PageRenderTime 55ms CodeModel.GetById 17ms RepoModel.GetById 1ms app.codeStats 0ms

/Doc/library/socket.rst

https://bitbucket.org/yomgui/cpython
ReStructuredText | 1542 lines | 1074 code | 468 blank | 0 comment | 0 complexity | 73ee6d5ac799df50a19fda50ad211249 MD5 | raw file
Possible License(s): 0BSD
  1. :mod:`socket` --- Low-level networking interface
  2. ================================================
  3. .. module:: socket
  4. :synopsis: Low-level networking interface.
  5. This module provides access to the BSD *socket* interface. It is available on
  6. all modern Unix systems, Windows, MacOS, and probably additional platforms.
  7. .. note::
  8. Some behavior may be platform dependent, since calls are made to the operating
  9. system socket APIs.
  10. .. index:: object: socket
  11. The Python interface is a straightforward transliteration of the Unix system
  12. call and library interface for sockets to Python's object-oriented style: the
  13. :func:`.socket` function returns a :dfn:`socket object` whose methods implement
  14. the various socket system calls. Parameter types are somewhat higher-level than
  15. in the C interface: as with :meth:`read` and :meth:`write` operations on Python
  16. files, buffer allocation on receive operations is automatic, and buffer length
  17. is implicit on send operations.
  18. .. seealso::
  19. Module :mod:`socketserver`
  20. Classes that simplify writing network servers.
  21. Module :mod:`ssl`
  22. A TLS/SSL wrapper for socket objects.
  23. Socket families
  24. ---------------
  25. Depending on the system and the build options, various socket families
  26. are supported by this module.
  27. The address format required by a particular socket object is automatically
  28. selected based on the address family specified when the socket object was
  29. created. Socket addresses are represented as follows:
  30. - The address of an :const:`AF_UNIX` socket bound to a file system node
  31. is represented as a string, using the file system encoding and the
  32. ``'surrogateescape'`` error handler (see :pep:`383`). An address in
  33. Linux's abstract namespace is returned as a :class:`bytes` object with
  34. an initial null byte; note that sockets in this namespace can
  35. communicate with normal file system sockets, so programs intended to
  36. run on Linux may need to deal with both types of address. A string or
  37. :class:`bytes` object can be used for either type of address when
  38. passing it as an argument.
  39. .. versionchanged:: 3.3
  40. Previously, :const:`AF_UNIX` socket paths were assumed to use UTF-8
  41. encoding.
  42. - A pair ``(host, port)`` is used for the :const:`AF_INET` address family,
  43. where *host* is a string representing either a hostname in Internet domain
  44. notation like ``'daring.cwi.nl'`` or an IPv4 address like ``'100.50.200.5'``,
  45. and *port* is an integer.
  46. - For :const:`AF_INET6` address family, a four-tuple ``(host, port, flowinfo,
  47. scopeid)`` is used, where *flowinfo* and *scopeid* represent the ``sin6_flowinfo``
  48. and ``sin6_scope_id`` members in :const:`struct sockaddr_in6` in C. For
  49. :mod:`socket` module methods, *flowinfo* and *scopeid* can be omitted just for
  50. backward compatibility. Note, however, omission of *scopeid* can cause problems
  51. in manipulating scoped IPv6 addresses.
  52. - :const:`AF_NETLINK` sockets are represented as pairs ``(pid, groups)``.
  53. - Linux-only support for TIPC is available using the :const:`AF_TIPC`
  54. address family. TIPC is an open, non-IP based networked protocol designed
  55. for use in clustered computer environments. Addresses are represented by a
  56. tuple, and the fields depend on the address type. The general tuple form is
  57. ``(addr_type, v1, v2, v3 [, scope])``, where:
  58. - *addr_type* is one of :const:`TIPC_ADDR_NAMESEQ`, :const:`TIPC_ADDR_NAME`,
  59. or :const:`TIPC_ADDR_ID`.
  60. - *scope* is one of :const:`TIPC_ZONE_SCOPE`, :const:`TIPC_CLUSTER_SCOPE`, and
  61. :const:`TIPC_NODE_SCOPE`.
  62. - If *addr_type* is :const:`TIPC_ADDR_NAME`, then *v1* is the server type, *v2* is
  63. the port identifier, and *v3* should be 0.
  64. If *addr_type* is :const:`TIPC_ADDR_NAMESEQ`, then *v1* is the server type, *v2*
  65. is the lower port number, and *v3* is the upper port number.
  66. If *addr_type* is :const:`TIPC_ADDR_ID`, then *v1* is the node, *v2* is the
  67. reference, and *v3* should be set to 0.
  68. - A tuple ``(interface, )`` is used for the :const:`AF_CAN` address family,
  69. where *interface* is a string representing a network interface name like
  70. ``'can0'``. The network interface name ``''`` can be used to receive packets
  71. from all network interfaces of this family.
  72. - A string or a tuple ``(id, unit)`` is used for the :const:`SYSPROTO_CONTROL`
  73. protocol of the :const:`PF_SYSTEM` family. The string is the name of a
  74. kernel control using a dynamically-assigned ID. The tuple can be used if ID
  75. and unit number of the kernel control are known or if a registered ID is
  76. used.
  77. .. versionadded:: 3.3
  78. - Certain other address families (:const:`AF_BLUETOOTH`, :const:`AF_PACKET`,
  79. :const:`AF_CAN`) support specific representations.
  80. .. XXX document them!
  81. For IPv4 addresses, two special forms are accepted instead of a host address:
  82. the empty string represents :const:`INADDR_ANY`, and the string
  83. ``'<broadcast>'`` represents :const:`INADDR_BROADCAST`. This behavior is not
  84. compatible with IPv6, therefore, you may want to avoid these if you intend
  85. to support IPv6 with your Python programs.
  86. If you use a hostname in the *host* portion of IPv4/v6 socket address, the
  87. program may show a nondeterministic behavior, as Python uses the first address
  88. returned from the DNS resolution. The socket address will be resolved
  89. differently into an actual IPv4/v6 address, depending on the results from DNS
  90. resolution and/or the host configuration. For deterministic behavior use a
  91. numeric address in *host* portion.
  92. All errors raise exceptions. The normal exceptions for invalid argument types
  93. and out-of-memory conditions can be raised; starting from Python 3.3, errors
  94. related to socket or address semantics raise :exc:`OSError` or one of its
  95. subclasses (they used to raise :exc:`socket.error`).
  96. Non-blocking mode is supported through :meth:`~socket.setblocking`. A
  97. generalization of this based on timeouts is supported through
  98. :meth:`~socket.settimeout`.
  99. Module contents
  100. ---------------
  101. The module :mod:`socket` exports the following elements.
  102. Exceptions
  103. ^^^^^^^^^^
  104. .. exception:: error
  105. A deprecated alias of :exc:`OSError`.
  106. .. versionchanged:: 3.3
  107. Following :pep:`3151`, this class was made an alias of :exc:`OSError`.
  108. .. exception:: herror
  109. A subclass of :exc:`OSError`, this exception is raised for
  110. address-related errors, i.e. for functions that use *h_errno* in the POSIX
  111. C API, including :func:`gethostbyname_ex` and :func:`gethostbyaddr`.
  112. The accompanying value is a pair ``(h_errno, string)`` representing an
  113. error returned by a library call. *h_errno* is a numeric value, while
  114. *string* represents the description of *h_errno*, as returned by the
  115. :c:func:`hstrerror` C function.
  116. .. versionchanged:: 3.3
  117. This class was made a subclass of :exc:`OSError`.
  118. .. exception:: gaierror
  119. A subclass of :exc:`OSError`, this exception is raised for
  120. address-related errors by :func:`getaddrinfo` and :func:`getnameinfo`.
  121. The accompanying value is a pair ``(error, string)`` representing an error
  122. returned by a library call. *string* represents the description of
  123. *error*, as returned by the :c:func:`gai_strerror` C function. The
  124. numeric *error* value will match one of the :const:`EAI_\*` constants
  125. defined in this module.
  126. .. versionchanged:: 3.3
  127. This class was made a subclass of :exc:`OSError`.
  128. .. exception:: timeout
  129. A subclass of :exc:`OSError`, this exception is raised when a timeout
  130. occurs on a socket which has had timeouts enabled via a prior call to
  131. :meth:`~socket.settimeout` (or implicitly through
  132. :func:`~socket.setdefaulttimeout`). The accompanying value is a string
  133. whose value is currently always "timed out".
  134. .. versionchanged:: 3.3
  135. This class was made a subclass of :exc:`OSError`.
  136. Constants
  137. ^^^^^^^^^
  138. The AF_* and SOCK_* constants are now :class:`AddressFamily` and
  139. :class:`SocketKind` :class:`.IntEnum` collections.
  140. .. versionadded:: 3.4
  141. .. data:: AF_UNIX
  142. AF_INET
  143. AF_INET6
  144. These constants represent the address (and protocol) families, used for the
  145. first argument to :func:`.socket`. If the :const:`AF_UNIX` constant is not
  146. defined then this protocol is unsupported. More constants may be available
  147. depending on the system.
  148. .. data:: SOCK_STREAM
  149. SOCK_DGRAM
  150. SOCK_RAW
  151. SOCK_RDM
  152. SOCK_SEQPACKET
  153. These constants represent the socket types, used for the second argument to
  154. :func:`.socket`. More constants may be available depending on the system.
  155. (Only :const:`SOCK_STREAM` and :const:`SOCK_DGRAM` appear to be generally
  156. useful.)
  157. .. data:: SOCK_CLOEXEC
  158. SOCK_NONBLOCK
  159. These two constants, if defined, can be combined with the socket types and
  160. allow you to set some flags atomically (thus avoiding possible race
  161. conditions and the need for separate calls).
  162. .. seealso::
  163. `Secure File Descriptor Handling <http://udrepper.livejournal.com/20407.html>`_
  164. for a more thorough explanation.
  165. Availability: Linux >= 2.6.27.
  166. .. versionadded:: 3.2
  167. .. data:: SO_*
  168. SOMAXCONN
  169. MSG_*
  170. SOL_*
  171. SCM_*
  172. IPPROTO_*
  173. IPPORT_*
  174. INADDR_*
  175. IP_*
  176. IPV6_*
  177. EAI_*
  178. AI_*
  179. NI_*
  180. TCP_*
  181. Many constants of these forms, documented in the Unix documentation on sockets
  182. and/or the IP protocol, are also defined in the socket module. They are
  183. generally used in arguments to the :meth:`setsockopt` and :meth:`getsockopt`
  184. methods of socket objects. In most cases, only those symbols that are defined
  185. in the Unix header files are defined; for a few symbols, default values are
  186. provided.
  187. .. data:: AF_CAN
  188. PF_CAN
  189. SOL_CAN_*
  190. CAN_*
  191. Many constants of these forms, documented in the Linux documentation, are
  192. also defined in the socket module.
  193. Availability: Linux >= 2.6.25.
  194. .. versionadded:: 3.3
  195. .. data:: CAN_BCM
  196. CAN_BCM_*
  197. CAN_BCM, in the CAN protocol family, is the broadcast manager (BCM) protocol.
  198. Broadcast manager constants, documented in the Linux documentation, are also
  199. defined in the socket module.
  200. Availability: Linux >= 2.6.25.
  201. .. versionadded:: 3.4
  202. .. data:: AF_RDS
  203. PF_RDS
  204. SOL_RDS
  205. RDS_*
  206. Many constants of these forms, documented in the Linux documentation, are
  207. also defined in the socket module.
  208. Availability: Linux >= 2.6.30.
  209. .. versionadded:: 3.3
  210. .. data:: SIO_*
  211. RCVALL_*
  212. Constants for Windows' WSAIoctl(). The constants are used as arguments to the
  213. :meth:`~socket.socket.ioctl` method of socket objects.
  214. .. data:: TIPC_*
  215. TIPC related constants, matching the ones exported by the C socket API. See
  216. the TIPC documentation for more information.
  217. .. data:: AF_LINK
  218. Availability: BSD, OSX.
  219. .. versionadded:: 3.4
  220. .. data:: has_ipv6
  221. This constant contains a boolean value which indicates if IPv6 is supported on
  222. this platform.
  223. Functions
  224. ^^^^^^^^^
  225. Creating sockets
  226. ''''''''''''''''
  227. The following functions all create :ref:`socket objects <socket-objects>`.
  228. .. function:: socket(family=AF_INET, type=SOCK_STREAM, proto=0, fileno=None)
  229. Create a new socket using the given address family, socket type and protocol
  230. number. The address family should be :const:`AF_INET` (the default),
  231. :const:`AF_INET6`, :const:`AF_UNIX`, :const:`AF_CAN` or :const:`AF_RDS`. The
  232. socket type should be :const:`SOCK_STREAM` (the default),
  233. :const:`SOCK_DGRAM`, :const:`SOCK_RAW` or perhaps one of the other ``SOCK_``
  234. constants. The protocol number is usually zero and may be omitted or in the
  235. case where the address family is :const:`AF_CAN` the protocol should be one
  236. of :const:`CAN_RAW` or :const:`CAN_BCM`.
  237. The newly created socket is :ref:`non-inheritable <fd_inheritance>`.
  238. .. versionchanged:: 3.3
  239. The AF_CAN family was added.
  240. The AF_RDS family was added.
  241. .. versionchanged:: 3.4
  242. The CAN_BCM protocol was added.
  243. .. versionchanged:: 3.4
  244. The returned socket is now non-inheritable.
  245. .. function:: socketpair([family[, type[, proto]]])
  246. Build a pair of connected socket objects using the given address family, socket
  247. type, and protocol number. Address family, socket type, and protocol number are
  248. as for the :func:`.socket` function above. The default family is :const:`AF_UNIX`
  249. if defined on the platform; otherwise, the default is :const:`AF_INET`.
  250. The newly created sockets are :ref:`non-inheritable <fd_inheritance>`.
  251. .. versionchanged:: 3.2
  252. The returned socket objects now support the whole socket API, rather
  253. than a subset.
  254. .. versionchanged:: 3.4
  255. The returned sockets are now non-inheritable.
  256. .. versionchanged:: 3.5
  257. Windows support added.
  258. .. function:: create_connection(address[, timeout[, source_address]])
  259. Connect to a TCP service listening on the Internet *address* (a 2-tuple
  260. ``(host, port)``), and return the socket object. This is a higher-level
  261. function than :meth:`socket.connect`: if *host* is a non-numeric hostname,
  262. it will try to resolve it for both :data:`AF_INET` and :data:`AF_INET6`,
  263. and then try to connect to all possible addresses in turn until a
  264. connection succeeds. This makes it easy to write clients that are
  265. compatible to both IPv4 and IPv6.
  266. Passing the optional *timeout* parameter will set the timeout on the
  267. socket instance before attempting to connect. If no *timeout* is
  268. supplied, the global default timeout setting returned by
  269. :func:`getdefaulttimeout` is used.
  270. If supplied, *source_address* must be a 2-tuple ``(host, port)`` for the
  271. socket to bind to as its source address before connecting. If host or port
  272. are '' or 0 respectively the OS default behavior will be used.
  273. .. versionchanged:: 3.2
  274. *source_address* was added.
  275. .. versionchanged:: 3.2
  276. support for the :keyword:`with` statement was added.
  277. .. function:: fromfd(fd, family, type, proto=0)
  278. Duplicate the file descriptor *fd* (an integer as returned by a file object's
  279. :meth:`fileno` method) and build a socket object from the result. Address
  280. family, socket type and protocol number are as for the :func:`.socket` function
  281. above. The file descriptor should refer to a socket, but this is not checked ---
  282. subsequent operations on the object may fail if the file descriptor is invalid.
  283. This function is rarely needed, but can be used to get or set socket options on
  284. a socket passed to a program as standard input or output (such as a server
  285. started by the Unix inet daemon). The socket is assumed to be in blocking mode.
  286. The newly created socket is :ref:`non-inheritable <fd_inheritance>`.
  287. .. versionchanged:: 3.4
  288. The returned socket is now non-inheritable.
  289. .. function:: fromshare(data)
  290. Instantiate a socket from data obtained from the :meth:`socket.share`
  291. method. The socket is assumed to be in blocking mode.
  292. Availability: Windows.
  293. .. versionadded:: 3.3
  294. .. data:: SocketType
  295. This is a Python type object that represents the socket object type. It is the
  296. same as ``type(socket(...))``.
  297. Other functions
  298. '''''''''''''''
  299. The :mod:`socket` module also offers various network-related services:
  300. .. function:: getaddrinfo(host, port, family=0, type=0, proto=0, flags=0)
  301. Translate the *host*/*port* argument into a sequence of 5-tuples that contain
  302. all the necessary arguments for creating a socket connected to that service.
  303. *host* is a domain name, a string representation of an IPv4/v6 address
  304. or ``None``. *port* is a string service name such as ``'http'``, a numeric
  305. port number or ``None``. By passing ``None`` as the value of *host*
  306. and *port*, you can pass ``NULL`` to the underlying C API.
  307. The *family*, *type* and *proto* arguments can be optionally specified
  308. in order to narrow the list of addresses returned. Passing zero as a
  309. value for each of these arguments selects the full range of results.
  310. The *flags* argument can be one or several of the ``AI_*`` constants,
  311. and will influence how results are computed and returned.
  312. For example, :const:`AI_NUMERICHOST` will disable domain name resolution
  313. and will raise an error if *host* is a domain name.
  314. The function returns a list of 5-tuples with the following structure:
  315. ``(family, type, proto, canonname, sockaddr)``
  316. In these tuples, *family*, *type*, *proto* are all integers and are
  317. meant to be passed to the :func:`.socket` function. *canonname* will be
  318. a string representing the canonical name of the *host* if
  319. :const:`AI_CANONNAME` is part of the *flags* argument; else *canonname*
  320. will be empty. *sockaddr* is a tuple describing a socket address, whose
  321. format depends on the returned *family* (a ``(address, port)`` 2-tuple for
  322. :const:`AF_INET`, a ``(address, port, flow info, scope id)`` 4-tuple for
  323. :const:`AF_INET6`), and is meant to be passed to the :meth:`socket.connect`
  324. method.
  325. The following example fetches address information for a hypothetical TCP
  326. connection to ``www.python.org`` on port 80 (results may differ on your
  327. system if IPv6 isn't enabled)::
  328. >>> socket.getaddrinfo("www.python.org", 80, proto=socket.IPPROTO_TCP)
  329. [(2, 1, 6, '', ('82.94.164.162', 80)),
  330. (10, 1, 6, '', ('2001:888:2000:d::a2', 80, 0, 0))]
  331. .. versionchanged:: 3.2
  332. parameters can now be passed using keyword arguments.
  333. .. function:: getfqdn([name])
  334. Return a fully qualified domain name for *name*. If *name* is omitted or empty,
  335. it is interpreted as the local host. To find the fully qualified name, the
  336. hostname returned by :func:`gethostbyaddr` is checked, followed by aliases for the
  337. host, if available. The first name which includes a period is selected. In
  338. case no fully qualified domain name is available, the hostname as returned by
  339. :func:`gethostname` is returned.
  340. .. function:: gethostbyname(hostname)
  341. Translate a host name to IPv4 address format. The IPv4 address is returned as a
  342. string, such as ``'100.50.200.5'``. If the host name is an IPv4 address itself
  343. it is returned unchanged. See :func:`gethostbyname_ex` for a more complete
  344. interface. :func:`gethostbyname` does not support IPv6 name resolution, and
  345. :func:`getaddrinfo` should be used instead for IPv4/v6 dual stack support.
  346. .. function:: gethostbyname_ex(hostname)
  347. Translate a host name to IPv4 address format, extended interface. Return a
  348. triple ``(hostname, aliaslist, ipaddrlist)`` where *hostname* is the primary
  349. host name responding to the given *ip_address*, *aliaslist* is a (possibly
  350. empty) list of alternative host names for the same address, and *ipaddrlist* is
  351. a list of IPv4 addresses for the same interface on the same host (often but not
  352. always a single address). :func:`gethostbyname_ex` does not support IPv6 name
  353. resolution, and :func:`getaddrinfo` should be used instead for IPv4/v6 dual
  354. stack support.
  355. .. function:: gethostname()
  356. Return a string containing the hostname of the machine where the Python
  357. interpreter is currently executing.
  358. If you want to know the current machine's IP address, you may want to use
  359. ``gethostbyname(gethostname())``. This operation assumes that there is a
  360. valid address-to-host mapping for the host, and the assumption does not
  361. always hold.
  362. Note: :func:`gethostname` doesn't always return the fully qualified domain
  363. name; use ``getfqdn()`` (see above).
  364. .. function:: gethostbyaddr(ip_address)
  365. Return a triple ``(hostname, aliaslist, ipaddrlist)`` where *hostname* is the
  366. primary host name responding to the given *ip_address*, *aliaslist* is a
  367. (possibly empty) list of alternative host names for the same address, and
  368. *ipaddrlist* is a list of IPv4/v6 addresses for the same interface on the same
  369. host (most likely containing only a single address). To find the fully qualified
  370. domain name, use the function :func:`getfqdn`. :func:`gethostbyaddr` supports
  371. both IPv4 and IPv6.
  372. .. function:: getnameinfo(sockaddr, flags)
  373. Translate a socket address *sockaddr* into a 2-tuple ``(host, port)``. Depending
  374. on the settings of *flags*, the result can contain a fully-qualified domain name
  375. or numeric address representation in *host*. Similarly, *port* can contain a
  376. string port name or a numeric port number.
  377. .. function:: getprotobyname(protocolname)
  378. Translate an Internet protocol name (for example, ``'icmp'``) to a constant
  379. suitable for passing as the (optional) third argument to the :func:`.socket`
  380. function. This is usually only needed for sockets opened in "raw" mode
  381. (:const:`SOCK_RAW`); for the normal socket modes, the correct protocol is chosen
  382. automatically if the protocol is omitted or zero.
  383. .. function:: getservbyname(servicename[, protocolname])
  384. Translate an Internet service name and protocol name to a port number for that
  385. service. The optional protocol name, if given, should be ``'tcp'`` or
  386. ``'udp'``, otherwise any protocol will match.
  387. .. function:: getservbyport(port[, protocolname])
  388. Translate an Internet port number and protocol name to a service name for that
  389. service. The optional protocol name, if given, should be ``'tcp'`` or
  390. ``'udp'``, otherwise any protocol will match.
  391. .. function:: ntohl(x)
  392. Convert 32-bit positive integers from network to host byte order. On machines
  393. where the host byte order is the same as network byte order, this is a no-op;
  394. otherwise, it performs a 4-byte swap operation.
  395. .. function:: ntohs(x)
  396. Convert 16-bit positive integers from network to host byte order. On machines
  397. where the host byte order is the same as network byte order, this is a no-op;
  398. otherwise, it performs a 2-byte swap operation.
  399. .. function:: htonl(x)
  400. Convert 32-bit positive integers from host to network byte order. On machines
  401. where the host byte order is the same as network byte order, this is a no-op;
  402. otherwise, it performs a 4-byte swap operation.
  403. .. function:: htons(x)
  404. Convert 16-bit positive integers from host to network byte order. On machines
  405. where the host byte order is the same as network byte order, this is a no-op;
  406. otherwise, it performs a 2-byte swap operation.
  407. .. function:: inet_aton(ip_string)
  408. Convert an IPv4 address from dotted-quad string format (for example,
  409. '123.45.67.89') to 32-bit packed binary format, as a bytes object four characters in
  410. length. This is useful when conversing with a program that uses the standard C
  411. library and needs objects of type :c:type:`struct in_addr`, which is the C type
  412. for the 32-bit packed binary this function returns.
  413. :func:`inet_aton` also accepts strings with less than three dots; see the
  414. Unix manual page :manpage:`inet(3)` for details.
  415. If the IPv4 address string passed to this function is invalid,
  416. :exc:`OSError` will be raised. Note that exactly what is valid depends on
  417. the underlying C implementation of :c:func:`inet_aton`.
  418. :func:`inet_aton` does not support IPv6, and :func:`inet_pton` should be used
  419. instead for IPv4/v6 dual stack support.
  420. .. function:: inet_ntoa(packed_ip)
  421. Convert a 32-bit packed IPv4 address (a bytes object four characters in
  422. length) to its standard dotted-quad string representation (for example,
  423. '123.45.67.89'). This is useful when conversing with a program that uses the
  424. standard C library and needs objects of type :c:type:`struct in_addr`, which
  425. is the C type for the 32-bit packed binary data this function takes as an
  426. argument.
  427. If the byte sequence passed to this function is not exactly 4 bytes in
  428. length, :exc:`OSError` will be raised. :func:`inet_ntoa` does not
  429. support IPv6, and :func:`inet_ntop` should be used instead for IPv4/v6 dual
  430. stack support.
  431. .. function:: inet_pton(address_family, ip_string)
  432. Convert an IP address from its family-specific string format to a packed,
  433. binary format. :func:`inet_pton` is useful when a library or network protocol
  434. calls for an object of type :c:type:`struct in_addr` (similar to
  435. :func:`inet_aton`) or :c:type:`struct in6_addr`.
  436. Supported values for *address_family* are currently :const:`AF_INET` and
  437. :const:`AF_INET6`. If the IP address string *ip_string* is invalid,
  438. :exc:`OSError` will be raised. Note that exactly what is valid depends on
  439. both the value of *address_family* and the underlying implementation of
  440. :c:func:`inet_pton`.
  441. Availability: Unix (maybe not all platforms), Windows.
  442. .. versionchanged:: 3.4
  443. Windows support added
  444. .. function:: inet_ntop(address_family, packed_ip)
  445. Convert a packed IP address (a bytes object of some number of characters) to its
  446. standard, family-specific string representation (for example, ``'7.10.0.5'`` or
  447. ``'5aef:2b::8'``). :func:`inet_ntop` is useful when a library or network protocol
  448. returns an object of type :c:type:`struct in_addr` (similar to :func:`inet_ntoa`)
  449. or :c:type:`struct in6_addr`.
  450. Supported values for *address_family* are currently :const:`AF_INET` and
  451. :const:`AF_INET6`. If the string *packed_ip* is not the correct length for the
  452. specified address family, :exc:`ValueError` will be raised. A
  453. :exc:`OSError` is raised for errors from the call to :func:`inet_ntop`.
  454. Availability: Unix (maybe not all platforms), Windows.
  455. .. versionchanged:: 3.4
  456. Windows support added
  457. ..
  458. XXX: Are sendmsg(), recvmsg() and CMSG_*() available on any
  459. non-Unix platforms? The old (obsolete?) 4.2BSD form of the
  460. interface, in which struct msghdr has no msg_control or
  461. msg_controllen members, is not currently supported.
  462. .. function:: CMSG_LEN(length)
  463. Return the total length, without trailing padding, of an ancillary
  464. data item with associated data of the given *length*. This value
  465. can often be used as the buffer size for :meth:`~socket.recvmsg` to
  466. receive a single item of ancillary data, but :rfc:`3542` requires
  467. portable applications to use :func:`CMSG_SPACE` and thus include
  468. space for padding, even when the item will be the last in the
  469. buffer. Raises :exc:`OverflowError` if *length* is outside the
  470. permissible range of values.
  471. Availability: most Unix platforms, possibly others.
  472. .. versionadded:: 3.3
  473. .. function:: CMSG_SPACE(length)
  474. Return the buffer size needed for :meth:`~socket.recvmsg` to
  475. receive an ancillary data item with associated data of the given
  476. *length*, along with any trailing padding. The buffer space needed
  477. to receive multiple items is the sum of the :func:`CMSG_SPACE`
  478. values for their associated data lengths. Raises
  479. :exc:`OverflowError` if *length* is outside the permissible range
  480. of values.
  481. Note that some systems might support ancillary data without
  482. providing this function. Also note that setting the buffer size
  483. using the results of this function may not precisely limit the
  484. amount of ancillary data that can be received, since additional
  485. data may be able to fit into the padding area.
  486. Availability: most Unix platforms, possibly others.
  487. .. versionadded:: 3.3
  488. .. function:: getdefaulttimeout()
  489. Return the default timeout in seconds (float) for new socket objects. A value
  490. of ``None`` indicates that new socket objects have no timeout. When the socket
  491. module is first imported, the default is ``None``.
  492. .. function:: setdefaulttimeout(timeout)
  493. Set the default timeout in seconds (float) for new socket objects. When
  494. the socket module is first imported, the default is ``None``. See
  495. :meth:`~socket.settimeout` for possible values and their respective
  496. meanings.
  497. .. function:: sethostname(name)
  498. Set the machine's hostname to *name*. This will raise a
  499. :exc:`OSError` if you don't have enough rights.
  500. Availability: Unix.
  501. .. versionadded:: 3.3
  502. .. function:: if_nameindex()
  503. Return a list of network interface information
  504. (index int, name string) tuples.
  505. :exc:`OSError` if the system call fails.
  506. Availability: Unix.
  507. .. versionadded:: 3.3
  508. .. function:: if_nametoindex(if_name)
  509. Return a network interface index number corresponding to an
  510. interface name.
  511. :exc:`OSError` if no interface with the given name exists.
  512. Availability: Unix.
  513. .. versionadded:: 3.3
  514. .. function:: if_indextoname(if_index)
  515. Return a network interface name corresponding to a
  516. interface index number.
  517. :exc:`OSError` if no interface with the given index exists.
  518. Availability: Unix.
  519. .. versionadded:: 3.3
  520. .. _socket-objects:
  521. Socket Objects
  522. --------------
  523. Socket objects have the following methods. Except for
  524. :meth:`~socket.makefile`, these correspond to Unix system calls applicable
  525. to sockets.
  526. .. method:: socket.accept()
  527. Accept a connection. The socket must be bound to an address and listening for
  528. connections. The return value is a pair ``(conn, address)`` where *conn* is a
  529. *new* socket object usable to send and receive data on the connection, and
  530. *address* is the address bound to the socket on the other end of the connection.
  531. The newly created socket is :ref:`non-inheritable <fd_inheritance>`.
  532. .. versionchanged:: 3.4
  533. The socket is now non-inheritable.
  534. .. method:: socket.bind(address)
  535. Bind the socket to *address*. The socket must not already be bound. (The format
  536. of *address* depends on the address family --- see above.)
  537. .. method:: socket.close()
  538. Mark the socket closed. The underlying system resource (e.g. a file
  539. descriptor) is also closed when all file objects from :meth:`makefile()`
  540. are closed. Once that happens, all future operations on the socket
  541. object will fail. The remote end will receive no more data (after
  542. queued data is flushed).
  543. Sockets are automatically closed when they are garbage-collected, but
  544. it is recommended to :meth:`close` them explicitly, or to use a
  545. :keyword:`with` statement around them.
  546. .. note::
  547. :meth:`close()` releases the resource associated with a connection but
  548. does not necessarily close the connection immediately. If you want
  549. to close the connection in a timely fashion, call :meth:`shutdown()`
  550. before :meth:`close()`.
  551. .. method:: socket.connect(address)
  552. Connect to a remote socket at *address*. (The format of *address* depends on the
  553. address family --- see above.)
  554. .. method:: socket.connect_ex(address)
  555. Like ``connect(address)``, but return an error indicator instead of raising an
  556. exception for errors returned by the C-level :c:func:`connect` call (other
  557. problems, such as "host not found," can still raise exceptions). The error
  558. indicator is ``0`` if the operation succeeded, otherwise the value of the
  559. :c:data:`errno` variable. This is useful to support, for example, asynchronous
  560. connects.
  561. .. method:: socket.detach()
  562. Put the socket object into closed state without actually closing the
  563. underlying file descriptor. The file descriptor is returned, and can
  564. be reused for other purposes.
  565. .. versionadded:: 3.2
  566. .. method:: socket.dup()
  567. Duplicate the socket.
  568. The newly created socket is :ref:`non-inheritable <fd_inheritance>`.
  569. .. versionchanged:: 3.4
  570. The socket is now non-inheritable.
  571. .. method:: socket.fileno()
  572. Return the socket's file descriptor (a small integer). This is useful with
  573. :func:`select.select`.
  574. Under Windows the small integer returned by this method cannot be used where a
  575. file descriptor can be used (such as :func:`os.fdopen`). Unix does not have
  576. this limitation.
  577. .. method:: socket.get_inheritable()
  578. Get the :ref:`inheritable flag <fd_inheritance>` of the socket's file
  579. descriptor or socket's handle: ``True`` if the socket can be inherited in
  580. child processes, ``False`` if it cannot.
  581. .. versionadded:: 3.4
  582. .. method:: socket.getpeername()
  583. Return the remote address to which the socket is connected. This is useful to
  584. find out the port number of a remote IPv4/v6 socket, for instance. (The format
  585. of the address returned depends on the address family --- see above.) On some
  586. systems this function is not supported.
  587. .. method:: socket.getsockname()
  588. Return the socket's own address. This is useful to find out the port number of
  589. an IPv4/v6 socket, for instance. (The format of the address returned depends on
  590. the address family --- see above.)
  591. .. method:: socket.getsockopt(level, optname[, buflen])
  592. Return the value of the given socket option (see the Unix man page
  593. :manpage:`getsockopt(2)`). The needed symbolic constants (:const:`SO_\*` etc.)
  594. are defined in this module. If *buflen* is absent, an integer option is assumed
  595. and its integer value is returned by the function. If *buflen* is present, it
  596. specifies the maximum length of the buffer used to receive the option in, and
  597. this buffer is returned as a bytes object. It is up to the caller to decode the
  598. contents of the buffer (see the optional built-in module :mod:`struct` for a way
  599. to decode C structures encoded as byte strings).
  600. .. method:: socket.gettimeout()
  601. Return the timeout in seconds (float) associated with socket operations,
  602. or ``None`` if no timeout is set. This reflects the last call to
  603. :meth:`setblocking` or :meth:`settimeout`.
  604. .. method:: socket.ioctl(control, option)
  605. :platform: Windows
  606. The :meth:`ioctl` method is a limited interface to the WSAIoctl system
  607. interface. Please refer to the `Win32 documentation
  608. <http://msdn.microsoft.com/en-us/library/ms741621%28VS.85%29.aspx>`_ for more
  609. information.
  610. On other platforms, the generic :func:`fcntl.fcntl` and :func:`fcntl.ioctl`
  611. functions may be used; they accept a socket object as their first argument.
  612. .. method:: socket.listen([backlog])
  613. Enable a server to accept connections. If *backlog* is specified, it must
  614. be at least 0 (if it is lower, it is set to 0); it specifies the number of
  615. unaccepted connections that the system will allow before refusing new
  616. connections. If not specified, a default reasonable value is chosen.
  617. .. versionchanged:: 3.5
  618. The *backlog* parameter is now optional.
  619. .. method:: socket.makefile(mode='r', buffering=None, *, encoding=None, \
  620. errors=None, newline=None)
  621. .. index:: single: I/O control; buffering
  622. Return a :term:`file object` associated with the socket. The exact returned
  623. type depends on the arguments given to :meth:`makefile`. These arguments are
  624. interpreted the same way as by the built-in :func:`open` function.
  625. The socket must be in blocking mode; it can have a timeout, but the file
  626. object's internal buffer may end up in a inconsistent state if a timeout
  627. occurs.
  628. Closing the file object returned by :meth:`makefile` won't close the
  629. original socket unless all other file objects have been closed and
  630. :meth:`socket.close` has been called on the socket object.
  631. .. note::
  632. On Windows, the file-like object created by :meth:`makefile` cannot be
  633. used where a file object with a file descriptor is expected, such as the
  634. stream arguments of :meth:`subprocess.Popen`.
  635. .. method:: socket.recv(bufsize[, flags])
  636. Receive data from the socket. The return value is a bytes object representing the
  637. data received. The maximum amount of data to be received at once is specified
  638. by *bufsize*. See the Unix manual page :manpage:`recv(2)` for the meaning of
  639. the optional argument *flags*; it defaults to zero.
  640. .. note::
  641. For best match with hardware and network realities, the value of *bufsize*
  642. should be a relatively small power of 2, for example, 4096.
  643. .. method:: socket.recvfrom(bufsize[, flags])
  644. Receive data from the socket. The return value is a pair ``(bytes, address)``
  645. where *bytes* is a bytes object representing the data received and *address* is the
  646. address of the socket sending the data. See the Unix manual page
  647. :manpage:`recv(2)` for the meaning of the optional argument *flags*; it defaults
  648. to zero. (The format of *address* depends on the address family --- see above.)
  649. .. method:: socket.recvmsg(bufsize[, ancbufsize[, flags]])
  650. Receive normal data (up to *bufsize* bytes) and ancillary data from
  651. the socket. The *ancbufsize* argument sets the size in bytes of
  652. the internal buffer used to receive the ancillary data; it defaults
  653. to 0, meaning that no ancillary data will be received. Appropriate
  654. buffer sizes for ancillary data can be calculated using
  655. :func:`CMSG_SPACE` or :func:`CMSG_LEN`, and items which do not fit
  656. into the buffer might be truncated or discarded. The *flags*
  657. argument defaults to 0 and has the same meaning as for
  658. :meth:`recv`.
  659. The return value is a 4-tuple: ``(data, ancdata, msg_flags,
  660. address)``. The *data* item is a :class:`bytes` object holding the
  661. non-ancillary data received. The *ancdata* item is a list of zero
  662. or more tuples ``(cmsg_level, cmsg_type, cmsg_data)`` representing
  663. the ancillary data (control messages) received: *cmsg_level* and
  664. *cmsg_type* are integers specifying the protocol level and
  665. protocol-specific type respectively, and *cmsg_data* is a
  666. :class:`bytes` object holding the associated data. The *msg_flags*
  667. item is the bitwise OR of various flags indicating conditions on
  668. the received message; see your system documentation for details.
  669. If the receiving socket is unconnected, *address* is the address of
  670. the sending socket, if available; otherwise, its value is
  671. unspecified.
  672. On some systems, :meth:`sendmsg` and :meth:`recvmsg` can be used to
  673. pass file descriptors between processes over an :const:`AF_UNIX`
  674. socket. When this facility is used (it is often restricted to
  675. :const:`SOCK_STREAM` sockets), :meth:`recvmsg` will return, in its
  676. ancillary data, items of the form ``(socket.SOL_SOCKET,
  677. socket.SCM_RIGHTS, fds)``, where *fds* is a :class:`bytes` object
  678. representing the new file descriptors as a binary array of the
  679. native C :c:type:`int` type. If :meth:`recvmsg` raises an
  680. exception after the system call returns, it will first attempt to
  681. close any file descriptors received via this mechanism.
  682. Some systems do not indicate the truncated length of ancillary data
  683. items which have been only partially received. If an item appears
  684. to extend beyond the end of the buffer, :meth:`recvmsg` will issue
  685. a :exc:`RuntimeWarning`, and will return the part of it which is
  686. inside the buffer provided it has not been truncated before the
  687. start of its associated data.
  688. On systems which support the :const:`SCM_RIGHTS` mechanism, the
  689. following function will receive up to *maxfds* file descriptors,
  690. returning the message data and a list containing the descriptors
  691. (while ignoring unexpected conditions such as unrelated control
  692. messages being received). See also :meth:`sendmsg`. ::
  693. import socket, array
  694. def recv_fds(sock, msglen, maxfds):
  695. fds = array.array("i") # Array of ints
  696. msg, ancdata, flags, addr = sock.recvmsg(msglen, socket.CMSG_LEN(maxfds * fds.itemsize))
  697. for cmsg_level, cmsg_type, cmsg_data in ancdata:
  698. if (cmsg_level == socket.SOL_SOCKET and cmsg_type == socket.SCM_RIGHTS):
  699. # Append data, ignoring any truncated integers at the end.
  700. fds.fromstring(cmsg_data[:len(cmsg_data) - (len(cmsg_data) % fds.itemsize)])
  701. return msg, list(fds)
  702. Availability: most Unix platforms, possibly others.
  703. .. versionadded:: 3.3
  704. .. method:: socket.recvmsg_into(buffers[, ancbufsize[, flags]])
  705. Receive normal data and ancillary data from the socket, behaving as
  706. :meth:`recvmsg` would, but scatter the non-ancillary data into a
  707. series of buffers instead of returning a new bytes object. The
  708. *buffers* argument must be an iterable of objects that export
  709. writable buffers (e.g. :class:`bytearray` objects); these will be
  710. filled with successive chunks of the non-ancillary data until it
  711. has all been written or there are no more buffers. The operating
  712. system may set a limit (:func:`~os.sysconf` value ``SC_IOV_MAX``)
  713. on the number of buffers that can be used. The *ancbufsize* and
  714. *flags* arguments have the same meaning as for :meth:`recvmsg`.
  715. The return value is a 4-tuple: ``(nbytes, ancdata, msg_flags,
  716. address)``, where *nbytes* is the total number of bytes of
  717. non-ancillary data written into the buffers, and *ancdata*,
  718. *msg_flags* and *address* are the same as for :meth:`recvmsg`.
  719. Example::
  720. >>> import socket
  721. >>> s1, s2 = socket.socketpair()
  722. >>> b1 = bytearray(b'----')
  723. >>> b2 = bytearray(b'0123456789')
  724. >>> b3 = bytearray(b'--------------')
  725. >>> s1.send(b'Mary had a little lamb')
  726. 22
  727. >>> s2.recvmsg_into([b1, memoryview(b2)[2:9], b3])
  728. (22, [], 0, None)
  729. >>> [b1, b2, b3]
  730. [bytearray(b'Mary'), bytearray(b'01 had a 9'), bytearray(b'little lamb---')]
  731. Availability: most Unix platforms, possibly others.
  732. .. versionadded:: 3.3
  733. .. method:: socket.recvfrom_into(buffer[, nbytes[, flags]])
  734. Receive data from the socket, writing it into *buffer* instead of creating a
  735. new bytestring. The return value is a pair ``(nbytes, address)`` where *nbytes* is
  736. the number of bytes received and *address* is the address of the socket sending
  737. the data. See the Unix manual page :manpage:`recv(2)` for the meaning of the
  738. optional argument *flags*; it defaults to zero. (The format of *address*
  739. depends on the address family --- see above.)
  740. .. method:: socket.recv_into(buffer[, nbytes[, flags]])
  741. Receive up to *nbytes* bytes from the socket, storing the data into a buffer
  742. rather than creating a new bytestring. If *nbytes* is not specified (or 0),
  743. receive up to the size available in the given buffer. Returns the number of
  744. bytes received. See the Unix manual page :manpage:`recv(2)` for the meaning
  745. of the optional argument *flags*; it defaults to zero.
  746. .. method:: socket.send(bytes[, flags])
  747. Send data to the socket. The socket must be connected to a remote socket. The
  748. optional *flags* argument has the same meaning as for :meth:`recv` above.
  749. Returns the number of bytes sent. Applications are responsible for checking that
  750. all data has been sent; if only some of the data was transmitted, the
  751. application needs to attempt delivery of the remaining data. For further
  752. information on this topic, consult the :ref:`socket-howto`.
  753. .. method:: socket.sendall(bytes[, flags])
  754. Send data to the socket. The socket must be connected to a remote socket. The
  755. optional *flags* argument has the same meaning as for :meth:`recv` above.
  756. Unlike :meth:`send`, this method continues to send data from *bytes* until
  757. either all data has been sent or an error occurs. ``None`` is returned on
  758. success. On error, an exception is raised, and there is no way to determine how
  759. much data, if any, was successfully sent.
  760. .. method:: socket.sendto(bytes, address)
  761. socket.sendto(bytes, flags, address)
  762. Send data to the socket. The socket should not be connected to a remote socket,
  763. since the destination socket is specified by *address*. The optional *flags*
  764. argument has the same meaning as for :meth:`recv` above. Return the number of
  765. bytes sent. (The format of *address* depends on the address family --- see
  766. above.)
  767. .. method:: socket.sendmsg(buffers[, ancdata[, flags[, address]]])
  768. Send normal and ancillary data to the socket, gathering the
  769. non-ancillary data from a series of buffers and concatenating it
  770. into a single message. The *buffers* argument specifies the
  771. non-ancillary data as an iterable of buffer-compatible objects
  772. (e.g. :class:`bytes` objects); the operating system may set a limit
  773. (:func:`~os.sysconf` value ``SC_IOV_MAX``) on the number of buffers
  774. that can be used. The *ancdata* argument specifies the ancillary
  775. data (control messages) as an iterable of zero or more tuples
  776. ``(cmsg_level, cmsg_type, cmsg_data)``, where *cmsg_level* and
  777. *cmsg_type* are integers specifying the protocol level and
  778. protocol-specific type respectively, and *cmsg_data* is a
  779. buffer-compatible object holding the associated data. Note that
  780. some systems (in particular, systems without :func:`CMSG_SPACE`)
  781. might support sending only one control message per call. The
  782. *flags* argument defaults to 0 and has the same meaning as for
  783. :meth:`send`. If *address* is supplied and not ``None``, it sets a
  784. destination address for the message. The return value is the
  785. number of bytes of non-ancillary data sent.
  786. The following function sends the list of file descriptors *fds*
  787. over an :const:`AF_UNIX` socket, on systems which support the
  788. :const:`SCM_RIGHTS` mechanism. See also :meth:`recvmsg`. ::
  789. import socket, array
  790. def send_fds(sock, msg, fds):
  791. return sock.sendmsg([msg], [(socket.SOL_SOCKET, socket.SCM_RIGHTS, array.array("i", fds))])
  792. Availability: most Unix platforms, possibly others.
  793. .. versionadded:: 3.3
  794. .. method:: socket.sendfile(file, offset=0, count=None)
  795. Send a file until EOF is reached by using high-performance
  796. :mod:`os.sendfile` and return the total number of bytes which were sent.
  797. *file* must be a regular file object opened in binary mode. If
  798. :mod:`os.sendfile` is not available (e.g. Windows) or *file* is not a
  799. regular file :meth:`send` will be used instead. *offset* tells from where to
  800. start reading the file. If specified, *count* is the total number of bytes
  801. to transmit as opposed to sending the file until EOF is reached. File
  802. position is updated on return or also in case of error in which case
  803. :meth:`file.tell() <io.IOBase.tell>` can be used to figure out the number of
  804. bytes which were sent. The socket must be of :const:`SOCK_STREAM` type. Non-
  805. blocking sockets are not supported.
  806. .. versionadded:: 3.5
  807. .. method:: socket.set_inheritable(inheritable)
  808. Set the :ref:`inheritable flag <fd_inheritance>` of the socket's file
  809. descriptor or socket's handle.
  810. .. versionadded:: 3.4
  811. .. method:: socket.setblocking(flag)
  812. Set blocking or non-blocking mode of the socket: if *flag* is false, the
  813. socket is set to non-blocking, else to blocking mode.
  814. This method is a shorthand for certain :meth:`~socket.settimeout` calls:
  815. * ``sock.setblocking(True)`` is equivalent to ``sock.settimeout(None)``
  816. * ``sock.setblocking(False)`` is equivalent to ``sock.settimeout(0.0)``
  817. .. method:: socket.settimeout(value)
  818. Set a timeout on blocking socket operations. The *value* argument can be a
  819. nonnegative floating point number expressing seconds, or ``None``.
  820. If a non-zero value is given, subsequent socket operations will raise a
  821. :exc:`timeout` exception if the timeout period *value* has elapsed before
  822. the operation has completed. If zero is given, the socket is put in
  823. non-blocking mode. If ``None`` is given, the socket is put in blocking mode.
  824. For further information, please consult the :ref:`notes on socket timeouts <socket-timeouts>`.
  825. .. method:: socket.setsockopt(level, optname, value)
  826. .. index:: module: struct
  827. Set the value of the given socket option (see the Unix manual page
  828. :manpage:`setsockopt(2)`). The needed symbolic constants are defined in the
  829. :mod:`socket` module (:const:`SO_\*` etc.). The value can be an integer or a
  830. bytes object representing a buffer. In the latter case it is up to the caller to
  831. ensure that the bytestring contains the proper bits (see the optional built-in
  832. module :mod:`struct` for a way to encode C structures as bytestrings).
  833. .. method:: socket.shutdown(how)
  834. Shut down one or both halves of the connection. If *how* is :const:`SHUT_RD`,
  835. further receives are disallowed. If *how* is :const:`SHUT_WR`, further sends
  836. are disallowed. If *how* is :const:`SHUT_RDWR`, further sends and receives are
  837. disallowed.
  838. .. method:: socket.share(process_id)
  839. Duplicate a socket and prepare it for sharing with a target process. The
  840. target process must be provided with *process_id*. The resulting bytes object
  841. can then be passed to the target process using some form of interprocess
  842. communication and the socket can be recreated there using :func:`fromshare`.
  843. Once this method has been called, it is safe to close the socket since
  844. the operating system has already duplicated it for the target process.
  845. Availability: Windows.
  846. .. versionadded:: 3.3
  847. Note that there are no methods :meth:`read` or :meth:`write`; use
  848. :meth:`~socket.recv` and :meth:`~socket.send` without *flags* argument instead.
  849. Socket objects also have these (read-only) attributes that correspond to the
  850. values given to the :class:`socket` constructor.
  851. .. attribute:: socket.family
  852. The socket family.
  853. .. attribute:: socket.type
  854. The socket type.
  855. .. attribute:: socket.proto
  856. The socket protocol.
  857. .. _socket-timeouts:
  858. Notes on socket timeouts
  859. ------------------------
  860. A socket object can be in one of three modes: blocking, non-blocking, or
  861. timeout. Sockets are by default always created in blocking mode, but this
  862. can be changed by calling :func:`setdefaulttimeout`.
  863. * In *blocking mode*, operations block until complete or the system returns
  864. an error (such as connection timed out).
  865. * In *non-blocking mode*, operations fail (with an error that is unfortunately
  866. system-dependent) if they cannot be completed immediately: functions from the
  867. :mod:`select` can be used to know when and whether a socket is available for
  868. reading or writing.
  869. * In *timeout mode*, operations fail if they cannot be completed within the
  870. timeout specified for the socket (they raise a :exc:`timeout` exception)
  871. or if the system returns an error.
  872. .. note::
  873. At the operating system level, sockets in *timeout mode* are internally set
  874. in non-blocking mode. Also, the blocking and timeout modes are shared between
  875. file descriptors and socket objects that refer to the same network endpoint.
  876. This implementation detail can have visible consequences if e.g. you decide
  877. to use the :meth:`~socket.fileno()` of a socket.
  878. Timeouts and the ``connect`` method
  879. ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  880. The :meth:`~socket.connect` operation is also subject to the timeout
  881. setting, and in general it is recommended to call :meth:`~socket.settimeout`
  882. before calling :meth:`~socket.connect` or pass a timeout parameter to
  883. :meth:`create_connection`. However, the system network stack may also
  884. return a connection timeout error of its own regardless of any Python socket
  885. timeout setting.
  886. Timeouts and the ``accept`` method
  887. ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  888. If :func:`getdefaulttimeout` is not :const:`None`, sockets returned by
  889. the :meth:`~socket.accept` method inherit that timeout. Otherwise, the
  890. behaviour depends on settings of the listening socket:
  891. * if the listening socket is in *blocking mode* or in *timeout mode*,
  892. the socket returned by :meth:`~socket.accept` is in *blocking mode*;
  893. * if the listening socket is in *non-blocking mode*, whether the socket
  894. returned by :meth:`~socket.accept` is in blocking or non-blocking mode
  895. is operating system-dependent. If you want to ensure cross-platform
  896. behaviour, it is recommended you manually override this setting.
  897. .. _socket-example:
  898. Example
  899. -------
  900. Here are four minimal example programs using the TCP/IP protocol: a server that
  901. echoes all data that it receives back (servicing only one client), and a client
  902. using it. Note that a server must perform the sequence :func:`.socket`,
  903. :meth:`~socket.bind`, :meth:`~socket.listen`, :meth:`~socket.accept` (possibly
  904. repeating the :meth:`~socket.accept` to service more than one client), while a
  905. client only needs the sequence :func:`.socket`, :meth:`~socket.connect`. Also
  906. note that the server does not :meth:`~socket.sendall`/:meth:`~socket.recv` on
  907. the socket it is listening on but on the new socket returned by
  908. :meth:`~socket.accept`.
  909. The first two examples support IPv4 only. ::
  910. # Echo server program
  911. import socket
  912. HOST = '' # Symbolic name meaning all available interfaces
  913. PORT = 50007 # Arbitrary non-privileged port
  914. s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
  915. s.bind((HOST, PORT))
  916. s.listen(1)
  917. conn, addr = s.accept()
  918. print('Connected by', addr)
  919. while True:
  920. data = conn.recv(1024)
  921. if not data: break
  922. conn.sendall(data)
  923. conn.close()
  924. ::
  925. # Echo client program
  926. import socket
  927. HOST = 'daring.cwi.nl' # The remote host
  928. PORT = 50007 # The same port as used by the server
  929. s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
  930. s.connect((HOST, PORT))
  931. s.sendall(b'Hello, world')
  932. data = s.recv(1024)
  933. s.close()
  934. print('Received', repr(data))
  935. The next two examples are identical to the above two, but support both IPv4 and
  936. IPv6. The server side will listen to the first address family available (it
  937. should listen to both instead). On most of IPv6-ready systems, IPv6 will take
  938. precedence and the server may not accept IPv4 traffic. The client side will try
  939. to connect to the all addresses returned as a result of the name resolution, and
  940. sends traffic to the first one connected successfully. ::
  941. # Echo server program
  942. import socket
  943. import sys
  944. HOST = None # Symbolic name meaning all available interfaces
  945. PORT = 50007 # Arbitrary non-privileged port
  946. s = None
  947. for res in socket.getaddrinfo(HOST, PORT, socket.AF_UNSPEC,
  948. socket.SOCK_STREAM, 0, socket.AI_PASSIVE):
  949. af, socktype, proto, canonname, sa = res
  950. try:
  951. s = socket.socket(af, socktype, proto)
  952. except OSError as msg:
  953. s = None
  954. continue
  955. try:
  956. s.bind(sa)
  957. s.listen(1)
  958. except OSError as msg:
  959. s.close()
  960. s = None
  961. continue
  962. break
  963. if s is None:
  964. print('could not open socket')
  965. sys.exit(1)
  966. conn, addr = s.accept()
  967. print('Connected by', addr)
  968. while True:
  969. data = conn.recv(1024)
  970. if not data: break
  971. conn.send(data)
  972. conn.close()
  973. ::
  974. # Echo client program
  975. import socket
  976. import sys
  977. HOST = 'daring.cwi.nl' # The remote host
  978. PORT = 50007 # The same port as used by the server
  979. s = None
  980. for res in socket.getaddrinfo(HOST, PORT, socket.AF_UNSPEC, socket.SOCK_STREAM):
  981. af, socktype, proto, canonname, sa = res
  982. try:
  983. s = socket.socket(af, socktype, proto)
  984. except OSError as msg:
  985. s = None
  986. continue
  987. try:
  988. s.connect(sa)
  989. except OSError as msg:
  990. s.close()
  991. s = None
  992. continue
  993. break
  994. if s is None:
  995. print('could not open socket')
  996. sys.exit(1)
  997. s.sendall(b'Hello, world')
  998. data = s.recv(1024)
  999. s.close()
  1000. print('Received', repr(data))
  1001. The next example shows how to write a very simple network sniffer with raw
  1002. sockets on Windows. The example requires administrator privileges to modify
  1003. the interface::
  1004. import socket
  1005. # the public network interface
  1006. HOST = socket.gethostbyname(socket.gethostname())
  1007. # create a raw socket and bind it to the public interface
  1008. s = socket.socket(socket.AF_INET, socket.SOCK_RAW, socket.IPPROTO_IP)
  1009. s.bind((HOST, 0))
  1010. # Include IP headers
  1011. s.setsockopt(socket.IPPROTO_IP, socket.IP_HDRINCL, 1)
  1012. # receive all packages
  1013. s.ioctl(socket.SIO_RCVALL, socket.RCVALL_ON)
  1014. # receive a package
  1015. print(s.recvfrom(65565))
  1016. # disabled promiscuous mode
  1017. s.ioctl(socket.SIO_RCVALL, socket.RCVALL_OFF)
  1018. The last example shows how to use the socket interface to communicate to a CAN
  1019. network using the raw socket protocol. To use CAN with the broadcast
  1020. manager protocol instead, open a socket with::
  1021. socket.socket(socket.AF_CAN, socket.SOCK_DGRAM, socket.CAN_BCM)
  1022. After binding (:const:`CAN_RAW`) or connecting (:const:`CAN_BCM`) the socket, you
  1023. can use the :meth:`socket.send`, and the :meth:`socket.recv` operations (and
  1024. their counterparts) on the socket object as usual.
  1025. This example might require special privileges::
  1026. import socket
  1027. import struct
  1028. # CAN frame packing/unpacking (see 'struct can_frame' in <linux/can.h>)
  1029. can_frame_fmt = "=IB3x8s"
  1030. can_frame_size = struct.calcsize(can_frame_fmt)
  1031. def build_can_frame(can_id, data):
  1032. can_dlc = len(data)
  1033. data = data.ljust(8, b'\x00')
  1034. return struct.pack(can_frame_fmt, can_id, can_dlc, data)
  1035. def dissect_can_frame(frame):
  1036. can_id, can_dlc, data = struct.unpack(can_frame_fmt, frame)
  1037. return (can_id, can_dlc, data[:can_dlc])
  1038. # create a raw socket and bind it to the 'vcan0' interface
  1039. s = socket.socket(socket.AF_CAN, socket.SOCK_RAW, socket.CAN_RAW)
  1040. s.bind(('vcan0',))
  1041. while True:
  1042. cf, addr = s.recvfrom(can_frame_size)
  1043. print('Received: can_id=%x, can_dlc=%x, data=%s' % dissect_can_frame(cf))
  1044. try:
  1045. s.send(cf)
  1046. except OSError:
  1047. print('Error sending CAN frame')
  1048. try:
  1049. s.send(build_can_frame(0x01, b'\x01\x02\x03'))
  1050. except OSError:
  1051. print('Error sending CAN frame')
  1052. Running an example several times with too small delay between executions, could
  1053. lead to this error::
  1054. OSError: [Errno 98] Address already in use
  1055. This is because the previous execution has left the socket in a ``TIME_WAIT``
  1056. state, and can't be immediately reused.
  1057. There is a :mod:`socket` flag to set, in order to prevent this,
  1058. :data:`socket.SO_REUSEADDR`::
  1059. s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
  1060. s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
  1061. s.bind((HOST, PORT))
  1062. the :data:`SO_REUSEADDR` flag tells the kernel to reuse a local socket in
  1063. ``TIME_WAIT`` state, without waiting for its natural timeout to expire.
  1064. .. seealso::
  1065. For an introduction to socket programming (in C), see the following papers:
  1066. - *An Introductory 4.3BSD Interprocess Communication Tutorial*, by Stuart Sechrest
  1067. - *An Advanced 4.3BSD Interprocess Communication Tutorial*, by Samuel J. Leffler et
  1068. al,
  1069. both in the UNIX Programmer's Manual, Supplementary Documents 1 (sections
  1070. PS1:7 and PS1:8). The platform-specific reference material for the various
  1071. socket-related system calls are also a valuable source of information on the
  1072. details of socket semantics. For Unix, refer to the manual pages; for Windows,
  1073. see the WinSock (or Winsock 2) specification. For IPv6-ready APIs, readers may
  1074. want to refer to :rfc:`3493` titled Basic Socket Interface Extensions for IPv6.