError as msg: 1274 s.close() 1275 s = None 1276 continue 1277 break 1278 if s is None: 1279 print('could not open socket') 1280 sys.exit(1) 1281 conn, addr = s.accept() 1282 print('Connected by', addr) 1283 while True: 1284 data = conn.recv(1024) 1285 if not data: break 1286 conn.send(data) 1287 conn.close() 1288 1289:: 1290 1291 # Echo client program 1292 import socket 1293 import sys 1294 1295 HOST = 'daring.cwi.nl' # The remote host 1296 PORT = 50007 # The same port as used by the server 1297 s = None 1298 for res in socket.getaddrinfo(HOST, PORT, socket.AF_UNSPEC, socket.SOCK_STREAM): 1299 af, socktype, proto, canonname, sa = res 1300 try: 1301 s = socket.socket(af, socktype, proto) 1302 except OSError as msg: 1303 s = None 1304 continue 1305 try: 1306 s.connect(sa) 1307 except OSError as msg: 1308 s.close() 1309 s = None 1310 continue 1311 break 1312 if s is None: 1313 print('could not open socket') 1314 sys.exit(1) 1315 s.sendall(b'Hello, world') 1316 data = s.recv(1024) 1317 s.close() 1318 print('Received', repr(data)) 1319 1320 1321The next example shows how to write a very simple network sniffer with raw 1322sockets on Windows. The example requires administrator privileges to modify 1323the interface:: 1324 1325 import socket 1326 1327 # the public network interface 1328 HOST = socket.gethostbyname(socket.gethostname()) 1329 1330 # create a raw socket and bind it to the public interface 1331 s = socket.socket(socket.AF_INET, socket.SOCK_RAW, socket.IPPROTO_IP) 1332 s.bind((HOST, 0)) 1333 1334 # Include IP headers 1335 s.setsockopt(socket.IPPROTO_IP, socket.IP_HDRINCL, 1) 1336 1337 # receive all packages 1338 s.ioctl(socket.SIO_RCVALL, socket.RCVALL_ON) 1339 1340 # receive a package 1341 print(s.recvfrom(65565)) 1342 1343 # disabled promiscuous mode 1344 s.ioctl(socket.SIO_RCVALL, socket.RCVALL_OFF) 1345 1346The last example shows how to use the socket interface to communicate to a CAN 1347network using the raw socket protocol. To use CAN with the broadcast 1348manager protocol instead, open a socket with:: 1349 1350 socket.socket(socket.AF_CAN, socket.SOCK_DGRAM, socket.CAN_BCM) 1351 1352After binding (:const:`CAN_RAW`) or connecting (:const:`CAN_BCM`) the socket, you 1353can use the :meth:`socket.send`, and the :meth:`socket.recv` operations (and 1354their counterparts) on the socket object as usual. 1355 1356This example might require special priviledge:: 1357 1358 import socket 1359 import struct 1360 1361 1362 # CAN frame packing/unpacking (see 'struct can_frame' in <linux/can.h>) 1363 1364 can_frame_fmt = "=IB3x8s" 1365 can_frame_size = struct.calcsize(can_frame_fmt) 1366 1367 def build_can_frame(can_id, data): 1368 can_dlc = len(data) 1369 data = data.ljust(8, b'\x00') 1370 return struct.pack(can_frame_fmt, can_id, can_dlc, data) 1371 1372 def dissect_can_frame(frame): 1373 can_id, can_dlc, data = struct.unpack(can_frame_fmt, frame) 1374 return (can_id, can_dlc, data[:can_dlc]) 1375 1376 1377 # create a raw socket and bind it to the 'vcan0' interface 1378 s = socket.socket(socket.AF_CAN, socket.SOCK_RAW, socket.CAN_RAW) 1379 s.bind(('vcan0',)) 1380 1381 while True: 1382 cf, addr = s.recvfrom(can_frame_size) 1383 1384 print('Received: can_id=%x, can_dlc=%x, data=%s' % dissect_can_frame(cf)) 1385 1386 try: 1387 s.send(cf) 1388 except OSError: 1389 print('Error sending CAN frame') 1390 1391 try: 1392 s.send(build_can_frame(0x01, b'\x01\x02\x03')) 1393 except OSError: 1394 print('Error sending CAN frame') 1395 1396Running an example several times with too small delay between executions, could 1397lead to this error:: 1398 1399 OSError: [Errno 98] Address already in use 1400 1401This is because the previous execution has left the socket in a ``TIME_WAIT`` 1402state, and can't be immediately reused. 1403 1404There is a :mod:`socket` flag to set, in order to prevent this, 1405:data:`socket.SO_REUSEADDR`:: 1406 1407 s = socket.socket(socket.AF_INET, socket.SOCK_STREAM) 1408 s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1) 1409 s.bind((HOST, PORT)) 1410 1411the :data:`SO_REUSEADDR` flag tells the kernel to reuse a local socket in 1412``TIME_WAIT`` state, without waiting for its natural timeout to expire. 1413 1414 1415.. seealso:: 1416 1417 For an introduction to socket programming (in C), see the following papers: 1418 1419 - *An Introductory 4.3BSD Interprocess Communication Tutorial*, by Stuart Sechrest 1420 1421 - *An Advanced 4.3BSD Interprocess Communication Tutorial*, by Samuel J. Leffler et 1422 al, 1423 1424 both in the UNIX Programmer's Manual, Supplementary Documents 1 (sections 1425 PS1:7 and PS1:8). The platform-specific reference material for the various 1426 socket-related system calls are also a valuable source of information on the 1427 details of socket semantics. For Unix, refer to the manual pages; for Windows, 1428 see the WinSock (or Winsock 2) specification. For IPv6-ready APIs, readers may 1429 want to refer to :rfc:`3493` titled Basic Socket Interface Extensions for IPv6. 1430