Stowage/cpython
2
0
Fork 0
mirror of https://github.com/python/cpython.git synced 2025-11-04 07:31:38 +00:00
Code Issues Projects Releases Packages Wiki Activity
cpython/Modules/socketmodule.c

5355 lines
133 KiB
C
Raw Normal View History

Initial revision
1991-06-25 21:36:08 +00:00
/* Socket module */
/*
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
Initial revision
1991-06-25 21:36:08 +00:00
This module provides an interface to Berkeley socket IPC.
Limitations:
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
- Only AF_INET, AF_INET6 and AF_UNIX address families are supported in a
Issue #1646: Make socket support TIPC. The socket module now has support for TIPC under Linux, see http://tipc.sf.net/ for more information. Thanks to Alberto Bertogli for the patch
2008-01-07 16:12:44 +00:00
portable manner, though AF_PACKET, AF_NETLINK and AF_TIPC are supported
under Linux.
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
- No read/write operations (use sendall/recv or makefile instead).
- Additional restrictions apply on some non-Unix platforms (compensated
for by socket.py).
Initial revision
1991-06-25 21:36:08 +00:00
fixed up comments describing interface
1995-03-16 15:43:47 +00:00
Module interface:
Initial revision
1991-06-25 21:36:08 +00:00
fixed up comments describing interface
1995-03-16 15:43:47 +00:00
- socket.error: exception raised for socket specific errors
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
- socket.gaierror: exception raised for getaddrinfo/getnameinfo errors,
a subclass of socket.error
- socket.herror: exception raised for gethostby* errors,
a subclass of socket.error
Update the module doc comment.
2002-09-03 19:10:18 +00:00
- socket.fromfd(fd, family, type[, proto]) --> new socket object (created
from an existing file descriptor)
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
- socket.gethostbyname(hostname) --> host IP address (string: 'dd.dd.dd.dd')
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
- socket.gethostbyaddr(IP address) --> (hostname, [alias, ...], [IP addr, ...])
fixed up comments describing interface
1995-03-16 15:43:47 +00:00
- socket.gethostname() --> host name (string: 'spam' or 'spam.domain.com')
Added the new getprotobyname() call to the module overview at the top.
1996-12-19 16:42:52 +00:00
- socket.getprotobyname(protocolname) --> protocol number
Added socket.getservbyport(), and make its second argument and that of getservbyname() optional. Update the tests and the docs.
2004-06-28 00:50:43 +00:00
- socket.getservbyname(servicename[, protocolname]) --> port number
- socket.getservbyport(portnumber[, protocolname]) --> service name
Update the module doc comment.
2002-09-03 19:10:18 +00:00
- socket.socket([family[, type [, proto]]]) --> new socket object
Patch #1003700: Add socketpair function to socket module.
2004-08-09 04:51:41 +00:00
- socket.socketpair([family[, type [, proto]]]) --> (socket, socket)
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
- socket.ntohs(16 bit value) --> new int object
- socket.ntohl(32 bit value) --> new int object
- socket.htons(16 bit value) --> new int object
- socket.htonl(32 bit value) --> new int object
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
- socket.getaddrinfo(host, port [, family, socktype, proto, flags])
--> List of (family, socktype, proto, canonname, sockaddr)
- socket.getnameinfo(sockaddr, flags) --> (host, port)
fixed up comments describing interface
1995-03-16 15:43:47 +00:00
- socket.AF_INET, socket.SOCK_STREAM, etc.: constants from <socket.h>
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
- socket.has_ipv6: boolean value indicating if IPv6 is supported
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
- socket.inet_aton(IP address) -> 32-bit packed IP representation
- socket.inet_ntoa(packed IP) -> IP address string
Add default timeout functionality. This adds setdefaulttimeout() and getdefaulttimeout() functions to the socket and _socket modules, and appropriate tests.
2002-07-18 17:08:35 +00:00
- socket.getdefaulttimeout() -> None | float
- socket.setdefaulttimeout(None | float)
Initial revision
1991-06-25 21:36:08 +00:00
- an Internet socket address is a pair (hostname, port)
where hostname can be anything recognized by gethostbyname()
(including the dd.dd.dd.dd notation) and port is in host byte order
- where a hostname is returned, the dd.dd.dd.dd notation is used
fixed up comments describing interface
1995-03-16 15:43:47 +00:00
- a UNIX domain socket address is a string specifying the pathname
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
- an AF_PACKET socket address is a tuple containing a string
specifying the ethernet interface and an integer specifying
the Ethernet protocol number to be received. For example:
fix a couple last-minute bugs in the raw socket support
2001-02-02 19:55:17 +00:00
("eth0",0x1234). Optional 3rd,4th,5th elements in the tuple
[Patch #947352 from Jason Andryuk] Add support for AF_PACKET hardware addresses
2004-07-10 23:39:35 +00:00
specify packet-type and ha-type/addr.
Issue #1646: Make socket support TIPC. The socket module now has support for TIPC under Linux, see http://tipc.sf.net/ for more information. Thanks to Alberto Bertogli for the patch
2008-01-07 16:12:44 +00:00
- an AF_TIPC socket address is expressed as
(addr_type, v1, v2, v3 [, scope]); where addr_type can be one of:
TIPC_ADDR_NAMESEQ, TIPC_ADDR_NAME, and TIPC_ADDR_ID;
and scope can be one of:
TIPC_ZONE_SCOPE, TIPC_CLUSTER_SCOPE, and TIPC_NODE_SCOPE.
The meaning of v1, v2 and v3 depends on the value of addr_type:
if addr_type is TIPC_ADDR_NAME:
v1 is the server type
v2 is the port identifier
v3 is ignored
if addr_type is TIPC_ADDR_NAMESEQ:
v1 is the server type
v2 is the lower port number
v3 is the upper port number
if addr_type is TIPC_ADDR_ID:
v1 is the node
v2 is the ref
v3 is ignored
Initial revision
1991-06-25 21:36:08 +00:00
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
Local naming conventions:
- names starting with sock_ are socket object methods
- names starting with socket_ are module-level functions
- names starting with PySocket are exported through socketmodule.h
Initial revision
1991-06-25 21:36:08 +00:00
*/
Patch 1471925 - Weak linking support for OSX This patch causes several symbols in the socket and posix module to be weakly linked on OSX and disables usage of ftime on OSX. These changes make it possible to use a binary build on OSX 10.4 on a 10.3 system.
2006-04-23 11:59:25 +00:00
#ifdef __APPLE__
/*
* inet_aton is not available on OSX 10.3, yet we want to use a binary
* that was build on 10.4 or later to work on that release, weak linking
* comes to the rescue.
*/
# pragma weak inet_aton
#endif
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
#include "Python.h"
RFE #567972: Socket objects' family, type and proto properties are now exposed via new get...() methods.
2006-03-17 19:17:34 +00:00
#include "structmember.h"
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
#undef MAX
#define MAX(x, y) ((x) < (y) ? (y) : (x))
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
/* Socket object documentation */
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(sock_doc,
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
"socket([family[, type[, proto]]]) -> socket object\n\
\n\
Open a socket of the given type. The family argument specifies the\n\
address family; it defaults to AF_INET. The type argument specifies\n\
whether this is a stream (SOCK_STREAM, this is the default)\n\
or datagram (SOCK_DGRAM) socket. The protocol argument defaults to 0,\n\
specifying the default protocol. Keyword arguments are accepted.\n\
\n\
A socket object represents one endpoint of a network connection.\n\
\n\
Methods of socket objects (keyword arguments not allowed):\n\
\n\
accept() -- accept a connection, returning new socket and client address\n\
bind(addr) -- bind the socket to a local address\n\
close() -- close the socket\n\
connect(addr) -- connect the socket to a remote address\n\
connect_ex(addr) -- connect, return an error code instead of an exception\n\
dup() -- return a new socket object identical to the current one [*]\n\
fileno() -- return underlying file descriptor\n\
getpeername() -- return remote address [*]\n\
getsockname() -- return local address\n\
getsockopt(level, optname[, buflen]) -- get socket options\n\
gettimeout() -- return timeout or None\n\
listen(n) -- start listening for incoming connections\n\
makefile([mode, [bufsize]]) -- return a file object for the socket [*]\n\
recv(buflen[, flags]) -- receive data\n\
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
recv_into(buffer[, nbytes[, flags]]) -- receive data (into a buffer)\n\
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
recvfrom(buflen[, flags]) -- receive data and sender\'s address\n\
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
recvfrom_into(buffer[, nbytes, [, flags])\n\
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
-- receive data and sender\'s address (into a buffer)\n\
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sendall(data[, flags]) -- send all data\n\
send(data[, flags]) -- send data, may not send all of it\n\
sendto(data[, flags], addr) -- send data to a given address\n\
setblocking(0 | 1) -- set or clear the blocking I/O flag\n\
setsockopt(level, optname, value) -- set socket options\n\
settimeout(None | float) -- set or clear the timeout\n\
shutdown(how) -- shut down traffic in one or both directions\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
[*] not available on all platforms!");
Clean up pass for the previous patches. - Use HAVE_GETHOSTBYNAME_R_6_ARG instead of testing for Linux and glibc2. - If gethostbyname takes 3 args, undefine HAVE_GETHOSTBYNAME_R -- don't know what code should be used. - New symbol USE_GETHOSTBYNAME_LOCK defined iff the lock should be used. - Modify the gethostbyaddr() code to also hold on to the lock until after it is safe to release, overlapping with the Python lock. (Note: I think that it could in theory be possible that Python code executed while gethostbyname_lock is held could attempt to reacquire the lock -- e.g. in a signal handler or destructor. I will simply say "don't do that then.")
1999-03-22 21:44:51 +00:00
Fix a bunch of typos in documentation, docstrings and comments. (From SF patch #810751)
2003-10-20 14:01:56 +00:00
/* XXX This is a terrible mess of platform-dependent preprocessor hacks.
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
I hope some day someone can clean this up please... */
(1) On Linux, we really need to trust the configure script to select the right variant of gethostbyname_r for us, since not all Linuxes are equal in this respect. Reported by Laurent Pointal. (2) On BeOS, Chris Herborth reports that instead of arpa/inet.h you must include net/netdb.h to get the inet_ntoa() and inet_addr() prototypes.
1999-09-15 22:01:40 +00:00
/* Hacks for gethostbyname_r(). On some non-Linux platforms, the configure
script doesn't get this right, so we hardcode some platform checks below.
On the other hand, not all Linux versions agree, so there the settings
computed by the configure script are needed! */
#ifndef linux
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# undef HAVE_GETHOSTBYNAME_R_3_ARG
# undef HAVE_GETHOSTBYNAME_R_5_ARG
# undef HAVE_GETHOSTBYNAME_R_6_ARG
(1) On Linux, we really need to trust the configure script to select the right variant of gethostbyname_r for us, since not all Linuxes are equal in this respect. Reported by Laurent Pointal. (2) On BeOS, Chris Herborth reports that instead of arpa/inet.h you must include net/netdb.h to get the inet_ntoa() and inet_addr() prototypes.
1999-09-15 22:01:40 +00:00
#endif
Implement two suggestions by Jonathan Giddy: (1) in AIX, clear the data struct before calling gethostby{name,addr}_r(); (2) ignore the 3/5/6 args determinations made by the configure script and switch on platform identifiers instead: AIX, OSF have 3 args Sun, SGI have 5 args Linux has 6 args On all other platforms, undef HAVE_GETHOSTBYNAME_R altogether.
1999-03-24 17:24:33 +00:00
Put back __osf__ support for gethostbyname_r(); the real bug was that it was being used even without threads. This of course might be an all-platform problem so now we only use the _r variant when we are using threads.
1999-04-13 04:07:32 +00:00
#ifndef WITH_THREAD
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# undef HAVE_GETHOSTBYNAME_R
Put back __osf__ support for gethostbyname_r(); the real bug was that it was being used even without threads. This of course might be an all-platform problem so now we only use the _r variant when we are using threads.
1999-04-13 04:07:32 +00:00
#endif
Implement two suggestions by Jonathan Giddy: (1) in AIX, clear the data struct before calling gethostby{name,addr}_r(); (2) ignore the 3/5/6 args determinations made by the configure script and switch on platform identifiers instead: AIX, OSF have 3 args Sun, SGI have 5 args Linux has 6 args On all other platforms, undef HAVE_GETHOSTBYNAME_R altogether.
1999-03-24 17:24:33 +00:00
#ifdef HAVE_GETHOSTBYNAME_R
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# if defined(_AIX) || defined(__osf__)
# define HAVE_GETHOSTBYNAME_R_3_ARG
# elif defined(__sun) || defined(__sgi)
# define HAVE_GETHOSTBYNAME_R_5_ARG
# elif defined(linux)
(1) On Linux, we really need to trust the configure script to select the right variant of gethostbyname_r for us, since not all Linuxes are equal in this respect. Reported by Laurent Pointal. (2) On BeOS, Chris Herborth reports that instead of arpa/inet.h you must include net/netdb.h to get the inet_ntoa() and inet_addr() prototypes.
1999-09-15 22:01:40 +00:00
/* Rely on the configure script */
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# else
# undef HAVE_GETHOSTBYNAME_R
# endif
Implement two suggestions by Jonathan Giddy: (1) in AIX, clear the data struct before calling gethostby{name,addr}_r(); (2) ignore the 3/5/6 args determinations made by the configure script and switch on platform identifiers instead: AIX, OSF have 3 args Sun, SGI have 5 args Linux has 6 args On all other platforms, undef HAVE_GETHOSTBYNAME_R altogether.
1999-03-24 17:24:33 +00:00
#endif
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
#if !defined(HAVE_GETHOSTBYNAME_R) && defined(WITH_THREAD) && \
!defined(MS_WINDOWS)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# define USE_GETHOSTBYNAME_LOCK
Clean up pass for the previous patches. - Use HAVE_GETHOSTBYNAME_R_6_ARG instead of testing for Linux and glibc2. - If gethostbyname takes 3 args, undefine HAVE_GETHOSTBYNAME_R -- don't know what code should be used. - New symbol USE_GETHOSTBYNAME_LOCK defined iff the lock should be used. - Modify the gethostbyaddr() code to also hold on to the lock until after it is safe to release, overlapping with the Python lock. (Note: I think that it could in theory be possible that Python code executed while gethostbyname_lock is held could attempt to reacquire the lock -- e.g. in a signal handler or destructor. I will simply say "don't do that then.")
1999-03-22 21:44:51 +00:00
#endif
Patch #1288833: Removed thread lock from socket.getaddrinfo on FreeBSD 5.3 and later versions which got thread-safe getaddrinfo(3). (Reported by Maxim Sobolev)
2005-09-24 14:58:47 +00:00
/* To use __FreeBSD_version */
#ifdef HAVE_SYS_PARAM_H
#include <sys/param.h>
#endif
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
/* On systems on which getaddrinfo() is believed to not be thread-safe,
Addition to [731644]: also use the lock when using the getaddrinfo emulation (ie. when HAVE_GETADDRINFO isn't defined).
2003-05-09 08:03:44 +00:00
(this includes the getaddrinfo emulation) protect access with a lock. */
Patch #1288833: Removed thread lock from socket.getaddrinfo on FreeBSD 5.3 and later versions which got thread-safe getaddrinfo(3). (Reported by Maxim Sobolev)
2005-09-24 14:58:47 +00:00
#if defined(WITH_THREAD) && (defined(__APPLE__) || \
(defined(__FreeBSD__) && __FreeBSD_version+0 < 503000) || \
Patch #1516912: improve Modules support for OpenVMS.
2006-07-10 01:18:57 +00:00
defined(__OpenBSD__) || defined(__NetBSD__) || \
defined(__VMS) || !defined(HAVE_GETADDRINFO))
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
#define USE_GETADDRINFO_LOCK
#endif
#ifdef USE_GETADDRINFO_LOCK
#define ACQUIRE_GETADDRINFO_LOCK PyThread_acquire_lock(netdb_lock, 1);
#define RELEASE_GETADDRINFO_LOCK PyThread_release_lock(netdb_lock);
#else
#define ACQUIRE_GETADDRINFO_LOCK
#define RELEASE_GETADDRINFO_LOCK
#endif
#if defined(USE_GETHOSTBYNAME_LOCK) || defined(USE_GETADDRINFO_LOCK)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# include "pythread.h"
Threading speedup patches by Christopher Lindblad <cjl@infoseek.com>.
1998-04-09 20:56:35 +00:00
#endif
Initial revision
1991-06-25 21:36:08 +00:00
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
#if defined(PYCC_VACPP)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# include <types.h>
# include <io.h>
# include <sys/ioctl.h>
# include <utils.h>
# include <ctype.h>
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
#endif
Don't include <ioctl.h> unless on __VMS.
2002-12-06 12:57:26 +00:00
#if defined(__VMS)
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
# include <ioctl.h>
#endif
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
#if defined(PYOS_OS2)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# define INCL_DOS
# define INCL_DOSERRORS
# define INCL_NOPMAPI
# include <os2.h>
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
#endif
Get socketmodule compiling on IRIX 6.5.10. I don't think the fix here is very good, but I'm not sure what would be better. In particular, we should not be defining _SGIAPI, but lots of things break if we remove it.
2003-07-17 16:58:48 +00:00
#if defined(__sgi) && _COMPILER_VERSION>700 && !_SGIAPI
Patch #608999: Fix portability problems with MIPSPro 7.x
2002-09-19 08:03:21 +00:00
/* make sure that the reentrant (gethostbyaddr_r etc)
functions are declared correctly if compiling with
MIPSPro 7.x in ANSI C mode (default) */
Get socketmodule compiling on IRIX 6.5.10. I don't think the fix here is very good, but I'm not sure what would be better. In particular, we should not be defining _SGIAPI, but lots of things break if we remove it.
2003-07-17 16:58:48 +00:00
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
/* XXX Using _SGIAPI is the wrong thing,
Get socketmodule compiling on IRIX 6.5.10. I don't think the fix here is very good, but I'm not sure what would be better. In particular, we should not be defining _SGIAPI, but lots of things break if we remove it.
2003-07-17 16:58:48 +00:00
but I don't know what the right thing is. */
Patch for compilation on IRIX from rwgk on http://python.org/sf/728330
2004-09-13 17:48:41 +00:00
#undef _SGIAPI /* to avoid warning */
Patch #608999: Fix portability problems with MIPSPro 7.x
2002-09-19 08:03:21 +00:00
#define _SGIAPI 1
Get socketmodule compiling on IRIX 6.5.10. I don't think the fix here is very good, but I'm not sure what would be better. In particular, we should not be defining _SGIAPI, but lots of things break if we remove it.
2003-07-17 16:58:48 +00:00
Patch for compilation on IRIX from rwgk on http://python.org/sf/728330
2004-09-13 17:48:41 +00:00
#undef _XOPEN_SOURCE
#include <sys/socket.h>
#include <sys/types.h>
#include <netinet/in.h>
#ifdef _SS_ALIGNSIZE
#define HAVE_GETADDRINFO 1
#define HAVE_GETNAMEINFO 1
#endif
Get socketmodule compiling on IRIX 6.5.10. I don't think the fix here is very good, but I'm not sure what would be better. In particular, we should not be defining _SGIAPI, but lots of things break if we remove it.
2003-07-17 16:58:48 +00:00
#define HAVE_INET_PTON
#include <netdb.h>
Patch #608999: Fix portability problems with MIPSPro 7.x
2002-09-19 08:03:21 +00:00
#endif
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
/* Irix 6.5 fails to define this variable at all. This is needed
for both GCC and SGI's compiler. I'd say that the SGI headers
Bug #854823: socketmodule now builds on Sun platforms even when INET_ADDRSTRLEN is not defined.
2006-02-20 09:42:33 +00:00
are just busted. Same thing for Solaris. */
#if (defined(__sgi) || defined(sun)) && !defined(INET_ADDRSTRLEN)
made the SGI INET_ADDRSTRLEN define apply for any compiler on Irix. Both the SGI compiler and GCC seem to need this.
2003-10-04 08:00:49 +00:00
#define INET_ADDRSTRLEN 16
#endif
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
/* Generic includes */
Patch #1495999: Part two of Windows CE changes. - update header checks, using autoconf - provide dummies for getenv, environ, and GetVersion - adjust MSC_VER check in socketmodule.c
2006-06-10 12:23:46 +00:00
#ifdef HAVE_SYS_TYPES_H
Merge alpha100 branch back to main trunk
1994-08-01 11:34:53 +00:00
#include <sys/types.h>
Patch #1495999: Part two of Windows CE changes. - update header checks, using autoconf - provide dummies for getenv, environ, and GetVersion - adjust MSC_VER check in socketmodule.c
2006-06-10 12:23:46 +00:00
#endif
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
Also fix the comment.
2002-02-16 18:47:07 +00:00
/* Generic socket object definitions and includes */
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
#define PySocket_BUILDING_SOCKET
Fix the name of the header file.
2002-02-16 18:44:52 +00:00
#include "socketmodule.h"
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
/* Addressing includes */
Slightly different Windows ifdefs
1996-06-28 20:15:15 +00:00
#ifndef MS_WINDOWS
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
/* Non-MS WINDOWS includes */
# include <netdb.h>
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
(1) On Linux, we really need to trust the configure script to select the right variant of gethostbyname_r for us, since not all Linuxes are equal in this respect. Reported by Laurent Pointal. (2) On BeOS, Chris Herborth reports that instead of arpa/inet.h you must include net/netdb.h to get the inet_ntoa() and inet_addr() prototypes.
1999-09-15 22:01:40 +00:00
/* Headers needed for inet_ntoa() and inet_addr() */
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# ifdef __BEOS__
# include <net/netdb.h>
# elif defined(PYOS_OS2) && defined(PYCC_VACPP)
# include <netdb.h>
SF patch 473749 compile under OS/2 VA C++, from Michael Muller. Changes enabling Python to compile under OS/2 Visual Age C++.
2001-11-05 02:45:59 +00:00
typedef size_t socklen_t;
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# else
# include <arpa/inet.h>
# endif
# ifndef RISCOS
# include <fcntl.h>
# else
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
# include <sys/ioctl.h>
# include <socklib.h>
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# define NO_DUP
SF patch #474590 -- RISC OS support
2001-10-24 20:42:55 +00:00
int h_errno; /* not used */
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
# define INET_ADDRSTRLEN 16
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# endif
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
Merge alpha100 branch back to main trunk
1994-08-01 11:34:53 +00:00
#else
* Microscopic corrections to make things compile on the Cray APP. * Removed one use of $> in Makefile and warned about others. Added configurable lines in Makefile to change CC and AR.
1993-04-07 14:06:14 +00:00
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
/* MS_WINDOWS includes */
Patch #1495999: Part two of Windows CE changes. - update header checks, using autoconf - provide dummies for getenv, environ, and GetVersion - adjust MSC_VER check in socketmodule.c
2006-06-10 12:23:46 +00:00
# ifdef HAVE_FCNTL_H
# include <fcntl.h>
# endif
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
#endif
remove support for missing ANSI C header files (limits.h, stddef.h, etc).
2004-02-10 16:50:21 +00:00
#include <stddef.h>
Provide a definition for offsetof.
2001-06-24 21:28:42 +00:00
#ifndef offsetof
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
# define offsetof(type, member) ((size_t)(&((type *)0)->member))
Provide a definition for offsetof.
2001-06-24 21:28:42 +00:00
#endif
Use O_NONBLOCK rather than O_NDELAY, so we get POSIX non-blocking I/O. On HPUX, Solaris, Tru64 (Dec UNIX), and IRIX (I think), O_NONBLOCK is the POSIX version of non-blocking I/O which is what we want. On Linux and FreeBSD (at least), O_NONBLOCK and O_NDELAY are the same. So this change should have no negative effect on those platforms. Tested on Linux, Solaris, HPUX. Thanks to Anders Qvist for diagnosing this problem.
2002-11-02 19:55:21 +00:00
#ifndef O_NONBLOCK
# define O_NONBLOCK O_NDELAY
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
#endif
Apply patch from http://python.org/sf/728330 to fix socket module compilation on Solaris 2.6, HP-UX 11, AIX 5.1 and (possibly) some IRIX versions.
2004-09-07 17:48:26 +00:00
/* include Python's addrinfo.h unless it causes trouble */
#if defined(__sgi) && _COMPILER_VERSION>700 && defined(_SS_ALIGNSIZE)
/* Do not include addinfo.h on some newer IRIX versions.
* _SS_ALIGNSIZE is defined in sys/socket.h by 6.5.21,
* for example, but not by 6.5.10.
*/
Patch #1495999: Part two of Windows CE changes. - update header checks, using autoconf - provide dummies for getenv, environ, and GetVersion - adjust MSC_VER check in socketmodule.c
2006-06-10 12:23:46 +00:00
#elif defined(_MSC_VER) && _MSC_VER>1201
Apply patch from http://python.org/sf/728330 to fix socket module compilation on Solaris 2.6, HP-UX 11, AIX 5.1 and (possibly) some IRIX versions.
2004-09-07 17:48:26 +00:00
/* Do not include addrinfo.h for MSVC7 or greater. 'addrinfo' and
* EAI_* constants are defined in (the already included) ws2tcpip.h.
*/
#else
# include "addrinfo.h"
#endif
Restore compilation on MSVC++ 6.0
2004-08-03 08:52:46 +00:00
Emulate inet_{pton,ntop} on systems that don't provide it.
2001-06-24 21:18:26 +00:00
#ifndef HAVE_INET_PTON
Backport of the PCbuild9 directory from the py3k branch. I've finished the last task for the PCbuild9 directory today. I don't think there is much left to do. Now you can all play around with the shiny new VS 2008 and try the PGO builds. I was able to get a speed improvement of about 10% on py3k. Have fun! :)
2007-11-22 11:21:16 +00:00
#if !defined(NTDDI_VERSION) || (NTDDI_VERSION < NTDDI_LONGHORN)
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
int inet_pton(int af, const char *src, void *dst);
Silence warnings in MSVC++: hide unused variables, add constness back to inet_pton/ntop, convert htons argument to u_short.
2001-07-21 09:42:15 +00:00
const char *inet_ntop(int af, const void *src, char *dst, socklen_t size);
Emulate inet_{pton,ntop} on systems that don't provide it.
2001-06-24 21:18:26 +00:00
#endif
Backport of the PCbuild9 directory from the py3k branch. I've finished the last task for the PCbuild9 directory today. I don't think there is much left to do. Now you can all play around with the shiny new VS 2008 and try the PGO builds. I was able to get a speed improvement of about 10% on py3k. Have fun! :)
2007-11-22 11:21:16 +00:00
#endif
Emulate inet_{pton,ntop} on systems that don't provide it.
2001-06-24 21:18:26 +00:00
Do not use the system getaddrinfo on Mac OS X. Fixes bug #445928. Since getnameinfo is not implemented, remove __APPLE__ check here.
2001-08-03 10:02:29 +00:00
#ifdef __APPLE__
/* On OS X, getaddrinfo returns no error indication of lookup
failure, so we must use the emulation instead of the libinfo
implementation. Unfortunately, performing an autoconf test
for this bug would require DNS access for the machine performing
the configuration, which is not acceptable. Therefore, we
determine the bug just by checking for __APPLE__. If this bug
gets ever fixed, perhaps checking for sys/version.h would be
appropriate, which is 10/0 on the system with the bug. */
Mac OS X Jaguar (developer preview) seems to have a working getaddrinfo().
2002-07-02 14:40:42 +00:00
#ifndef HAVE_GETNAMEINFO
/* This bug seems to be fixed in Jaguar. Ths easiest way I could
Find to check for Jaguar is that it has getnameinfo(), which
older releases don't have */
Do not use the system getaddrinfo on Mac OS X. Fixes bug #445928. Since getnameinfo is not implemented, remove __APPLE__ check here.
2001-08-03 10:02:29 +00:00
#undef HAVE_GETADDRINFO
#endif
Patch 1471925 - Weak linking support for OSX This patch causes several symbols in the socket and posix module to be weakly linked on OSX and disables usage of ftime on OSX. These changes make it possible to use a binary build on OSX 10.4 on a 10.3 system.
2006-04-23 11:59:25 +00:00
#ifdef HAVE_INET_ATON
#define USE_INET_ATON_WEAKLINK
#endif
Mac OS X Jaguar (developer preview) seems to have a working getaddrinfo().
2002-07-02 14:40:42 +00:00
#endif
Do not use the system getaddrinfo on Mac OS X. Fixes bug #445928. Since getnameinfo is not implemented, remove __APPLE__ check here.
2001-08-03 10:02:29 +00:00
Patch #401196: Configuration machinery for IPv6. Contributed by Jun-ichiro "itojun" Hagino. get{addr,name}info emulation code taken from WIDE.
2001-06-23 16:30:13 +00:00
/* I know this is a bad practice, but it is the easiest... */
Do not use the system getaddrinfo on Mac OS X. Fixes bug #445928. Since getnameinfo is not implemented, remove __APPLE__ check here.
2001-08-03 10:02:29 +00:00
#if !defined(HAVE_GETADDRINFO)
Patch #725942: Always rename emulation functions.
2003-05-01 05:20:46 +00:00
/* avoid clashes with the C library definition of the symbol. */
#define getaddrinfo fake_getaddrinfo
#define gai_strerror fake_gai_strerror
#define freeaddrinfo fake_freeaddrinfo
Patch #401196: Configuration machinery for IPv6. Contributed by Jun-ichiro "itojun" Hagino. get{addr,name}info emulation code taken from WIDE.
2001-06-23 16:30:13 +00:00
#include "getaddrinfo.c"
#endif
Do not use the system getaddrinfo on Mac OS X. Fixes bug #445928. Since getnameinfo is not implemented, remove __APPLE__ check here.
2001-08-03 10:02:29 +00:00
#if !defined(HAVE_GETNAMEINFO)
Patch #725942: Always rename emulation functions.
2003-05-01 05:20:46 +00:00
#define getnameinfo fake_getnameinfo
Patch #401196: Configuration machinery for IPv6. Contributed by Jun-ichiro "itojun" Hagino. get{addr,name}info emulation code taken from WIDE.
2001-06-23 16:30:13 +00:00
#include "getnameinfo.c"
#endif
Changes for BeOS, QNX and long long, by Chris Herborth.
1998-08-04 22:53:56 +00:00
#if defined(MS_WINDOWS) || defined(__BEOS__)
/* BeOS suffers from the same socket dichotomy as Win32... - [cjh] */
Changes to make it work and compile on NT, by Mark Hammond. (NT changes for posixmodule.c re-invented by Guido.)
1994-08-18 15:42:46 +00:00
/* seem to be a few differences in the API */
Use an explicit macro SOCKETCLOSE which expands to closesocket (on Windows), soclose (on OS2), or to close (everywhere else). Hopefully this fixes a new compilation error that I suddenly get on Windows because the macro definition for close -> closesocket apparently was done before including io.h, which contains a prototype for close. (No idea why this wasn't an error before.)
2000-04-21 20:33:00 +00:00
#define SOCKETCLOSE closesocket
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
#define NO_DUP /* Actually it exists on NT 3.5, but what the heck... */
Changes to make it work and compile on NT, by Mark Hammond. (NT changes for posixmodule.c re-invented by Guido.)
1994-08-18 15:42:46 +00:00
#endif
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#ifdef MS_WIN32
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
#define EAFNOSUPPORT WSAEAFNOSUPPORT
#define snprintf _snprintf
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
OS/2 EMX port changes (Modules part of patch #450267): Modules/ socketmodule.c EMX handles sockets like Posix, rather than use native APIs
2002-03-03 03:03:52 +00:00
#if defined(PYOS_OS2) && !defined(PYCC_GCC)
Use an explicit macro SOCKETCLOSE which expands to closesocket (on Windows), soclose (on OS2), or to close (everywhere else). Hopefully this fixes a new compilation error that I suddenly get on Windows because the macro definition for close -> closesocket apparently was done before including io.h, which contains a prototype for close. (No idea why this wasn't an error before.)
2000-04-21 20:33:00 +00:00
#define SOCKETCLOSE soclose
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
#define NO_DUP /* Sockets are Not Actual File Handles under OS/2 */
#endif
Use an explicit macro SOCKETCLOSE which expands to closesocket (on Windows), soclose (on OS2), or to close (everywhere else). Hopefully this fixes a new compilation error that I suddenly get on Windows because the macro definition for close -> closesocket apparently was done before including io.h, which contains a prototype for close. (No idea why this wasn't an error before.)
2000-04-21 20:33:00 +00:00
#ifndef SOCKETCLOSE
#define SOCKETCLOSE close
#endif
Apply patch for netbsd multiprocessing support
2009-03-31 18:12:35 +00:00
#if defined(HAVE_BLUETOOTH_H) || defined(HAVE_BLUETOOTH_BLUETOOTH_H) && !defined(__NetBSD__)
Add FreeBSD support for bluetooth sockets. (SF Patch #888148, reviewed by loewis)
2004-02-02 06:05:24 +00:00
#define USE_BLUETOOTH 1
#if defined(__FreeBSD__)
#define BTPROTO_L2CAP BLUETOOTH_PROTO_L2CAP
#define BTPROTO_RFCOMM BLUETOOTH_PROTO_RFCOMM
Patch #1432399: Add HCI sockets.
2007-02-14 10:07:37 +00:00
#define BTPROTO_HCI BLUETOOTH_PROTO_HCI
Fix build on FreeBSD. Bluetooth HCI API in FreeBSD is quite different from Linux's. Just fix the build for now but the code doesn't support the complete capability of HCI on FreeBSD yet.
2007-06-05 18:16:52 +00:00
#define SOL_HCI SOL_HCI_RAW
#define HCI_FILTER SO_HCI_RAW_FILTER
Add FreeBSD support for bluetooth sockets. (SF Patch #888148, reviewed by loewis)
2004-02-02 06:05:24 +00:00
#define sockaddr_l2 sockaddr_l2cap
#define sockaddr_rc sockaddr_rfcomm
Fix build on FreeBSD. Bluetooth HCI API in FreeBSD is quite different from Linux's. Just fix the build for now but the code doesn't support the complete capability of HCI on FreeBSD yet.
2007-06-05 18:16:52 +00:00
#define hci_dev hci_node
Add FreeBSD support for bluetooth sockets. (SF Patch #888148, reviewed by loewis)
2004-02-02 06:05:24 +00:00
#define _BT_L2_MEMB(sa, memb) ((sa)->l2cap_##memb)
#define _BT_RC_MEMB(sa, memb) ((sa)->rfcomm_##memb)
Patch #1432399: Add HCI sockets.
2007-02-14 10:07:37 +00:00
#define _BT_HCI_MEMB(sa, memb) ((sa)->hci_##memb)
Allow socketmodule to compile on NetBSD -current, whose bluetooth API differs from both Linux and FreeBSD. Accepted by Neal Norwitz.
2006-07-28 11:27:27 +00:00
#elif defined(__NetBSD__)
#define sockaddr_l2 sockaddr_bt
#define sockaddr_rc sockaddr_bt
Patch #1432399: Add HCI sockets.
2007-02-14 10:07:37 +00:00
#define sockaddr_hci sockaddr_bt
Allow socketmodule to compile on NetBSD -current, whose bluetooth API differs from both Linux and FreeBSD. Accepted by Neal Norwitz.
2006-07-28 11:27:27 +00:00
#define sockaddr_sco sockaddr_bt
#define _BT_L2_MEMB(sa, memb) ((sa)->bt_##memb)
#define _BT_RC_MEMB(sa, memb) ((sa)->bt_##memb)
Patch #1432399: Add HCI sockets.
2007-02-14 10:07:37 +00:00
#define _BT_HCI_MEMB(sa, memb) ((sa)->bt_##memb)
Allow socketmodule to compile on NetBSD -current, whose bluetooth API differs from both Linux and FreeBSD. Accepted by Neal Norwitz.
2006-07-28 11:27:27 +00:00
#define _BT_SCO_MEMB(sa, memb) ((sa)->bt_##memb)
Add FreeBSD support for bluetooth sockets. (SF Patch #888148, reviewed by loewis)
2004-02-02 06:05:24 +00:00
#else
#define _BT_L2_MEMB(sa, memb) ((sa)->l2_##memb)
#define _BT_RC_MEMB(sa, memb) ((sa)->rc_##memb)
Patch #1432399: Add HCI sockets.
2007-02-14 10:07:37 +00:00
#define _BT_HCI_MEMB(sa, memb) ((sa)->hci_##memb)
Add FreeBSD support for bluetooth sockets. (SF Patch #888148, reviewed by loewis)
2004-02-02 06:05:24 +00:00
#define _BT_SCO_MEMB(sa, memb) ((sa)->sco_##memb)
#endif
#endif
Patch #1516912: improve Modules support for OpenVMS.
2006-07-10 01:18:57 +00:00
#ifdef __VMS
/* TCP/IP Services for VMS uses a maximum send/recv buffer length */
#define SEGMENT_SIZE (32 * 1024 -1)
#endif
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
#define SAS2SA(x) ((struct sockaddr *)(x))
Patch #724588: Check whether the address of hstrerror and inet_pton can be taken, and define NI_MAX{HOST|SERV} if necessary.
2003-05-03 10:12:45 +00:00
/*
* Constants for getnameinfo()
*/
#if !defined(NI_MAXHOST)
#define NI_MAXHOST 1025
#endif
#if !defined(NI_MAXSERV)
#define NI_MAXSERV 32
#endif
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
/* XXX There's a problem here: *static* functions are not supposed to have
a Py prefix (or use CapitalizedWords). Later... */
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* Global variable holding the exception type for errors detected
by this module (but not argument type or memory errors, etc.). */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
static PyObject *socket_error;
static PyObject *socket_herror;
static PyObject *socket_gaierror;
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
static PyObject *socket_timeout;
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
#ifdef RISCOS
/* Global variable which is !=0 if Python is running in a RISC OS taskwindow */
static int taskwindow;
#endif
Moved the declaration of PySocketSock_Type from socketmodule.h to socketmodule.c. No code outside of the .c file references it, so it doesn't belong the .h file (at least not yet ...), and declaring it an imported symbol in the .h file can't be made to work on Windows (it's a cross-DLL symbol then) without substantial code rewriting. Also repaired the comment that goes along with the decl, to stop referring to names and functions that haven't existed for 7 years <wink>. socketmodule.c compiles cleanly on Windows again. The test_socket dies at once, though (later).
2002-02-17 04:13:21 +00:00
/* A forward reference to the socket type object.
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
The sock_type variable contains pointers to various functions,
some of which call new_sockobject(), which uses sock_type, so
Moved the declaration of PySocketSock_Type from socketmodule.h to socketmodule.c. No code outside of the .c file references it, so it doesn't belong the .h file (at least not yet ...), and declaring it an imported symbol in the .h file can't be made to work on Windows (it's a cross-DLL symbol then) without substantial code rewriting. Also repaired the comment that goes along with the decl, to stop referring to names and functions that haven't existed for 7 years <wink>. socketmodule.c compiles cleanly on Windows again. The test_socket dies at once, though (later).
2002-02-17 04:13:21 +00:00
there has to be a circular reference. */
staticforward bites the dust. The staticforward define was needed to support certain broken C compilers (notably SCO ODT 3.0, perhaps early AIX as well) botched the static keyword when it was used with a forward declaration of a static initialized structure. Standard C allows the forward declaration with static, and we've decided to stop catering to broken C compilers. (In fact, we expect that the compilers are all fixed eight years later.) I'm leaving staticforward and statichere defined in object.h as static. This is only for backwards compatibility with C extensions that might still use it. XXX I haven't updated the documentation.
2002-07-17 16:30:39 +00:00
static PyTypeObject sock_type;
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
#1494314: Fix a regression with high-numbered sockets in 2.4.3. This means that select() on sockets > FD_SETSIZE (typically 1024) work again. The patch makes sockets use poll() internally where available.
2006-07-11 02:04:09 +00:00
#if defined(HAVE_POLL_H)
#include <poll.h>
#elif defined(HAVE_SYS_POLL_H)
#include <sys/poll.h>
#endif
Introduce Py_SOCKET_FD_CAN_BE_GE_FD_SETSIZE. Proposed by Tim Peters.
2006-02-11 09:27:05 +00:00
#ifdef Py_SOCKET_FD_CAN_BE_GE_FD_SETSIZE
/* Platform can select file descriptors beyond FD_SETSIZE */
#define IS_SELECTABLE(s) 1
#1494314: Fix a regression with high-numbered sockets in 2.4.3. This means that select() on sockets > FD_SETSIZE (typically 1024) work again. The patch makes sockets use poll() internally where available.
2006-07-11 02:04:09 +00:00
#elif defined(HAVE_POLL)
/* Instead of select(), we'll use poll() since poll() works on any fd. */
#define IS_SELECTABLE(s) 1
/* Can we call select() with this socket without a buffer overrun? */
Introduce Py_SOCKET_FD_CAN_BE_GE_FD_SETSIZE. Proposed by Tim Peters.
2006-02-11 09:27:05 +00:00
#else
/* POSIX says selecting file descriptors beyond FD_SETSIZE
#1494314: Fix a regression with high-numbered sockets in 2.4.3. This means that select() on sockets > FD_SETSIZE (typically 1024) work again. The patch makes sockets use poll() internally where available.
2006-07-11 02:04:09 +00:00
has undefined behaviour. If there's no timeout left, we don't have to
call select, so it's a safe, little white lie. */
#define IS_SELECTABLE(s) ((s)->sock_fd < FD_SETSIZE || s->sock_timeout <= 0.0)
Introduce Py_SOCKET_FD_CAN_BE_GE_FD_SETSIZE. Proposed by Tim Peters.
2006-02-11 09:27:05 +00:00
#endif
Bug #876637, prevent stack corruption when socket descriptor is larger than FD_SETSIZE. This can only be acheived with ulimit -n SOME_NUMBER_BIGGER_THAN_FD_SETSIZE which is typically only available to root. Since this wouldn't normally be run in a test (ie, run as root), it doesn't seem too worthwhile to add a normal test. The bug report has one version of a test. I've written another. Not sure what the best thing to do is. Do the check before calling internal_select() because we can't set an error in between Py_BEGIN_ALLOW_THREADS and Py_END_ALLOW_THREADS. This seemed the clearest solution, ie handle before calling internal_select() rather than inside. Plus there is at least one place outside of internal_select() that needed to be handled. Will backport.
2006-02-07 07:04:46 +00:00
static PyObject*
select_error(void)
{
PyErr_SetString(socket_error, "unable to select on socket");
return NULL;
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* Convenience function to raise an error according to errno
and return a NULL pointer from a function. */
Initial revision
1991-06-25 21:36:08 +00:00
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
set_error(void)
Initial revision
1991-06-25 21:36:08 +00:00
{
changed windows test symbols; removed debug printf from sockets
1996-06-26 18:22:49 +00:00
#ifdef MS_WINDOWS
Patch #100926 - Better error messages for socket exceptions on Windows. [Slight style differences from posted patch]
2000-07-24 01:45:11 +00:00
int err_no = WSAGetLastError();
A patch from issue 1378 by roudkerk: Currently on Windows set_error() make use of a large array which maps socket error numbers to error messages. This patch removes that array and just lets PyErr_SetExcFromWindowsErr() generate the message by using the Win32 function FormatMessage().
2007-11-16 00:24:44 +00:00
/* PyErr_SetExcFromWindowsErr() invokes FormatMessage() which
recognizes the error codes used by both GetLastError() and
WSAGetLastError */
if (err_no)
return PyErr_SetExcFromWindowsErr(socket_error, err_no);
* Modules/xxmodule.c: integrated with xxobject.c by Jack * Modules/(posix,socket}module.c: more NT changes
1994-09-29 09:50:09 +00:00
#endif
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
OS/2 EMX port changes (Modules part of patch #450267): Modules/ socketmodule.c EMX handles sockets like Posix, rather than use native APIs
2002-03-03 03:03:52 +00:00
#if defined(PYOS_OS2) && !defined(PYCC_GCC)
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
if (sock_errno() != NO_ERROR) {
APIRET rc;
ULONG msglen;
char outbuf[100];
int myerrorcode = sock_errno();
/* Retrieve socket-related error message from MPTN.MSG file */
rc = DosGetMessage(NULL, 0, outbuf, sizeof(outbuf),
myerrorcode - SOCBASEERR + 26,
"mptn.msg",
&msglen);
if (rc == NO_ERROR) {
PyObject *v;
/* OS/2 doesn't guarantee a terminator */
outbuf[msglen] = '\0';
if (strlen(outbuf) > 0) {
/* If non-empty msg, trim CRLF */
char *lastc = &outbuf[ strlen(outbuf)-1 ];
Fix SF bug #1072182, problems with signed characters. Most of these can be backported.
2005-12-19 06:05:18 +00:00
while (lastc > outbuf &&
isspace(Py_CHARMASK(*lastc))) {
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
/* Trim trailing whitespace (CRLF) */
*lastc-- = '\0';
}
}
v = Py_BuildValue("(is)", myerrorcode, outbuf);
if (v != NULL) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetObject(socket_error, v);
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
Py_DECREF(v);
}
return NULL;
}
}
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
#endif
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
#if defined(RISCOS)
if (_inet_error.errnum != NULL) {
PyObject *v;
v = Py_BuildValue("(is)", errno, _inet_err());
if (v != NULL) {
PyErr_SetObject(socket_error, v);
Py_DECREF(v);
}
return NULL;
}
#endif
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
return PyErr_SetFromErrno(socket_error);
Initial revision
1991-06-25 21:36:08 +00:00
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
set_herror(int h_error)
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
{
PyObject *v;
#ifdef HAVE_HSTRERROR
v = Py_BuildValue("(is)", h_error, (char *)hstrerror(h_error));
#else
v = Py_BuildValue("(is)", h_error, "host not found");
#endif
if (v != NULL) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetObject(socket_herror, v);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
Py_DECREF(v);
}
return NULL;
}
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
set_gaierror(int error)
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
{
PyObject *v;
Patch #520062: Support IPv6 with VC.NET.
2002-03-01 08:31:07 +00:00
#ifdef EAI_SYSTEM
/* EAI_SYSTEM is not available on Windows XP. */
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
if (error == EAI_SYSTEM)
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
return set_error();
Patch #520062: Support IPv6 with VC.NET.
2002-03-01 08:31:07 +00:00
#endif
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
Fix portability problems with glibc 2.0, as reported in #449157.
2001-08-15 17:14:33 +00:00
#ifdef HAVE_GAI_STRERROR
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
v = Py_BuildValue("(is)", error, gai_strerror(error));
Fix portability problems with glibc 2.0, as reported in #449157.
2001-08-15 17:14:33 +00:00
#else
v = Py_BuildValue("(is)", error, "getaddrinfo failed");
#endif
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
if (v != NULL) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetObject(socket_gaierror, v);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
Py_DECREF(v);
}
return NULL;
}
Patch #1516912: improve Modules support for OpenVMS.
2006-07-10 01:18:57 +00:00
#ifdef __VMS
/* Function to send in segments */
static int
sendsegmented(int sock_fd, char *buf, int len, int flags)
{
int n = 0;
int remaining = len;
while (remaining > 0) {
unsigned int segment;
segment = (remaining >= SEGMENT_SIZE ? SEGMENT_SIZE : remaining);
n = send(sock_fd, buf, segment, flags);
if (n < 0) {
return n;
}
remaining -= segment;
buf += segment;
} /* end while */
return len;
}
#endif
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
/* Function to perform the setting of socket blocking mode
internally. block = (1 | 0). */
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
static int
internal_setblocking(PySocketSockObject *s, int block)
{
#ifndef RISCOS
#ifndef MS_WINDOWS
int delay_flag;
#endif
#endif
Py_BEGIN_ALLOW_THREADS
#ifdef __BEOS__
block = !block;
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
setsockopt(s->sock_fd, SOL_SOCKET, SO_NONBLOCK,
(void *)(&block), sizeof(int));
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
#else
#ifndef RISCOS
#ifndef MS_WINDOWS
#if defined(PYOS_OS2) && !defined(PYCC_GCC)
block = !block;
ioctl(s->sock_fd, FIONBIO, (caddr_t)&block, sizeof(block));
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
#elif defined(__VMS)
block = !block;
Patch #1516912: improve Modules support for OpenVMS.
2006-07-10 01:18:57 +00:00
ioctl(s->sock_fd, FIONBIO, (unsigned int *)&block);
#else /* !PYOS_OS2 && !__VMS */
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
delay_flag = fcntl(s->sock_fd, F_GETFL, 0);
if (block)
Use O_NONBLOCK rather than O_NDELAY, so we get POSIX non-blocking I/O. On HPUX, Solaris, Tru64 (Dec UNIX), and IRIX (I think), O_NONBLOCK is the POSIX version of non-blocking I/O which is what we want. On Linux and FreeBSD (at least), O_NONBLOCK and O_NDELAY are the same. So this change should have no negative effect on those platforms. Tested on Linux, Solaris, HPUX. Thanks to Anders Qvist for diagnosing this problem.
2002-11-02 19:55:21 +00:00
delay_flag &= (~O_NONBLOCK);
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
else
Use O_NONBLOCK rather than O_NDELAY, so we get POSIX non-blocking I/O. On HPUX, Solaris, Tru64 (Dec UNIX), and IRIX (I think), O_NONBLOCK is the POSIX version of non-blocking I/O which is what we want. On Linux and FreeBSD (at least), O_NONBLOCK and O_NDELAY are the same. So this change should have no negative effect on those platforms. Tested on Linux, Solaris, HPUX. Thanks to Anders Qvist for diagnosing this problem.
2002-11-02 19:55:21 +00:00
delay_flag |= O_NONBLOCK;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
fcntl(s->sock_fd, F_SETFL, delay_flag);
#endif /* !PYOS_OS2 */
#else /* MS_WINDOWS */
block = !block;
ioctlsocket(s->sock_fd, FIONBIO, (u_long*)&block);
#endif /* MS_WINDOWS */
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
#else /* RISCOS */
block = !block;
socketioctl(s->sock_fd, FIONBIO, (u_long*)&block);
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
#endif /* RISCOS */
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
#endif /* __BEOS__ */
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Py_END_ALLOW_THREADS
/* Since these don't return anything */
return 1;
}
#1494314: Fix a regression with high-numbered sockets in 2.4.3. This means that select() on sockets > FD_SETSIZE (typically 1024) work again. The patch makes sockets use poll() internally where available.
2006-07-11 02:04:09 +00:00
/* Do a select()/poll() on the socket, if necessary (sock_timeout > 0).
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
The argument writing indicates the direction.
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
This does not raise an exception; we'll let our caller do that
after they've reacquired the interpreter lock.
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
Returns 1 on timeout, -1 on error, 0 otherwise. */
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
static int
Correct several blunders in the timeout code, mostly my own fault (for not testing it -- apparently test_timeout.py doesn't test anything useful): In internal_select(): - The tv_usec part of the timeout for select() was calculated wrong. - The first argument to select() was one too low. - The sense of the direction argument to internal_select() was inverted. In PySocketSock_settimeout(): - The calls to internal_setblocking() were swapped. Also, repaired some comments and fixed the test for the return value of internal_select() in sendall -- this was in the original patch.
2002-06-07 01:42:47 +00:00
internal_select(PySocketSockObject *s, int writing)
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
{
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
int n;
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
Bail out early from internal_select() when socket file descriptor closed. Prevents core dump.
2002-07-19 12:44:59 +00:00
/* Nothing to do unless we're in timeout mode (not non-blocking) */
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
if (s->sock_timeout <= 0.0)
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
return 0;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Bail out early from internal_select() when socket file descriptor closed. Prevents core dump.
2002-07-19 12:44:59 +00:00
/* Guard against closed socket */
if (s->sock_fd < 0)
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
return 0;
Bail out early from internal_select() when socket file descriptor closed. Prevents core dump.
2002-07-19 12:44:59 +00:00
#1494314: Fix a regression with high-numbered sockets in 2.4.3. This means that select() on sockets > FD_SETSIZE (typically 1024) work again. The patch makes sockets use poll() internally where available.
2006-07-11 02:04:09 +00:00
/* Prefer poll, if available, since you can poll() any fd
* which can't be done with select(). */
#ifdef HAVE_POLL
{
struct pollfd pollfd;
int timeout;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
#1494314: Fix a regression with high-numbered sockets in 2.4.3. This means that select() on sockets > FD_SETSIZE (typically 1024) work again. The patch makes sockets use poll() internally where available.
2006-07-11 02:04:09 +00:00
pollfd.fd = s->sock_fd;
pollfd.events = writing ? POLLOUT : POLLIN;
/* s->sock_timeout is in seconds, timeout in ms */
timeout = (int)(s->sock_timeout * 1000 + 0.5);
n = poll(&pollfd, 1, timeout);
}
#else
{
/* Construct the arguments to select */
fd_set fds;
struct timeval tv;
tv.tv_sec = (int)s->sock_timeout;
tv.tv_usec = (int)((s->sock_timeout - tv.tv_sec) * 1e6);
FD_ZERO(&fds);
FD_SET(s->sock_fd, &fds);
/* See if the socket is ready */
if (writing)
n = select(s->sock_fd+1, NULL, &fds, NULL, &tv);
else
n = select(s->sock_fd+1, &fds, NULL, NULL, &tv);
}
#endif
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
if (n < 0)
return -1;
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
if (n == 0)
return 1;
return 0;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
}
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
/* Initialize a new socket object. */
A Python float is a C double; redeclare defaulttimeout as such; stops compiler wngs on Windows.
2002-07-18 22:38:44 +00:00
static double defaulttimeout = -1.0; /* Default timeout for new sockets */
Add default timeout functionality. This adds setdefaulttimeout() and getdefaulttimeout() functions to the socket and _socket modules, and appropriate tests.
2002-07-18 17:08:35 +00:00
Replace DL_IMPORT with PyMODINIT_FUNC and remove "/export:init..." link command line for Windows builds. This should allow MSVC to import and build the Python MSVC6 project files without error.
2002-07-23 06:31:15 +00:00
PyMODINIT_FUNC
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
init_sockobject(PySocketSockObject *s,
SOCKET_T fd, int family, int type, int proto)
{
#ifdef RISCOS
int block = 1;
#endif
s->sock_fd = fd;
s->sock_family = family;
s->sock_type = type;
s->sock_proto = proto;
Add default timeout functionality. This adds setdefaulttimeout() and getdefaulttimeout() functions to the socket and _socket modules, and appropriate tests.
2002-07-18 17:08:35 +00:00
s->sock_timeout = defaulttimeout;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
s->errorhandler = &set_error;
Add default timeout functionality. This adds setdefaulttimeout() and getdefaulttimeout() functions to the socket and _socket modules, and appropriate tests.
2002-07-18 17:08:35 +00:00
if (defaulttimeout >= 0.0)
internal_setblocking(s, 0);
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
#ifdef RISCOS
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
if (taskwindow)
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
socketioctl(s->sock_fd, 0x80046679, (u_long*)&block);
#endif
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* Create a new socket object.
This just creates the object and initializes it.
If the creation fails, return NULL and set an exception (implicit
in NEWOBJ()). */
new style
1994-10-10 17:59:00 +00:00
static PySocketSockObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
new_sockobject(SOCKET_T fd, int family, int type, int proto)
Initial revision
1991-06-25 21:36:08 +00:00
{
new style
1994-10-10 17:59:00 +00:00
PySocketSockObject *s;
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
s = (PySocketSockObject *)
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyType_GenericNew(&sock_type, NULL, NULL);
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
if (s != NULL)
init_sockobject(s, fd, family, type, proto);
Initial revision
1991-06-25 21:36:08 +00:00
return s;
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
/* Lock to allow python interpreter to continue, but only allow one
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
thread to be in gethostbyname or getaddrinfo */
#if defined(USE_GETHOSTBYNAME_LOCK) || defined(USE_GETADDRINFO_LOCK)
PyThread_type_lock netdb_lock;
Threading speedup patches by Christopher Lindblad <cjl@infoseek.com>.
1998-04-09 20:56:35 +00:00
#endif
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* Convert a string specifying a host name or one of a few symbolic
names to a numeric IP address. This usually calls gethostbyname()
to do the work; the names "" and "<broadcast>" are special.
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
Return the length (IPv4 should be 4 bytes), or negative if
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
an error occurred; then an exception is raised. */
Initial revision
1991-06-25 21:36:08 +00:00
static int
Pass length of result structure into setipaddr. Fixes bug #565747. Committed to 2.2 branch.
2002-07-28 16:10:31 +00:00
setipaddr(char *name, struct sockaddr *addr_ret, size_t addr_ret_size, int af)
Initial revision
1991-06-25 21:36:08 +00:00
{
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
struct addrinfo hints, *res;
int error;
Patch 731209: Restore socketmodule's behaviour with dotted quad addresses to that of Python2.1. Such nnn.nnn.nnn.nnn addresses are just used directly, not passed to the resolver for a pointless lookup.
2003-05-02 15:40:46 +00:00
int d1, d2, d3, d4;
char ch;
Initial revision
1991-06-25 21:36:08 +00:00
Keep gcc -Wall happy.
1996-12-05 23:43:35 +00:00
memset((void *) addr_ret, '\0', sizeof(*addr_ret));
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
if (name[0] == '\0') {
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
int siz;
memset(&hints, 0, sizeof(hints));
hints.ai_family = af;
hints.ai_socktype = SOCK_DGRAM; /*dummy*/
hints.ai_flags = AI_PASSIVE;
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
Py_BEGIN_ALLOW_THREADS
ACQUIRE_GETADDRINFO_LOCK
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
error = getaddrinfo(NULL, "0", &hints, &res);
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
Py_END_ALLOW_THREADS
/* We assume that those thread-unsafe getaddrinfo() versions
*are* safe regarding their return value, ie. that a
subsequent call to getaddrinfo() does not destroy the
outcome of the first call. */
RELEASE_GETADDRINFO_LOCK
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
if (error) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
set_gaierror(error);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return -1;
}
switch (res->ai_family) {
case AF_INET:
siz = 4;
break;
Remove INET6 define. Use ENABLE_IPV6 instead.
2001-12-02 10:15:37 +00:00
#ifdef ENABLE_IPV6
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
case AF_INET6:
siz = 16;
break;
#endif
default:
freeaddrinfo(res);
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error,
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
"unsupported address family");
return -1;
}
if (res->ai_next) {
Fix memory leaks detecting in bug report #478003.
2001-11-07 08:31:03 +00:00
freeaddrinfo(res);
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error,
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
"wildcard resolved to multiple address");
return -1;
}
Pass length of result structure into setipaddr. Fixes bug #565747. Committed to 2.2 branch.
2002-07-28 16:10:31 +00:00
if (res->ai_addrlen < addr_ret_size)
addr_ret_size = res->ai_addrlen;
memcpy(addr_ret, res->ai_addr, addr_ret_size);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
freeaddrinfo(res);
return siz;
Initial revision
1991-06-25 21:36:08 +00:00
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
if (name[0] == '<' && strcmp(name, "<broadcast>") == 0) {
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
struct sockaddr_in *sin;
Patch #650422: Use Posix AF_ constants instead of PF_ ones.
2002-12-11 13:10:57 +00:00
if (af != AF_INET && af != AF_UNSPEC) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error,
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
"address family mismatched");
return -1;
}
sin = (struct sockaddr_in *)addr_ret;
memset((void *) sin, '\0', sizeof(*sin));
sin->sin_family = AF_INET;
#ifdef HAVE_SOCKADDR_SA_LEN
sin->sin_len = sizeof(*sin);
#endif
sin->sin_addr.s_addr = INADDR_BROADCAST;
return sizeof(sin->sin_addr);
Initial revision
1991-06-25 21:36:08 +00:00
}
Patch 731209: Restore socketmodule's behaviour with dotted quad addresses to that of Python2.1. Such nnn.nnn.nnn.nnn addresses are just used directly, not passed to the resolver for a pointless lookup.
2003-05-02 15:40:46 +00:00
if (sscanf(name, "%d.%d.%d.%d%c", &d1, &d2, &d3, &d4, &ch) == 4 &&
0 <= d1 && d1 <= 255 && 0 <= d2 && d2 <= 255 &&
0 <= d3 && d3 <= 255 && 0 <= d4 && d4 <= 255) {
struct sockaddr_in *sin;
sin = (struct sockaddr_in *)addr_ret;
sin->sin_addr.s_addr = htonl(
((long) d1 << 24) | ((long) d2 << 16) |
((long) d3 << 8) | ((long) d4 << 0));
sin->sin_family = AF_INET;
#ifdef HAVE_SOCKADDR_SA_LEN
sin->sin_len = sizeof(*sin);
#endif
return 4;
}
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
memset(&hints, 0, sizeof(hints));
hints.ai_family = af;
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
Py_BEGIN_ALLOW_THREADS
ACQUIRE_GETADDRINFO_LOCK
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
error = getaddrinfo(name, NULL, &hints, &res);
After discussion with itojun, it was clarified that Tru64 is in error, and that the work-around should be restricted to that system.
2001-10-25 09:04:03 +00:00
#if defined(__digital__) && defined(__unix__)
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
if (error == EAI_NONAME && af == AF_UNSPEC) {
/* On Tru64 V5.1, numeric-to-addr conversion fails
if no address family is given. Assume IPv4 for now.*/
hints.ai_family = AF_INET;
error = getaddrinfo(name, NULL, &hints, &res);
}
After discussion with itojun, it was clarified that Tru64 is in error, and that the work-around should be restricted to that system.
2001-10-25 09:04:03 +00:00
#endif
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
Py_END_ALLOW_THREADS
RELEASE_GETADDRINFO_LOCK /* see comment in setipaddr() */
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
if (error) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
set_gaierror(error);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return -1;
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
}
Pass length of result structure into setipaddr. Fixes bug #565747. Committed to 2.2 branch.
2002-07-28 16:10:31 +00:00
if (res->ai_addrlen < addr_ret_size)
addr_ret_size = res->ai_addrlen;
memcpy((char *) addr_ret, res->ai_addr, addr_ret_size);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
freeaddrinfo(res);
switch (addr_ret->sa_family) {
case AF_INET:
return 4;
Remove INET6 define. Use ENABLE_IPV6 instead.
2001-12-02 10:15:37 +00:00
#ifdef ENABLE_IPV6
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
case AF_INET6:
return 16;
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
#endif
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
default:
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error, "unknown address family");
Initial revision
1991-06-25 21:36:08 +00:00
return -1;
}
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* Create a string object representing an IP address.
This is always a string of the form 'dd.dd.dd.dd' (with variable
size numbers). */
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
makeipaddr(struct sockaddr *addr, int addrlen)
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
{
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
char buf[NI_MAXHOST];
int error;
error = getnameinfo(addr, addrlen, buf, sizeof(buf), NULL, 0,
NI_NUMERICHOST);
if (error) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
set_gaierror(error);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return NULL;
}
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromString(buf);
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
}
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
#ifdef USE_BLUETOOTH
/* Convert a string representation of a Bluetooth address into a numeric
address. Returns the length (6), or raises an exception and returns -1 if
an error occurred. */
static int
setbdaddr(char *name, bdaddr_t *bdaddr)
{
unsigned int b0, b1, b2, b3, b4, b5;
char ch;
int n;
n = sscanf(name, "%X:%X:%X:%X:%X:%X%c",
&b5, &b4, &b3, &b2, &b1, &b0, &ch);
if (n == 6 && (b0 | b1 | b2 | b3 | b4 | b5) < 256) {
bdaddr->b[0] = b0;
bdaddr->b[1] = b1;
bdaddr->b[2] = b2;
bdaddr->b[3] = b3;
bdaddr->b[4] = b4;
bdaddr->b[5] = b5;
return 6;
} else {
PyErr_SetString(socket_error, "bad bluetooth address");
return -1;
}
}
/* Create a string representation of the Bluetooth address. This is always a
string of the form 'XX:XX:XX:XX:XX:XX' where XX is a two digit hexadecimal
value (zero padded if necessary). */
static PyObject *
makebdaddr(bdaddr_t *bdaddr)
{
char buf[(6 * 2) + 5 + 1];
sprintf(buf, "%02X:%02X:%02X:%02X:%02X:%02X",
bdaddr->b[5], bdaddr->b[4], bdaddr->b[3],
bdaddr->b[2], bdaddr->b[1], bdaddr->b[0]);
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromString(buf);
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
}
#endif
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* Create an object representing the given socket address,
suitable for passing it back to bind(), connect() etc.
The family field of the sockaddr structure is inspected
to determine what kind of address it really is. */
Initial revision
1991-06-25 21:36:08 +00:00
/*ARGSUSED*/
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
makesockaddr(int sockfd, struct sockaddr *addr, int addrlen, int proto)
Initial revision
1991-06-25 21:36:08 +00:00
{
Sometimes, recvfrom() returns an empty address; this is not an error. Also, get rid of makepair() in favor of mkvalue().
1992-08-05 19:00:45 +00:00
if (addrlen == 0) {
/* No address -- may be recvfrom() from known socket */
new style
1994-10-10 17:59:00 +00:00
Py_INCREF(Py_None);
return Py_None;
Sometimes, recvfrom() returns an empty address; this is not an error. Also, get rid of makepair() in favor of mkvalue().
1992-08-05 19:00:45 +00:00
}
Changes for BeOS, QNX and long long, by Chris Herborth.
1998-08-04 22:53:56 +00:00
#ifdef __BEOS__
Patch by Charles G Waldman: 1) Adds MSG_DONTWAIT if defined (I needed this) 2) Spells "coreectly" correctly ;-)
2000-04-25 21:34:53 +00:00
/* XXX: BeOS version of accept() doesn't set family correctly */
Changes for BeOS, QNX and long long, by Chris Herborth.
1998-08-04 22:53:56 +00:00
addr->sa_family = AF_INET;
#endif
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
switch (addr->sa_family) {
case AF_INET:
{
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
struct sockaddr_in *a;
PyObject *addrobj = makeipaddr(addr, sizeof(*a));
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
PyObject *ret = NULL;
if (addrobj) {
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
a = (struct sockaddr_in *)addr;
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
ret = Py_BuildValue("Oi", addrobj, ntohs(a->sin_port));
Py_DECREF(addrobj);
}
* ceval.c: ifdef out the last argument passing compat hack. * Fixed memory leaks in socket, select and sv modules: mkvalue("O", v) does INCREF(v) so if v is brand new it should be XDECREF'd
1993-02-05 09:46:15 +00:00
return ret;
Initial revision
1991-06-25 21:36:08 +00:00
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
OS/2's TCP/IP stack supports AF_UNIX sockets, with the limitation that the socket name must start with /socket32/. Unlike Unix systems, this file never exists in the file system.
2004-04-04 07:13:49 +00:00
#if defined(AF_UNIX)
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
case AF_UNIX:
{
Initial revision
1991-06-25 21:36:08 +00:00
struct sockaddr_un *a = (struct sockaddr_un *) addr;
SF Patch #1062014: AF_UNIX sockets under Linux have a special abstract namespace that is now fully supported.
2006-04-19 11:50:27 +00:00
#ifdef linux
if (a->sun_path[0] == 0) { /* Linux abstract namespace */
patch 1754489 by vlahan: improve portability of address length calculation for AF_UNIX sockets
2008-01-23 14:07:13 +00:00
addrlen -= offsetof(struct sockaddr_un, sun_path);
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromStringAndSize(a->sun_path,
SF Patch #1062014: AF_UNIX sockets under Linux have a special abstract namespace that is now fully supported.
2006-04-19 11:50:27 +00:00
addrlen);
}
else
#endif /* linux */
{
/* regular NULL-terminated string */
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromString(a->sun_path);
SF Patch #1062014: AF_UNIX sockets under Linux have a special abstract namespace that is now fully supported.
2006-04-19 11:50:27 +00:00
}
Initial revision
1991-06-25 21:36:08 +00:00
}
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif /* AF_UNIX */
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
#if defined(AF_NETLINK)
case AF_NETLINK:
{
struct sockaddr_nl *a = (struct sockaddr_nl *) addr;
Change build format of netlink addresses.
2006-01-15 07:49:20 +00:00
return Py_BuildValue("II", a->nl_pid, a->nl_groups);
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
}
#endif /* AF_NETLINK */
Remove INET6 define. Use ENABLE_IPV6 instead.
2001-12-02 10:15:37 +00:00
#ifdef ENABLE_IPV6
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
case AF_INET6:
{
struct sockaddr_in6 *a;
PyObject *addrobj = makeipaddr(addr, sizeof(*a));
PyObject *ret = NULL;
if (addrobj) {
a = (struct sockaddr_in6 *)addr;
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
ret = Py_BuildValue("Oiii",
addrobj,
ntohs(a->sin6_port),
a->sin6_flowinfo,
a->sin6_scope_id);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
Py_DECREF(addrobj);
}
return ret;
}
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
#endif
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
#ifdef USE_BLUETOOTH
case AF_BLUETOOTH:
switch (proto) {
case BTPROTO_L2CAP:
{
struct sockaddr_l2 *a = (struct sockaddr_l2 *) addr;
PyObject *addrobj = makebdaddr(&_BT_L2_MEMB(a, bdaddr));
PyObject *ret = NULL;
if (addrobj) {
ret = Py_BuildValue("Oi",
addrobj,
_BT_L2_MEMB(a, psm));
Py_DECREF(addrobj);
}
return ret;
}
case BTPROTO_RFCOMM:
{
struct sockaddr_rc *a = (struct sockaddr_rc *) addr;
PyObject *addrobj = makebdaddr(&_BT_RC_MEMB(a, bdaddr));
PyObject *ret = NULL;
if (addrobj) {
ret = Py_BuildValue("Oi",
addrobj,
_BT_RC_MEMB(a, channel));
Py_DECREF(addrobj);
}
return ret;
}
Patch #1432399: Add HCI sockets.
2007-02-14 10:07:37 +00:00
case BTPROTO_HCI:
{
struct sockaddr_hci *a = (struct sockaddr_hci *) addr;
PyObject *ret = NULL;
ret = Py_BuildValue("i", _BT_HCI_MEMB(a, dev));
return ret;
}
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
#if !defined(__FreeBSD__)
case BTPROTO_SCO:
{
struct sockaddr_sco *a = (struct sockaddr_sco *) addr;
return makebdaddr(&_BT_SCO_MEMB(a, bdaddr));
}
#endif
}
#endif
Autotest for netpacket/packet.h, as it is not available on all Linux versions. Depend AF_PACKET on HAVE_NETPACKET_PACKET_H. Fixes #449157
2001-08-10 20:29:40 +00:00
#ifdef HAVE_NETPACKET_PACKET_H
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
case AF_PACKET:
{
struct sockaddr_ll *a = (struct sockaddr_ll *)addr;
char *ifname = "";
struct ifreq ifr;
fix a couple last-minute bugs in the raw socket support
2001-02-02 19:55:17 +00:00
/* need to look up interface name give index */
if (a->sll_ifindex) {
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
ifr.ifr_ifindex = a->sll_ifindex;
fix a couple last-minute bugs in the raw socket support
2001-02-02 19:55:17 +00:00
if (ioctl(sockfd, SIOCGIFNAME, &ifr) == 0)
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
ifname = ifr.ifr_name;
}
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
return Py_BuildValue("shbhs#",
ifname,
ntohs(a->sll_protocol),
a->sll_pkttype,
a->sll_hatype,
a->sll_addr,
a->sll_halen);
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
}
#endif
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
Issue #1646: Make socket support TIPC. The socket module now has support for TIPC under Linux, see http://tipc.sf.net/ for more information. Thanks to Alberto Bertogli for the patch
2008-01-07 16:12:44 +00:00
#ifdef HAVE_LINUX_TIPC_H
case AF_TIPC:
{
struct sockaddr_tipc *a = (struct sockaddr_tipc *) addr;
if (a->addrtype == TIPC_ADDR_NAMESEQ) {
return Py_BuildValue("IIIII",
a->addrtype,
a->addr.nameseq.type,
a->addr.nameseq.lower,
a->addr.nameseq.upper,
a->scope);
} else if (a->addrtype == TIPC_ADDR_NAME) {
return Py_BuildValue("IIIII",
a->addrtype,
a->addr.name.name.type,
a->addr.name.name.instance,
a->addr.name.name.instance,
a->scope);
} else if (a->addrtype == TIPC_ADDR_ID) {
return Py_BuildValue("IIIII",
a->addrtype,
a->addr.id.node,
a->addr.id.ref,
0,
a->scope);
} else {
PyErr_SetString(PyExc_TypeError,
"Invalid address type");
return NULL;
}
}
#endif
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* More cases here... */
default:
In makesockaddr(), if we don't know the address family, don't raise an exception -- return it as a tuple. Seems useful in promiscuous mode.
1997-05-07 17:41:48 +00:00
/* If we don't know the address family, don't raise an
exception -- return it as a tuple. */
return Py_BuildValue("is#",
addr->sa_family,
addr->sa_data,
sizeof(addr->sa_data));
Sometimes, recvfrom() returns an empty address; this is not an error. Also, get rid of makepair() in favor of mkvalue().
1992-08-05 19:00:45 +00:00
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
}
Initial revision
1991-06-25 21:36:08 +00:00
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* Parse a socket address argument according to the socket object's
address family. Return 1 if the address was in the proper format,
0 of not. The address is returned through addr_ret, its length
through len_ret. */
Initial revision
1991-06-25 21:36:08 +00:00
static int
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
getsockaddrarg(PySocketSockObject *s, PyObject *args,
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
struct sockaddr *addr_ret, int *len_ret)
Initial revision
1991-06-25 21:36:08 +00:00
{
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
switch (s->sock_family) {
OS/2's TCP/IP stack supports AF_UNIX sockets, with the limitation that the socket name must start with /socket32/. Unlike Unix systems, this file never exists in the file system.
2004-04-04 07:13:49 +00:00
#if defined(AF_UNIX)
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
case AF_UNIX:
{
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
struct sockaddr_un* addr;
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
char *path;
Initial revision
1991-06-25 21:36:08 +00:00
int len;
Use the t# format where appropriate. Greg Stein.
1998-10-08 02:25:24 +00:00
if (!PyArg_Parse(args, "t#", &path, &len))
Initial revision
1991-06-25 21:36:08 +00:00
return 0;
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
addr = (struct sockaddr_un*)addr_ret;
SF Patch #1062014: AF_UNIX sockets under Linux have a special abstract namespace that is now fully supported.
2006-04-19 11:50:27 +00:00
#ifdef linux
if (len > 0 && path[0] == 0) {
/* Linux abstract namespace extension */
if (len > sizeof addr->sun_path) {
PyErr_SetString(socket_error,
"AF_UNIX path too long");
return 0;
}
}
else
#endif /* linux */
{
/* regular NULL-terminated string */
if (len >= sizeof addr->sun_path) {
PyErr_SetString(socket_error,
"AF_UNIX path too long");
return 0;
}
addr->sun_path[len] = 0;
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
}
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
addr->sun_family = s->sock_family;
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
memcpy(addr->sun_path, path, len);
Fixes for AF_UNIX support on OS/2: - return the full size of the sockaddr_un structure, without which bind() fails with EINVAL; - set test_socketserver to use a socket name that meets the form required by the underlying implementation; - don't bother exercising the forking AF_UNIX tests on EMX - its fork() can't handle the stress.
2004-04-11 12:03:57 +00:00
#if defined(PYOS_OS2)
*len_ret = sizeof(*addr);
#else
patch 1754489 by vlahan: improve portability of address length calculation for AF_UNIX sockets
2008-01-23 14:07:13 +00:00
*len_ret = len + offsetof(struct sockaddr_un, sun_path);
Fixes for AF_UNIX support on OS/2: - return the full size of the sockaddr_un structure, without which bind() fails with EINVAL; - set test_socketserver to use a socket name that meets the form required by the underlying implementation; - don't bother exercising the forking AF_UNIX tests on EMX - its fork() can't handle the stress.
2004-04-11 12:03:57 +00:00
#endif
Initial revision
1991-06-25 21:36:08 +00:00
return 1;
}
Merge alpha100 branch back to main trunk
1994-08-01 11:34:53 +00:00
#endif /* AF_UNIX */
Initial revision
1991-06-25 21:36:08 +00:00
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
#if defined(AF_NETLINK)
case AF_NETLINK:
{
struct sockaddr_nl* addr;
int pid, groups;
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
addr = (struct sockaddr_nl *)addr_ret;
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
if (!PyTuple_Check(args)) {
PyErr_Format(
PyExc_TypeError,
"getsockaddrarg: "
"AF_NETLINK address must be tuple, not %.500s",
#1629: Renamed Py_Size, Py_Type and Py_Refcnt to Py_SIZE, Py_TYPE and Py_REFCNT. Macros for b/w compatibility are available.
2007-12-19 02:37:44 +00:00
Py_TYPE(args)->tp_name);
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
return 0;
}
if (!PyArg_ParseTuple(args, "II:getsockaddrarg", &pid, &groups))
return 0;
addr->nl_family = AF_NETLINK;
addr->nl_pid = pid;
addr->nl_groups = groups;
*len_ret = sizeof(*addr);
return 1;
}
#endif
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
case AF_INET:
{
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
struct sockaddr_in* addr;
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
char *host;
Release host name memory. Fixes #783312. Will backport to 2.3.
2003-08-07 11:55:15 +00:00
int port, result;
Patch by Michael Hudson to clarify the error message from getsockaddrarg when the address is not a tuple.
2000-12-01 13:13:11 +00:00
if (!PyTuple_Check(args)) {
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
PyErr_Format(
PyExc_TypeError,
"getsockaddrarg: "
"AF_INET address must be tuple, not %.500s",
#1629: Renamed Py_Size, Py_Type and Py_Refcnt to Py_SIZE, Py_TYPE and Py_REFCNT. Macros for b/w compatibility are available.
2007-12-19 02:37:44 +00:00
Py_TYPE(args)->tp_name);
Patch by Michael Hudson to clarify the error message from getsockaddrarg when the address is not a tuple.
2000-12-01 13:13:11 +00:00
return 0;
}
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
if (!PyArg_ParseTuple(args, "eti:getsockaddrarg",
Implement IDNA (Internationalized Domain Names in Applications).
2003-04-18 10:39:54 +00:00
"idna", &host, &port))
Initial revision
1991-06-25 21:36:08 +00:00
return 0;
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
addr=(struct sockaddr_in*)addr_ret;
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
result = setipaddr(host, (struct sockaddr *)addr,
Release host name memory. Fixes #783312. Will backport to 2.3.
2003-08-07 11:55:15 +00:00
sizeof(*addr), AF_INET);
PyMem_Free(host);
if (result < 0)
Initial revision
1991-06-25 21:36:08 +00:00
return 0;
- Issue #5104: The socket module now raises OverflowError when 16-bit port and protocol numbers are supplied outside the allowed 0-65536 range on bind() and getservbyport().
2009-01-31 22:57:30 +00:00
if (port < 0 || port > 0xffff) {
PyErr_SetString(
PyExc_OverflowError,
"getsockaddrarg: port must be 0-65535.");
return 0;
}
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
addr->sin_family = AF_INET;
Tweaks to keep the Microsoft compiler quier.
1997-04-09 19:24:53 +00:00
addr->sin_port = htons((short)port);
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
*len_ret = sizeof *addr;
Initial revision
1991-06-25 21:36:08 +00:00
return 1;
}
Remove INET6 define. Use ENABLE_IPV6 instead.
2001-12-02 10:15:37 +00:00
#ifdef ENABLE_IPV6
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
case AF_INET6:
{
struct sockaddr_in6* addr;
char *host;
Release host name memory. Fixes #783312. Will backport to 2.3.
2003-08-07 11:55:15 +00:00
int port, flowinfo, scope_id, result;
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
flowinfo = scope_id = 0;
Fix SF bug #1476111: SystemError in socket sendto. The AF_INET6 and AF_PACKET cases in getsockaddrarg were missing their own checks for tuple-ness of the address argument, which means a confusing SystemError was raised by PyArg_ParseTuple instead.
2006-04-25 15:08:10 +00:00
if (!PyTuple_Check(args)) {
PyErr_Format(
PyExc_TypeError,
"getsockaddrarg: "
"AF_INET6 address must be tuple, not %.500s",
#1629: Renamed Py_Size, Py_Type and Py_Refcnt to Py_SIZE, Py_TYPE and Py_REFCNT. Macros for b/w compatibility are available.
2007-12-19 02:37:44 +00:00
Py_TYPE(args)->tp_name);
Fix SF bug #1476111: SystemError in socket sendto. The AF_INET6 and AF_PACKET cases in getsockaddrarg were missing their own checks for tuple-ness of the address argument, which means a confusing SystemError was raised by PyArg_ParseTuple instead.
2006-04-25 15:08:10 +00:00
return 0;
}
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
if (!PyArg_ParseTuple(args, "eti|ii",
Implement IDNA (Internationalized Domain Names in Applications).
2003-04-18 10:39:54 +00:00
"idna", &host, &port, &flowinfo,
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
&scope_id)) {
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return 0;
}
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
addr = (struct sockaddr_in6*)addr_ret;
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
result = setipaddr(host, (struct sockaddr *)addr,
Release host name memory. Fixes #783312. Will backport to 2.3.
2003-08-07 11:55:15 +00:00
sizeof(*addr), AF_INET6);
PyMem_Free(host);
if (result < 0)
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return 0;
- Issue #5104: The socket module now raises OverflowError when 16-bit port and protocol numbers are supplied outside the allowed 0-65536 range on bind() and getservbyport().
2009-01-31 22:57:30 +00:00
if (port < 0 || port > 0xffff) {
PyErr_SetString(
PyExc_OverflowError,
"getsockaddrarg: port must be 0-65535.");
return 0;
}
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
addr->sin6_family = s->sock_family;
addr->sin6_port = htons((short)port);
addr->sin6_flowinfo = flowinfo;
addr->sin6_scope_id = scope_id;
*len_ret = sizeof *addr;
return 1;
}
#endif
Add FreeBSD support for bluetooth sockets. (SF Patch #888148, reviewed by loewis)
2004-02-02 06:05:24 +00:00
#ifdef USE_BLUETOOTH
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
case AF_BLUETOOTH:
{
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
switch (s->sock_proto) {
case BTPROTO_L2CAP:
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
{
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
struct sockaddr_l2 *addr;
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
char *straddr;
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
addr = (struct sockaddr_l2 *)addr_ret;
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
_BT_L2_MEMB(addr, family) = AF_BLUETOOTH;
if (!PyArg_ParseTuple(args, "si", &straddr,
&_BT_L2_MEMB(addr, psm))) {
PyErr_SetString(socket_error, "getsockaddrarg: "
"wrong format");
return 0;
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
}
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
if (setbdaddr(straddr, &_BT_L2_MEMB(addr, bdaddr)) < 0)
return 0;
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
*len_ret = sizeof *addr;
return 1;
}
case BTPROTO_RFCOMM:
{
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
struct sockaddr_rc *addr;
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
char *straddr;
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
addr = (struct sockaddr_rc *)addr_ret;
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
_BT_RC_MEMB(addr, family) = AF_BLUETOOTH;
if (!PyArg_ParseTuple(args, "si", &straddr,
&_BT_RC_MEMB(addr, channel))) {
PyErr_SetString(socket_error, "getsockaddrarg: "
"wrong format");
return 0;
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
}
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
if (setbdaddr(straddr, &_BT_RC_MEMB(addr, bdaddr)) < 0)
return 0;
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
*len_ret = sizeof *addr;
return 1;
}
Patch #1432399: Add HCI sockets.
2007-02-14 10:07:37 +00:00
case BTPROTO_HCI:
{
Update 1432399 to removal of _BT_SOCKADDR_MEMB.
2007-02-14 11:30:07 +00:00
struct sockaddr_hci *addr = (struct sockaddr_hci *)addr_ret;
Patch #1432399: Add HCI sockets.
2007-02-14 10:07:37 +00:00
_BT_HCI_MEMB(addr, family) = AF_BLUETOOTH;
if (!PyArg_ParseTuple(args, "i", &_BT_HCI_MEMB(addr, dev))) {
PyErr_SetString(socket_error, "getsockaddrarg: "
"wrong format");
return 0;
}
*len_ret = sizeof *addr;
return 1;
}
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
#if !defined(__FreeBSD__)
case BTPROTO_SCO:
{
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
struct sockaddr_sco *addr;
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
char *straddr;
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
addr = (struct sockaddr_sco *)addr_ret;
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
_BT_SCO_MEMB(addr, family) = AF_BLUETOOTH;
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
straddr = PyString_AsString(args);
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
if (straddr == NULL) {
PyErr_SetString(socket_error, "getsockaddrarg: "
"wrong format");
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
return 0;
}
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
if (setbdaddr(straddr, &_BT_SCO_MEMB(addr, bdaddr)) < 0)
return 0;
*len_ret = sizeof *addr;
return 1;
}
#endif
default:
PyErr_SetString(socket_error, "getsockaddrarg: unknown Bluetooth protocol");
return 0;
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
}
}
#endif
Autotest for netpacket/packet.h, as it is not available on all Linux versions. Depend AF_PACKET on HAVE_NETPACKET_PACKET_H. Fixes #449157
2001-08-10 20:29:40 +00:00
#ifdef HAVE_NETPACKET_PACKET_H
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
case AF_PACKET:
{
struct sockaddr_ll* addr;
struct ifreq ifr;
char *interfaceName;
int protoNumber;
int hatype = 0;
int pkttype = 0;
[Patch #947352 from Jason Andryuk] Add support for AF_PACKET hardware addresses
2004-07-10 23:39:35 +00:00
char *haddr = NULL;
unsigned int halen = 0;
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
Fix SF bug #1476111: SystemError in socket sendto. The AF_INET6 and AF_PACKET cases in getsockaddrarg were missing their own checks for tuple-ness of the address argument, which means a confusing SystemError was raised by PyArg_ParseTuple instead.
2006-04-25 15:08:10 +00:00
if (!PyTuple_Check(args)) {
PyErr_Format(
PyExc_TypeError,
"getsockaddrarg: "
"AF_PACKET address must be tuple, not %.500s",
#1629: Renamed Py_Size, Py_Type and Py_Refcnt to Py_SIZE, Py_TYPE and Py_REFCNT. Macros for b/w compatibility are available.
2007-12-19 02:37:44 +00:00
Py_TYPE(args)->tp_name);
Fix SF bug #1476111: SystemError in socket sendto. The AF_INET6 and AF_PACKET cases in getsockaddrarg were missing their own checks for tuple-ness of the address argument, which means a confusing SystemError was raised by PyArg_ParseTuple instead.
2006-04-25 15:08:10 +00:00
return 0;
}
[Patch #947352 from Jason Andryuk] Add support for AF_PACKET hardware addresses
2004-07-10 23:39:35 +00:00
if (!PyArg_ParseTuple(args, "si|iis#", &interfaceName,
&protoNumber, &pkttype, &hatype,
&haddr, &halen))
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
return 0;
strncpy(ifr.ifr_name, interfaceName, sizeof(ifr.ifr_name));
ifr.ifr_name[(sizeof(ifr.ifr_name))-1] = '\0';
fix a couple last-minute bugs in the raw socket support
2001-02-02 19:55:17 +00:00
if (ioctl(s->sock_fd, SIOCGIFINDEX, &ifr) < 0) {
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
s->errorhandler();
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
return 0;
fix a couple last-minute bugs in the raw socket support
2001-02-02 19:55:17 +00:00
}
[Patch #947352 from Jason Andryuk] Add support for AF_PACKET hardware addresses
2004-07-10 23:39:35 +00:00
if (halen > 8) {
PyErr_SetString(PyExc_ValueError,
"Hardware address must be 8 bytes or less");
return 0;
}
- Issue #5104: The socket module now raises OverflowError when 16-bit port and protocol numbers are supplied outside the allowed 0-65536 range on bind() and getservbyport().
2009-01-31 22:57:30 +00:00
if (protoNumber < 0 || protoNumber > 0xffff) {
PyErr_SetString(
PyExc_OverflowError,
"getsockaddrarg: protoNumber must be 0-65535.");
return 0;
}
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
addr = (struct sockaddr_ll*)addr_ret;
addr->sll_family = AF_PACKET;
addr->sll_protocol = htons((short)protoNumber);
addr->sll_ifindex = ifr.ifr_ifindex;
addr->sll_pkttype = pkttype;
addr->sll_hatype = hatype;
[Patch #947352 from Jason Andryuk] Add support for AF_PACKET hardware addresses
2004-07-10 23:39:35 +00:00
if (halen != 0) {
memcpy(&addr->sll_addr, haddr, halen);
}
addr->sll_halen = halen;
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
*len_ret = sizeof *addr;
return 1;
}
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
#endif
Issue #1646: Make socket support TIPC. The socket module now has support for TIPC under Linux, see http://tipc.sf.net/ for more information. Thanks to Alberto Bertogli for the patch
2008-01-07 16:12:44 +00:00
#ifdef HAVE_LINUX_TIPC_H
case AF_TIPC:
{
unsigned int atype, v1, v2, v3;
unsigned int scope = TIPC_CLUSTER_SCOPE;
struct sockaddr_tipc *addr;
if (!PyTuple_Check(args)) {
PyErr_Format(
PyExc_TypeError,
"getsockaddrarg: "
"AF_TIPC address must be tuple, not %.500s",
Py_TYPE(args)->tp_name);
return 0;
}
if (!PyArg_ParseTuple(args,
"IIII|I;Invalid TIPC address format",
&atype, &v1, &v2, &v3, &scope))
return 0;
addr = (struct sockaddr_tipc *) addr_ret;
memset(addr, 0, sizeof(struct sockaddr_tipc));
addr->family = AF_TIPC;
addr->scope = scope;
addr->addrtype = atype;
if (atype == TIPC_ADDR_NAMESEQ) {
addr->addr.nameseq.type = v1;
addr->addr.nameseq.lower = v2;
addr->addr.nameseq.upper = v3;
} else if (atype == TIPC_ADDR_NAME) {
addr->addr.name.name.type = v1;
addr->addr.name.name.instance = v2;
} else if (atype == TIPC_ADDR_ID) {
addr->addr.id.node = v1;
addr->addr.id.ref = v2;
} else {
/* Shouldn't happen */
PyErr_SetString(PyExc_TypeError, "Invalid address type");
return 0;
}
*len_ret = sizeof(*addr);
return 1;
}
#endif
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* More cases here... */
default:
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error, "getsockaddrarg: bad family");
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
return 0;
}
Initial revision
1991-06-25 21:36:08 +00:00
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
/* Get the address length according to the socket object's address family.
Some UNIX types want the exact size of the address structure
1992-06-12 10:39:36 +00:00
Return 1 if the family is known, 0 otherwise. The length is returned
through len_ret. */
static int
Errare humanum est. Changed parameter in getsockaddrlen from unsigned to socklen_t.
2000-07-11 23:00:12 +00:00
getsockaddrlen(PySocketSockObject *s, socklen_t *len_ret)
Some UNIX types want the exact size of the address structure
1992-06-12 10:39:36 +00:00
{
switch (s->sock_family) {
OS/2's TCP/IP stack supports AF_UNIX sockets, with the limitation that the socket name must start with /socket32/. Unlike Unix systems, this file never exists in the file system.
2004-04-04 07:13:49 +00:00
#if defined(AF_UNIX)
Some UNIX types want the exact size of the address structure
1992-06-12 10:39:36 +00:00
case AF_UNIX:
{
*len_ret = sizeof (struct sockaddr_un);
return 1;
}
Merge alpha100 branch back to main trunk
1994-08-01 11:34:53 +00:00
#endif /* AF_UNIX */
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
#if defined(AF_NETLINK)
case AF_NETLINK:
{
*len_ret = sizeof (struct sockaddr_nl);
return 1;
}
#endif
Some UNIX types want the exact size of the address structure
1992-06-12 10:39:36 +00:00
case AF_INET:
{
*len_ret = sizeof (struct sockaddr_in);
return 1;
}
Remove INET6 define. Use ENABLE_IPV6 instead.
2001-12-02 10:15:37 +00:00
#ifdef ENABLE_IPV6
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
case AF_INET6:
{
*len_ret = sizeof (struct sockaddr_in6);
return 1;
}
#endif
Add FreeBSD support for bluetooth sockets. (SF Patch #888148, reviewed by loewis)
2004-02-02 06:05:24 +00:00
#ifdef USE_BLUETOOTH
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
case AF_BLUETOOTH:
{
switch(s->sock_proto)
{
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
case BTPROTO_L2CAP:
*len_ret = sizeof (struct sockaddr_l2);
return 1;
case BTPROTO_RFCOMM:
*len_ret = sizeof (struct sockaddr_rc);
return 1;
Patch #1432399: Add HCI sockets.
2007-02-14 10:07:37 +00:00
case BTPROTO_HCI:
*len_ret = sizeof (struct sockaddr_hci);
return 1;
Add FreeBSD support for bluetooth sockets. (SF Patch #888148, reviewed by loewis)
2004-02-02 06:05:24 +00:00
#if !defined(__FreeBSD__)
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
case BTPROTO_SCO:
*len_ret = sizeof (struct sockaddr_sco);
return 1;
Add FreeBSD support for bluetooth sockets. (SF Patch #888148, reviewed by loewis)
2004-02-02 06:05:24 +00:00
#endif
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
default:
PyErr_SetString(socket_error, "getsockaddrlen: "
"unknown BT protocol");
return 0;
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
}
}
#endif
Autotest for netpacket/packet.h, as it is not available on all Linux versions. Depend AF_PACKET on HAVE_NETPACKET_PACKET_H. Fixes #449157
2001-08-10 20:29:40 +00:00
#ifdef HAVE_NETPACKET_PACKET_H
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
case AF_PACKET:
{
*len_ret = sizeof (struct sockaddr_ll);
return 1;
}
#endif
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
Issue #1646: Make socket support TIPC. The socket module now has support for TIPC under Linux, see http://tipc.sf.net/ for more information. Thanks to Alberto Bertogli for the patch
2008-01-07 16:12:44 +00:00
#ifdef HAVE_LINUX_TIPC_H
case AF_TIPC:
{
*len_ret = sizeof (struct sockaddr_tipc);
return 1;
}
#endif
Some UNIX types want the exact size of the address structure
1992-06-12 10:39:36 +00:00
/* More cases here... */
default:
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error, "getsockaddrlen: bad family");
Some UNIX types want the exact size of the address structure
1992-06-12 10:39:36 +00:00
return 0;
}
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* s.accept() method */
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_accept(PySocketSockObject *s)
Initial revision
1991-06-25 21:36:08 +00:00
{
- Changes donated by Elemental Security to make it work on AIX 5.3 with IBM's 64-bit compiler (SF patch #1284289). This also closes SF bug #105470: test_pwd fails on 64bit system (Opteron).
2005-09-14 18:09:42 +00:00
sock_addr_t addrbuf;
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
SOCKET_T newfd;
Jack Jansen: The GUSI 2.0 I/O library used on the Mac uses the socklen_t (unsigned int) for most size parameters. Apparently this is part of the UNIX 98 standard. [GvR: the changes to configure.in etc. that I just checked in make sure that socklen_t is defined everywhere, so I deleted the little part of Jack's mod to define socklen_t if not in GUSI2. I suppose I will have to add it to the Windows config.h in a minute.]
2000-04-24 15:16:03 +00:00
socklen_t addrlen;
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
PyObject *sock = NULL;
PyObject *addr = NULL;
PyObject *res = NULL;
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
int timeout;
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
Some UNIX types want the exact size of the address structure
1992-06-12 10:39:36 +00:00
if (!getsockaddrlen(s, &addrlen))
return NULL;
- Changes donated by Elemental Security to make it work on AIX 5.3 with IBM's 64-bit compiler (SF patch #1284289). This also closes SF bug #105470: test_pwd fails on 64bit system (Opteron).
2005-09-14 18:09:42 +00:00
memset(&addrbuf, 0, addrlen);
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
#ifdef MS_WINDOWS
newfd = INVALID_SOCKET;
#else
newfd = -1;
#endif
Bug #876637, prevent stack corruption when socket descriptor is larger than FD_SETSIZE. This can only be acheived with ulimit -n SOME_NUMBER_BIGGER_THAN_FD_SETSIZE which is typically only available to root. Since this wouldn't normally be run in a test (ie, run as root), it doesn't seem too worthwhile to add a normal test. The bug report has one version of a test. I've written another. Not sure what the best thing to do is. Do the check before calling internal_select() because we can't set an error in between Py_BEGIN_ALLOW_THREADS and Py_END_ALLOW_THREADS. This seemed the clearest solution, ie handle before calling internal_select() rather than inside. Plus there is at least one place outside of internal_select() that needed to be handled. Will backport.
2006-02-07 07:04:46 +00:00
if (!IS_SELECTABLE(s))
return select_error();
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
timeout = internal_select(s, 0);
if (!timeout)
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
newfd = accept(s->sock_fd, SAS2SA(&addrbuf), &addrlen);
new style
1994-10-10 17:59:00 +00:00
Py_END_ALLOW_THREADS
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
if (timeout == 1) {
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
PyErr_SetString(socket_timeout, "timed out");
return NULL;
}
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
#ifdef MS_WINDOWS
if (newfd == INVALID_SOCKET)
#else
Initial revision
1991-06-25 21:36:08 +00:00
if (newfd < 0)
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
#endif
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* Create the new object with unspecified family,
to avoid calls to bind() etc. on it. */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock = (PyObject *) new_sockobject(newfd,
s->sock_family,
s->sock_type,
s->sock_proto);
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
if (sock == NULL) {
Use an explicit macro SOCKETCLOSE which expands to closesocket (on Windows), soclose (on OS2), or to close (everywhere else). Hopefully this fixes a new compilation error that I suddenly get on Windows because the macro definition for close -> closesocket apparently was done before including io.h, which contains a prototype for close. (No idea why this wasn't an error before.)
2000-04-21 20:33:00 +00:00
SOCKETCLOSE(newfd);
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
goto finally;
}
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
addr = makesockaddr(s->sock_fd, SAS2SA(&addrbuf),
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
addrlen, s->sock_proto);
fix a couple last-minute bugs in the raw socket support
2001-02-02 19:55:17 +00:00
if (addr == NULL)
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
goto finally;
Simplify and speedup uses of Py_BuildValue(): * Py_BuildValue("(OOO)",a,b,c) --> PyTuple_Pack(3,a,b,c) * Py_BuildValue("()",a) --> PyTuple_New(0) * Py_BuildValue("O", a) --> Py_INCREF(a)
2003-10-12 19:09:37 +00:00
res = PyTuple_Pack(2, sock, addr);
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
finally:
new style
1994-10-10 17:59:00 +00:00
Py_XDECREF(sock);
Py_XDECREF(addr);
Initial revision
1991-06-25 21:36:08 +00:00
return res;
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(accept_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"accept() -> (socket object, address info)\n\
\n\
Wait for an incoming connection. Return a new socket representing the\n\
connection, and the address of the client. For IP sockets, the address\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
info is a pair (hostaddr, port).");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
/* s.setblocking(flag) method. Argument:
False -- non-blocking mode; same as settimeout(0)
True -- blocking mode; same as settimeout(None)
*/
Setup.in: added tkinter; rearranged the definition of PYTHONPATH so that the module-specific components are in the section for that module. cursesmodule.c: patched it so it actually works. tkintermodule.c: call Py_AtExit instead of atexit(). signalmodule.c: converted to new naming style; added BGN/END SAVE around pause() call. socketmodule.c: added setblocking() after Tommy Burnette.
1994-09-07 14:32:49 +00:00
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_setblocking(PySocketSockObject *s, PyObject *arg)
Setup.in: added tkinter; rearranged the definition of PYTHONPATH so that the module-specific components are in the section for that module. cursesmodule.c: patched it so it actually works. tkintermodule.c: call Py_AtExit instead of atexit(). signalmodule.c: converted to new naming style; added BGN/END SAVE around pause() call. socketmodule.c: added setblocking() after Tommy Burnette.
1994-09-07 14:32:49 +00:00
{
int block;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Convert socket methods to use METH_O and METH_NOARGS where possible.
2001-10-11 16:36:35 +00:00
block = PyInt_AsLong(arg);
if (block == -1 && PyErr_Occurred())
Setup.in: added tkinter; rearranged the definition of PYTHONPATH so that the module-specific components are in the section for that module. cursesmodule.c: patched it so it actually works. tkintermodule.c: call Py_AtExit instead of atexit(). signalmodule.c: converted to new naming style; added BGN/END SAVE around pause() call. socketmodule.c: added setblocking() after Tommy Burnette.
1994-09-07 14:32:49 +00:00
return NULL;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
s->sock_timeout = block ? -1.0 : 0.0;
I decided to change the interaction between setblocking() and settimeout(). Already, settimeout() canceled non-blocking mode; now, setblocking() also cancels the timeout. This is easier to document. (XXX should settimeout(0) be an alias for setblocking(0)? They seem to have roughly the same effect. Also, I'm not sure that the code in connect() and accept() is correct in all cases. We'll sort this out soon enough.)
2002-06-07 03:36:20 +00:00
internal_setblocking(s, block);
Setup.in: added tkinter; rearranged the definition of PYTHONPATH so that the module-specific components are in the section for that module. cursesmodule.c: patched it so it actually works. tkintermodule.c: call Py_AtExit instead of atexit(). signalmodule.c: converted to new naming style; added BGN/END SAVE around pause() call. socketmodule.c: added setblocking() after Tommy Burnette.
1994-09-07 14:32:49 +00:00
new style
1994-10-10 17:59:00 +00:00
Py_INCREF(Py_None);
return Py_None;
Setup.in: added tkinter; rearranged the definition of PYTHONPATH so that the module-specific components are in the section for that module. cursesmodule.c: patched it so it actually works. tkintermodule.c: call Py_AtExit instead of atexit(). signalmodule.c: converted to new naming style; added BGN/END SAVE around pause() call. socketmodule.c: added setblocking() after Tommy Burnette.
1994-09-07 14:32:49 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(setblocking_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"setblocking(flag)\n\
\n\
Set the socket to blocking (flag is true) or non-blocking (false).\n\
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
setblocking(True) is equivalent to settimeout(None);\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
setblocking(False) is equivalent to settimeout(0.0).");
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
/* s.settimeout(timeout) method. Argument:
None -- no timeout, blocking mode; same as setblocking(True)
0.0 -- non-blocking mode; same as setblocking(False)
> 0 -- timeout mode; operations time out after timeout seconds
< 0 -- illegal; raises an exception
*/
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_settimeout(PySocketSockObject *s, PyObject *arg)
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
{
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
double timeout;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
if (arg == Py_None)
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
timeout = -1.0;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
else {
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
timeout = PyFloat_AsDouble(arg);
if (timeout < 0.0) {
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
if (!PyErr_Occurred())
PyErr_SetString(PyExc_ValueError,
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
"Timeout value out of range");
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
return NULL;
}
}
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
s->sock_timeout = timeout;
internal_setblocking(s, timeout < 0.0);
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Py_INCREF(Py_None);
return Py_None;
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(settimeout_doc,
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
"settimeout(timeout)\n\
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
\n\
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
Set a timeout on socket operations. 'timeout' can be a float,\n\
giving in seconds, or None. Setting a timeout of None disables\n\
the timeout feature and is equivalent to setblocking(1).\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Setting a timeout of zero is the same as setblocking(0).");
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
/* s.gettimeout() method.
Returns the timeout associated with a socket. */
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_gettimeout(PySocketSockObject *s)
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
{
if (s->sock_timeout < 0.0) {
Py_INCREF(Py_None);
return Py_None;
}
else
return PyFloat_FromDouble(s->sock_timeout);
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(gettimeout_doc,
Clean up some docstrings. Some docstrings didn't show their return value; others were inconsistent in what to name the argument or return value; a few module-global functions had "socket." in front of their name, against convention.
2002-08-08 20:37:08 +00:00
"gettimeout() -> timeout\n\
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
\n\
Returns the timeout in floating seconds associated with socket \n\
operations. A timeout of None indicates that timeouts on socket \n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
operations are disabled.");
Setup.in: added tkinter; rearranged the definition of PYTHONPATH so that the module-specific components are in the section for that module. cursesmodule.c: patched it so it actually works. tkintermodule.c: call Py_AtExit instead of atexit(). signalmodule.c: converted to new naming style; added BGN/END SAVE around pause() call. socketmodule.c: added setblocking() after Tommy Burnette.
1994-09-07 14:32:49 +00:00
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
#ifdef RISCOS
/* s.sleeptaskw(1 | 0) method */
static PyObject *
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
sock_sleeptaskw(PySocketSockObject *s,PyObject *arg)
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
{
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
int block;
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
block = PyInt_AsLong(arg);
if (block == -1 && PyErr_Occurred())
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
return NULL;
Py_BEGIN_ALLOW_THREADS
socketioctl(s->sock_fd, 0x80046679, (u_long*)&block);
Py_END_ALLOW_THREADS
Py_INCREF(Py_None);
return Py_None;
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(sleeptaskw_doc,
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
"sleeptaskw(flag)\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Allow sleeps in taskwindows.");
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
#endif
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
/* s.setsockopt() method.
With an integer third argument, sets an integer option.
With a string third argument, sets an option from a buffer;
use optional built-in module 'struct' to encode the string. */
Added {get,set}sockopt(). allowbroadcast is now obsolete.
1992-06-05 15:11:30 +00:00
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_setsockopt(PySocketSockObject *s, PyObject *args)
Added {get,set}sockopt(). allowbroadcast is now obsolete.
1992-06-05 15:11:30 +00:00
{
int level;
int optname;
int res;
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
char *buf;
int buflen;
int flag;
Added {get,set}sockopt(). allowbroadcast is now obsolete.
1992-06-05 15:11:30 +00:00
Fix all routines to use PyArg_ParseTuple(), and add ":name" to the argument format strings. THIS WILL PROBABLY BREAK LOTS OF CODE!!! Also fixed a bogus string in an error message in getsockaddrlen().
2000-03-24 20:56:56 +00:00
if (PyArg_ParseTuple(args, "iii:setsockopt",
&level, &optname, &flag)) {
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
buf = (char *) &flag;
buflen = sizeof flag;
}
else {
new style
1994-10-10 17:59:00 +00:00
PyErr_Clear();
Fix all routines to use PyArg_ParseTuple(), and add ":name" to the argument format strings. THIS WILL PROBABLY BREAK LOTS OF CODE!!! Also fixed a bogus string in an error message in getsockaddrlen().
2000-03-24 20:56:56 +00:00
if (!PyArg_ParseTuple(args, "iis#:setsockopt",
&level, &optname, &buf, &buflen))
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
return NULL;
}
Use 'void' directly instead of the ANY #define, now that all code is ANSI C. Leave the actual #define in for API compatibility.
2000-07-25 12:56:38 +00:00
res = setsockopt(s->sock_fd, level, optname, (void *)buf, buflen);
Added {get,set}sockopt(). allowbroadcast is now obsolete.
1992-06-05 15:11:30 +00:00
if (res < 0)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
new style
1994-10-10 17:59:00 +00:00
Py_INCREF(Py_None);
return Py_None;
Added {get,set}sockopt(). allowbroadcast is now obsolete.
1992-06-05 15:11:30 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(setsockopt_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"setsockopt(level, option, value)\n\
\n\
Set a socket option. See the Unix manual for level and option.\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
The value argument can either be an integer or a string.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Added {get,set}sockopt(). allowbroadcast is now obsolete.
1992-06-05 15:11:30 +00:00
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
/* s.getsockopt() method.
With two arguments, retrieves an integer option.
With a third integer argument, retrieves a string buffer of that size;
use optional built-in module 'struct' to decode the string. */
Added {get,set}sockopt(). allowbroadcast is now obsolete.
1992-06-05 15:11:30 +00:00
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_getsockopt(PySocketSockObject *s, PyObject *args)
Added {get,set}sockopt(). allowbroadcast is now obsolete.
1992-06-05 15:11:30 +00:00
{
int level;
int optname;
int res;
new style
1994-10-10 17:59:00 +00:00
PyObject *buf;
Jack Jansen: The GUSI 2.0 I/O library used on the Mac uses the socklen_t (unsigned int) for most size parameters. Apparently this is part of the UNIX 98 standard. [GvR: the changes to configure.in etc. that I just checked in make sure that socklen_t is defined everywhere, so I deleted the little part of Jack's mod to define socklen_t if not in GUSI2. I suppose I will have to add it to the Windows config.h in a minute.]
2000-04-24 15:16:03 +00:00
socklen_t buflen = 0;
Added {get,set}sockopt(). allowbroadcast is now obsolete.
1992-06-05 15:11:30 +00:00
Changes for BeOS, QNX and long long, by Chris Herborth.
1998-08-04 22:53:56 +00:00
#ifdef __BEOS__
Fix a minor style consistency issue. When getting a string buffer for a string we just created, use PyString_AS_STRING() instead of PyString_AsString() to avoid the call overhead and extra type check.
2001-05-11 20:12:26 +00:00
/* We have incomplete socket support. */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error, "getsockopt not supported");
Changes for BeOS, QNX and long long, by Chris Herborth.
1998-08-04 22:53:56 +00:00
return NULL;
#else
Fix all routines to use PyArg_ParseTuple(), and add ":name" to the argument format strings. THIS WILL PROBABLY BREAK LOTS OF CODE!!! Also fixed a bogus string in an error message in getsockaddrlen().
2000-03-24 20:56:56 +00:00
if (!PyArg_ParseTuple(args, "ii|i:getsockopt",
&level, &optname, &buflen))
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
return NULL;
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
if (buflen == 0) {
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
int flag = 0;
Jack Jansen: The GUSI 2.0 I/O library used on the Mac uses the socklen_t (unsigned int) for most size parameters. Apparently this is part of the UNIX 98 standard. [GvR: the changes to configure.in etc. that I just checked in make sure that socklen_t is defined everywhere, so I deleted the little part of Jack's mod to define socklen_t if not in GUSI2. I suppose I will have to add it to the Windows config.h in a minute.]
2000-04-24 15:16:03 +00:00
socklen_t flagsize = sizeof flag;
* timemodule.c: Add hack for Solaris 2. * posixmodule.c: don't prototype getcwd() -- it's not portable... * mappingobject.c: double-check validity of last_name_char in dict{lookup,insert,remove}. * arraymodule.c: need memmove only for non-STDC Suns. * Makefile: comment out HTML_LIBS and XT_USE by default * pythonmain.c: don't prototype getopt() -- it's not standardized * socketmodule.c: cast flags arg to {get,set}sockopt() and addrbuf arg to recvfrom() to (ANY*). * pythonrun.c (initsigs): fix prototype, make it static * intobject.c (LONG_BIT): only #define it if not already defined * classobject.[ch]: remove all references to unused instance_convert() * mappingobject.c (getmappingsize): Don't return NULL in int function.
1993-11-23 17:53:17 +00:00
res = getsockopt(s->sock_fd, level, optname,
Use 'void' directly instead of the ANY #define, now that all code is ANSI C. Leave the actual #define in for API compatibility.
2000-07-25 12:56:38 +00:00
(void *)&flag, &flagsize);
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
if (res < 0)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
new style
1994-10-10 17:59:00 +00:00
return PyInt_FromLong(flag);
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
}
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
#ifdef __VMS
Patch #1516912: improve Modules support for OpenVMS.
2006-07-10 01:18:57 +00:00
/* socklen_t is unsigned so no negative test is needed,
test buflen == 0 is previously done */
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
if (buflen > 1024) {
#else
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
if (buflen <= 0 || buflen > 1024) {
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
#endif
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error,
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
"getsockopt buflen out of range");
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
return NULL;
}
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
buf = PyString_FromStringAndSize((char *)NULL, buflen);
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
if (buf == NULL)
return NULL;
* timemodule.c: Add hack for Solaris 2. * posixmodule.c: don't prototype getcwd() -- it's not portable... * mappingobject.c: double-check validity of last_name_char in dict{lookup,insert,remove}. * arraymodule.c: need memmove only for non-STDC Suns. * Makefile: comment out HTML_LIBS and XT_USE by default * pythonmain.c: don't prototype getopt() -- it's not standardized * socketmodule.c: cast flags arg to {get,set}sockopt() and addrbuf arg to recvfrom() to (ANY*). * pythonrun.c (initsigs): fix prototype, make it static * intobject.c (LONG_BIT): only #define it if not already defined * classobject.[ch]: remove all references to unused instance_convert() * mappingobject.c (getmappingsize): Don't return NULL in int function.
1993-11-23 17:53:17 +00:00
res = getsockopt(s->sock_fd, level, optname,
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
(void *)PyString_AS_STRING(buf), &buflen);
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
if (res < 0) {
new style
1994-10-10 17:59:00 +00:00
Py_DECREF(buf);
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
}
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
_PyString_Resize(&buf, buflen);
Made struct a standard built-in module; added string args to {set,get}sockopt; added gl.gversion().
1992-09-08 09:05:33 +00:00
return buf;
Changes for BeOS, QNX and long long, by Chris Herborth.
1998-08-04 22:53:56 +00:00
#endif /* __BEOS__ */
Added {get,set}sockopt(). allowbroadcast is now obsolete.
1992-06-05 15:11:30 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(getsockopt_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"getsockopt(level, option[, buffersize]) -> value\n\
\n\
Get a socket option. See the Unix manual for level and option.\n\
If a nonzero buffersize argument is given, the return value is a\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
string of that length; otherwise it is an integer.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Added {get,set}sockopt(). allowbroadcast is now obsolete.
1992-06-05 15:11:30 +00:00
Restore strict checking of socket address values; addresses passed to the bind(), connect(), and connect_ex() methods must be a single parameter. Originally part of 1.98, reverted in 1.100.
2000-07-01 03:40:12 +00:00
/* s.bind(sockaddr) method */
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_bind(PySocketSockObject *s, PyObject *addro)
Initial revision
1991-06-25 21:36:08 +00:00
{
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
sock_addr_t addrbuf;
Initial revision
1991-06-25 21:36:08 +00:00
int addrlen;
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
int res;
Convert socket methods to use METH_O and METH_NOARGS where possible.
2001-10-11 16:36:35 +00:00
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
if (!getsockaddrarg(s, addro, SAS2SA(&addrbuf), &addrlen))
Initial revision
1991-06-25 21:36:08 +00:00
return NULL;
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
res = bind(s->sock_fd, SAS2SA(&addrbuf), addrlen);
new style
1994-10-10 17:59:00 +00:00
Py_END_ALLOW_THREADS
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
if (res < 0)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
new style
1994-10-10 17:59:00 +00:00
Py_INCREF(Py_None);
return Py_None;
Initial revision
1991-06-25 21:36:08 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(bind_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"bind(address)\n\
\n\
Bind the socket to a local address. For IP sockets, the address is a\n\
SF patch 101137 from Grant Edwards Adds support for raw packets (AF_PACKET) under Linux. I haven't tested this code thoroughly; it compiles and the basic calls all work without crashing. Not sure what to actually do with raw sockets though. Not sure what other platforms this might be useful for.
2001-02-02 03:23:09 +00:00
pair (host, port); the host must refer to the local host. For raw packet\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
sockets the address is a tuple (ifname, proto [,pkttype [,hatype]])");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* s.close() method.
Set the file descriptor to -1 so operations tried subsequently
will surely fail. */
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_close(PySocketSockObject *s)
Initial revision
1991-06-25 21:36:08 +00:00
{
Adapted from a patch by Barry Scott, SF patch #102875 and SF bug #125981: closing sockets was not thread-safe.
2000-12-18 22:23:44 +00:00
SOCKET_T fd;
Convert socket methods to use METH_O and METH_NOARGS where possible.
2001-10-11 16:36:35 +00:00
Adapted from a patch by Barry Scott, SF patch #102875 and SF bug #125981: closing sockets was not thread-safe.
2000-12-18 22:23:44 +00:00
if ((fd = s->sock_fd) != -1) {
s->sock_fd = -1;
Don't close an already closed socket.
1996-10-12 14:07:22 +00:00
Py_BEGIN_ALLOW_THREADS
Adapted from a patch by Barry Scott, SF patch #102875 and SF bug #125981: closing sockets was not thread-safe.
2000-12-18 22:23:44 +00:00
(void) SOCKETCLOSE(fd);
Don't close an already closed socket.
1996-10-12 14:07:22 +00:00
Py_END_ALLOW_THREADS
}
new style
1994-10-10 17:59:00 +00:00
Py_INCREF(Py_None);
return Py_None;
Initial revision
1991-06-25 21:36:08 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(close_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"close()\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Close the socket. It cannot be used after this call.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Fix non-blocking connect() for Windows. Refactored the code that retries the connect() call in timeout mode so it can be shared between connect() and connect_ex(), and needs only a single #ifdef. The test for this was doing funky stuff I don't approve of, so I removed it in favor of a simpler test. This allowed me to implement a simpler, "purer" form of the timeout retry code. Hopefully that's enough (if you want to be fancy, use non-blocking mode and decode the errors yourself, like before).
2002-06-13 16:07:04 +00:00
static int
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
internal_connect(PySocketSockObject *s, struct sockaddr *addr, int addrlen,
int *timeoutp)
Fix non-blocking connect() for Windows. Refactored the code that retries the connect() call in timeout mode so it can be shared between connect() and connect_ex(), and needs only a single #ifdef. The test for this was doing funky stuff I don't approve of, so I removed it in favor of a simpler test. This allowed me to implement a simpler, "purer" form of the timeout retry code. Hopefully that's enough (if you want to be fancy, use non-blocking mode and decode the errors yourself, like before).
2002-06-13 16:07:04 +00:00
{
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
int res, timeout;
Fix non-blocking connect() for Windows. Refactored the code that retries the connect() call in timeout mode so it can be shared between connect() and connect_ex(), and needs only a single #ifdef. The test for this was doing funky stuff I don't approve of, so I removed it in favor of a simpler test. This allowed me to implement a simpler, "purer" form of the timeout retry code. Hopefully that's enough (if you want to be fancy, use non-blocking mode and decode the errors yourself, like before).
2002-06-13 16:07:04 +00:00
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
timeout = 0;
Fix non-blocking connect() for Windows. Refactored the code that retries the connect() call in timeout mode so it can be shared between connect() and connect_ex(), and needs only a single #ifdef. The test for this was doing funky stuff I don't approve of, so I removed it in favor of a simpler test. This allowed me to implement a simpler, "purer" form of the timeout retry code. Hopefully that's enough (if you want to be fancy, use non-blocking mode and decode the errors yourself, like before).
2002-06-13 16:07:04 +00:00
res = connect(s->sock_fd, addr, addrlen);
#ifdef MS_WINDOWS
if (s->sock_timeout > 0.0) {
Bug #876637, prevent stack corruption when socket descriptor is larger than FD_SETSIZE. This can only be acheived with ulimit -n SOME_NUMBER_BIGGER_THAN_FD_SETSIZE which is typically only available to root. Since this wouldn't normally be run in a test (ie, run as root), it doesn't seem too worthwhile to add a normal test. The bug report has one version of a test. I've written another. Not sure what the best thing to do is. Do the check before calling internal_select() because we can't set an error in between Py_BEGIN_ALLOW_THREADS and Py_END_ALLOW_THREADS. This seemed the clearest solution, ie handle before calling internal_select() rather than inside. Plus there is at least one place outside of internal_select() that needed to be handled. Will backport.
2006-02-07 07:04:46 +00:00
if (res < 0 && WSAGetLastError() == WSAEWOULDBLOCK &&
IS_SELECTABLE(s)) {
The connect timeout code wasn't working on Windows. Rather than trying to second-guess the various error returns of a second connect(), use select() to determine whether the socket becomes writable (which means connected).
2003-02-19 17:50:16 +00:00
/* This is a mess. Best solution: trust select */
fd_set fds;
Fix [ 1001018 ]: Windows: setdefaulttimeout causes unnecessary timeouts on connect error
2004-08-03 05:06:26 +00:00
fd_set fds_exc;
The connect timeout code wasn't working on Windows. Rather than trying to second-guess the various error returns of a second connect(), use select() to determine whether the socket becomes writable (which means connected).
2003-02-19 17:50:16 +00:00
struct timeval tv;
tv.tv_sec = (int)s->sock_timeout;
tv.tv_usec = (int)((s->sock_timeout - tv.tv_sec) * 1e6);
FD_ZERO(&fds);
FD_SET(s->sock_fd, &fds);
Fix [ 1001018 ]: Windows: setdefaulttimeout causes unnecessary timeouts on connect error
2004-08-03 05:06:26 +00:00
FD_ZERO(&fds_exc);
FD_SET(s->sock_fd, &fds_exc);
res = select(s->sock_fd+1, NULL, &fds, &fds_exc, &tv);
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
if (res == 0) {
The connect timeout code wasn't working on Windows. Rather than trying to second-guess the various error returns of a second connect(), use select() to determine whether the socket becomes writable (which means connected).
2003-02-19 17:50:16 +00:00
res = WSAEWOULDBLOCK;
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
timeout = 1;
Fix [ 1001018 ]: Windows: setdefaulttimeout causes unnecessary timeouts on connect error
2004-08-03 05:06:26 +00:00
} else if (res > 0) {
if (FD_ISSET(s->sock_fd, &fds))
/* The socket is in the writeable set - this
means connected */
res = 0;
else {
/* As per MS docs, we need to call getsockopt()
to get the underlying error */
int res_size = sizeof res;
/* It must be in the exception set */
assert(FD_ISSET(s->sock_fd, &fds_exc));
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
if (0 == getsockopt(s->sock_fd, SOL_SOCKET, SO_ERROR,
Fix [ 1001018 ]: Windows: setdefaulttimeout causes unnecessary timeouts on connect error
2004-08-03 05:06:26 +00:00
(char *)&res, &res_size))
/* getsockopt also clears WSAGetLastError,
so reset it back. */
WSASetLastError(res);
else
res = WSAGetLastError();
}
}
The connect timeout code wasn't working on Windows. Rather than trying to second-guess the various error returns of a second connect(), use select() to determine whether the socket becomes writable (which means connected).
2003-02-19 17:50:16 +00:00
/* else if (res < 0) an error occurred */
Fix non-blocking connect() for Windows. Refactored the code that retries the connect() call in timeout mode so it can be shared between connect() and connect_ex(), and needs only a single #ifdef. The test for this was doing funky stuff I don't approve of, so I removed it in favor of a simpler test. This allowed me to implement a simpler, "purer" form of the timeout retry code. Hopefully that's enough (if you want to be fancy, use non-blocking mode and decode the errors yourself, like before).
2002-06-13 16:07:04 +00:00
}
}
if (res < 0)
res = WSAGetLastError();
#else
if (s->sock_timeout > 0.0) {
Patch #1019808 from Federico Schwindt: Return correct socket error when a default timeout has been set, by using getsockopt() to get the error condition (instead of trying another connect() call, which seems to be a Linuxism). 2.5 bugfix candidate, assuming no one reports any problems with this change.
2008-01-19 20:47:59 +00:00
if (res < 0 && errno == EINPROGRESS && IS_SELECTABLE(s)) {
timeout = internal_select(s, 1);
if (timeout == 0) {
/* Bug #1019808: in case of an EINPROGRESS,
use getsockopt(SO_ERROR) to get the real
error. */
socklen_t res_size = sizeof res;
(void)getsockopt(s->sock_fd, SOL_SOCKET,
SO_ERROR, &res, &res_size);
if (res == EISCONN)
res = 0;
errno = res;
}
else if (timeout == -1) {
res = errno; /* had error */
}
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
else
res = EWOULDBLOCK; /* timed out */
Fix non-blocking connect() for Windows. Refactored the code that retries the connect() call in timeout mode so it can be shared between connect() and connect_ex(), and needs only a single #ifdef. The test for this was doing funky stuff I don't approve of, so I removed it in favor of a simpler test. This allowed me to implement a simpler, "purer" form of the timeout retry code. Hopefully that's enough (if you want to be fancy, use non-blocking mode and decode the errors yourself, like before).
2002-06-13 16:07:04 +00:00
}
}
if (res < 0)
res = errno;
#endif
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
*timeoutp = timeout;
Fix non-blocking connect() for Windows. Refactored the code that retries the connect() call in timeout mode so it can be shared between connect() and connect_ex(), and needs only a single #ifdef. The test for this was doing funky stuff I don't approve of, so I removed it in favor of a simpler test. This allowed me to implement a simpler, "purer" form of the timeout retry code. Hopefully that's enough (if you want to be fancy, use non-blocking mode and decode the errors yourself, like before).
2002-06-13 16:07:04 +00:00
return res;
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Restore strict checking of socket address values; addresses passed to the bind(), connect(), and connect_ex() methods must be a single parameter. Originally part of 1.98, reverted in 1.100.
2000-07-01 03:40:12 +00:00
/* s.connect(sockaddr) method */
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_connect(PySocketSockObject *s, PyObject *addro)
Initial revision
1991-06-25 21:36:08 +00:00
{
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
sock_addr_t addrbuf;
Initial revision
1991-06-25 21:36:08 +00:00
int addrlen;
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
int res;
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
int timeout;
Convert socket methods to use METH_O and METH_NOARGS where possible.
2001-10-11 16:36:35 +00:00
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
if (!getsockaddrarg(s, addro, SAS2SA(&addrbuf), &addrlen))
Initial revision
1991-06-25 21:36:08 +00:00
return NULL;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
res = internal_connect(s, SAS2SA(&addrbuf), addrlen, &timeout);
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
Py_END_ALLOW_THREADS
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
if (timeout == 1) {
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
PyErr_SetString(socket_timeout, "timed out");
return NULL;
}
Fix non-blocking connect() for Windows. Refactored the code that retries the connect() call in timeout mode so it can be shared between connect() and connect_ex(), and needs only a single #ifdef. The test for this was doing funky stuff I don't approve of, so I removed it in favor of a simpler test. This allowed me to implement a simpler, "purer" form of the timeout retry code. Hopefully that's enough (if you want to be fancy, use non-blocking mode and decode the errors yourself, like before).
2002-06-13 16:07:04 +00:00
if (res != 0)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
new style
1994-10-10 17:59:00 +00:00
Py_INCREF(Py_None);
return Py_None;
Initial revision
1991-06-25 21:36:08 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(connect_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"connect(address)\n\
\n\
Connect the socket to a remote address. For IP sockets, the address\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
is a pair (host, port).");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Restore strict checking of socket address values; addresses passed to the bind(), connect(), and connect_ex() methods must be a single parameter. Originally part of 1.98, reverted in 1.100.
2000-07-01 03:40:12 +00:00
/* s.connect_ex(sockaddr) method */
Add s.connect_ex() which returns errno instead of raising an exception.
1997-11-19 18:57:13 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_connect_ex(PySocketSockObject *s, PyObject *addro)
Add s.connect_ex() which returns errno instead of raising an exception.
1997-11-19 18:57:13 +00:00
{
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
sock_addr_t addrbuf;
Add s.connect_ex() which returns errno instead of raising an exception.
1997-11-19 18:57:13 +00:00
int addrlen;
int res;
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
int timeout;
Convert socket methods to use METH_O and METH_NOARGS where possible.
2001-10-11 16:36:35 +00:00
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
if (!getsockaddrarg(s, addro, SAS2SA(&addrbuf), &addrlen))
Add s.connect_ex() which returns errno instead of raising an exception.
1997-11-19 18:57:13 +00:00
return NULL;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Add s.connect_ex() which returns errno instead of raising an exception.
1997-11-19 18:57:13 +00:00
Py_BEGIN_ALLOW_THREADS
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
res = internal_connect(s, SAS2SA(&addrbuf), addrlen, &timeout);
Major overhaul of timeout sockets: - setblocking(0) and settimeout(0) are now equivalent, and ditto for setblocking(1) and settimeout(None). - Don't raise an exception from internal_select(); let the final call report the error (this means you will get an EAGAIN error instead of an ETIMEDOUT error -- I don't care). - Move the select to inside the Py_{BEGIN,END}_ALLOW_THREADS brackets, so other theads can run (this was a bug in the original code). - Redid the retry logic in connect() and connect_ex() to avoid masking errors. This probably doesn't work for Windows yet; I'll fix that next. It may also fail on other platforms, depending on what retrying a connect does; I need help with this. - Get rid of the retry logic in accept(). I don't think it was needed at all. But I may be wrong.
2002-06-13 15:07:44 +00:00
Py_END_ALLOW_THREADS
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
/* Signals are not errors (though they may raise exceptions). Adapted
from PyErr_SetFromErrnoWithFilenameObject(). */
#ifdef EINTR
if (res == EINTR && PyErr_CheckSignals())
return NULL;
#endif
Add s.connect_ex() which returns errno instead of raising an exception.
1997-11-19 18:57:13 +00:00
return PyInt_FromLong((long) res);
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(connect_ex_doc,
Clean up some docstrings. Some docstrings didn't show their return value; others were inconsistent in what to name the argument or return value; a few module-global functions had "socket." in front of their name, against convention.
2002-08-08 20:37:08 +00:00
"connect_ex(address) -> errno\n\
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
\n\
This is like connect(address), but returns an error code (the errno value)\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
instead of raising an exception when an error occurs.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Add s.connect_ex() which returns errno instead of raising an exception.
1997-11-19 18:57:13 +00:00
Changes for new UNIX-specific built-in module 'select' and new header for interfaces to variants of select() system call, "myselect.h". This includes adding fileno() methods to files, sockets and stdwin.
1992-06-23 09:07:03 +00:00
/* s.fileno() method */
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_fileno(PySocketSockObject *s)
Changes for new UNIX-specific built-in module 'select' and new header for interfaces to variants of select() system call, "myselect.h". This includes adding fileno() methods to files, sockets and stdwin.
1992-06-23 09:07:03 +00:00
{
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
#if SIZEOF_SOCKET_T <= SIZEOF_LONG
new style
1994-10-10 17:59:00 +00:00
return PyInt_FromLong((long) s->sock_fd);
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
#else
Rename LONG_LONG to PY_LONG_LONG. Fixes #710285.
2003-03-29 10:06:18 +00:00
return PyLong_FromLongLong((PY_LONG_LONG)s->sock_fd);
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
#endif
Changes for new UNIX-specific built-in module 'select' and new header for interfaces to variants of select() system call, "myselect.h". This includes adding fileno() methods to files, sockets and stdwin.
1992-06-23 09:07:03 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(fileno_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"fileno() -> integer\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Return the integer file descriptor of the socket.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Changes for new UNIX-specific built-in module 'select' and new header for interfaces to variants of select() system call, "myselect.h". This includes adding fileno() methods to files, sockets and stdwin.
1992-06-23 09:07:03 +00:00
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
#ifndef NO_DUP
/* s.dup() method */
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_dup(PySocketSockObject *s)
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
{
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
SOCKET_T newfd;
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
PyObject *sock;
Convert socket methods to use METH_O and METH_NOARGS where possible.
2001-10-11 16:36:35 +00:00
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
newfd = dup(s->sock_fd);
if (newfd < 0)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock = (PyObject *) new_sockobject(newfd,
s->sock_family,
s->sock_type,
s->sock_proto);
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
if (sock == NULL)
Use an explicit macro SOCKETCLOSE which expands to closesocket (on Windows), soclose (on OS2), or to close (everywhere else). Hopefully this fixes a new compilation error that I suddenly get on Windows because the macro definition for close -> closesocket apparently was done before including io.h, which contains a prototype for close. (No idea why this wasn't an error before.)
2000-04-21 20:33:00 +00:00
SOCKETCLOSE(newfd);
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
return sock;
}
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(dup_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"dup() -> socket object\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Return a new socket object connected to the same system resource.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
#endif
* config.c: Added audioop to lists. * Makefile: change default source directory * socketmodule.c: added getsockname and getpeername * bltinmodule.c: corrected typo in type() error message * Added new built-in functions str() and repr(): repr(x) == `x`; str(x) == x if x is a string, otherwise str(x) == repr(x). * Added joinfields to stropmodule.c (string.join calls joinfields now)
1992-11-26 08:54:07 +00:00
/* s.getsockname() method */
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_getsockname(PySocketSockObject *s)
* config.c: Added audioop to lists. * Makefile: change default source directory * socketmodule.c: added getsockname and getpeername * bltinmodule.c: corrected typo in type() error message * Added new built-in functions str() and repr(): repr(x) == `x`; str(x) == x if x is a string, otherwise str(x) == repr(x). * Added joinfields to stropmodule.c (string.join calls joinfields now)
1992-11-26 08:54:07 +00:00
{
- Changes donated by Elemental Security to make it work on AIX 5.3 with IBM's 64-bit compiler (SF patch #1284289). This also closes SF bug #105470: test_pwd fails on 64bit system (Opteron).
2005-09-14 18:09:42 +00:00
sock_addr_t addrbuf;
Jack Jansen: The GUSI 2.0 I/O library used on the Mac uses the socklen_t (unsigned int) for most size parameters. Apparently this is part of the UNIX 98 standard. [GvR: the changes to configure.in etc. that I just checked in make sure that socklen_t is defined everywhere, so I deleted the little part of Jack's mod to define socklen_t if not in GUSI2. I suppose I will have to add it to the Windows config.h in a minute.]
2000-04-24 15:16:03 +00:00
int res;
socklen_t addrlen;
* config.c: Added audioop to lists. * Makefile: change default source directory * socketmodule.c: added getsockname and getpeername * bltinmodule.c: corrected typo in type() error message * Added new built-in functions str() and repr(): repr(x) == `x`; str(x) == x if x is a string, otherwise str(x) == repr(x). * Added joinfields to stropmodule.c (string.join calls joinfields now)
1992-11-26 08:54:07 +00:00
if (!getsockaddrlen(s, &addrlen))
return NULL;
- Changes donated by Elemental Security to make it work on AIX 5.3 with IBM's 64-bit compiler (SF patch #1284289). This also closes SF bug #105470: test_pwd fails on 64bit system (Opteron).
2005-09-14 18:09:42 +00:00
memset(&addrbuf, 0, addrlen);
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
res = getsockname(s->sock_fd, SAS2SA(&addrbuf), &addrlen);
new style
1994-10-10 17:59:00 +00:00
Py_END_ALLOW_THREADS
* config.c: Added audioop to lists. * Makefile: change default source directory * socketmodule.c: added getsockname and getpeername * bltinmodule.c: corrected typo in type() error message * Added new built-in functions str() and repr(): repr(x) == `x`; str(x) == x if x is a string, otherwise str(x) == repr(x). * Added joinfields to stropmodule.c (string.join calls joinfields now)
1992-11-26 08:54:07 +00:00
if (res < 0)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
return makesockaddr(s->sock_fd, SAS2SA(&addrbuf), addrlen,
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
s->sock_proto);
* config.c: Added audioop to lists. * Makefile: change default source directory * socketmodule.c: added getsockname and getpeername * bltinmodule.c: corrected typo in type() error message * Added new built-in functions str() and repr(): repr(x) == `x`; str(x) == x if x is a string, otherwise str(x) == repr(x). * Added joinfields to stropmodule.c (string.join calls joinfields now)
1992-11-26 08:54:07 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(getsockname_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"getsockname() -> address info\n\
\n\
Return the address of the local endpoint. For IP sockets, the address\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
info is a pair (hostaddr, port).");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
* config.c: Added audioop to lists. * Makefile: change default source directory * socketmodule.c: added getsockname and getpeername * bltinmodule.c: corrected typo in type() error message * Added new built-in functions str() and repr(): repr(x) == `x`; str(x) == x if x is a string, otherwise str(x) == repr(x). * Added joinfields to stropmodule.c (string.join calls joinfields now)
1992-11-26 08:54:07 +00:00
Merge alpha100 branch back to main trunk
1994-08-01 11:34:53 +00:00
#ifdef HAVE_GETPEERNAME /* Cray APP doesn't have this :-( */
* config.c: Added audioop to lists. * Makefile: change default source directory * socketmodule.c: added getsockname and getpeername * bltinmodule.c: corrected typo in type() error message * Added new built-in functions str() and repr(): repr(x) == `x`; str(x) == x if x is a string, otherwise str(x) == repr(x). * Added joinfields to stropmodule.c (string.join calls joinfields now)
1992-11-26 08:54:07 +00:00
/* s.getpeername() method */
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_getpeername(PySocketSockObject *s)
* config.c: Added audioop to lists. * Makefile: change default source directory * socketmodule.c: added getsockname and getpeername * bltinmodule.c: corrected typo in type() error message * Added new built-in functions str() and repr(): repr(x) == `x`; str(x) == x if x is a string, otherwise str(x) == repr(x). * Added joinfields to stropmodule.c (string.join calls joinfields now)
1992-11-26 08:54:07 +00:00
{
- Changes donated by Elemental Security to make it work on AIX 5.3 with IBM's 64-bit compiler (SF patch #1284289). This also closes SF bug #105470: test_pwd fails on 64bit system (Opteron).
2005-09-14 18:09:42 +00:00
sock_addr_t addrbuf;
Jack Jansen: The GUSI 2.0 I/O library used on the Mac uses the socklen_t (unsigned int) for most size parameters. Apparently this is part of the UNIX 98 standard. [GvR: the changes to configure.in etc. that I just checked in make sure that socklen_t is defined everywhere, so I deleted the little part of Jack's mod to define socklen_t if not in GUSI2. I suppose I will have to add it to the Windows config.h in a minute.]
2000-04-24 15:16:03 +00:00
int res;
socklen_t addrlen;
* config.c: Added audioop to lists. * Makefile: change default source directory * socketmodule.c: added getsockname and getpeername * bltinmodule.c: corrected typo in type() error message * Added new built-in functions str() and repr(): repr(x) == `x`; str(x) == x if x is a string, otherwise str(x) == repr(x). * Added joinfields to stropmodule.c (string.join calls joinfields now)
1992-11-26 08:54:07 +00:00
if (!getsockaddrlen(s, &addrlen))
return NULL;
- Changes donated by Elemental Security to make it work on AIX 5.3 with IBM's 64-bit compiler (SF patch #1284289). This also closes SF bug #105470: test_pwd fails on 64bit system (Opteron).
2005-09-14 18:09:42 +00:00
memset(&addrbuf, 0, addrlen);
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
res = getpeername(s->sock_fd, SAS2SA(&addrbuf), &addrlen);
new style
1994-10-10 17:59:00 +00:00
Py_END_ALLOW_THREADS
* config.c: Added audioop to lists. * Makefile: change default source directory * socketmodule.c: added getsockname and getpeername * bltinmodule.c: corrected typo in type() error message * Added new built-in functions str() and repr(): repr(x) == `x`; str(x) == x if x is a string, otherwise str(x) == repr(x). * Added joinfields to stropmodule.c (string.join calls joinfields now)
1992-11-26 08:54:07 +00:00
if (res < 0)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
return makesockaddr(s->sock_fd, SAS2SA(&addrbuf), addrlen,
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
s->sock_proto);
* config.c: Added audioop to lists. * Makefile: change default source directory * socketmodule.c: added getsockname and getpeername * bltinmodule.c: corrected typo in type() error message * Added new built-in functions str() and repr(): repr(x) == `x`; str(x) == x if x is a string, otherwise str(x) == repr(x). * Added joinfields to stropmodule.c (string.join calls joinfields now)
1992-11-26 08:54:07 +00:00
}
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(getpeername_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"getpeername() -> address info\n\
\n\
Return the address of the remote endpoint. For IP sockets, the address\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
info is a pair (hostaddr, port).");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Merge alpha100 branch back to main trunk
1994-08-01 11:34:53 +00:00
#endif /* HAVE_GETPEERNAME */
* config.c: Added audioop to lists. * Makefile: change default source directory * socketmodule.c: added getsockname and getpeername * bltinmodule.c: corrected typo in type() error message * Added new built-in functions str() and repr(): repr(x) == `x`; str(x) == x if x is a string, otherwise str(x) == repr(x). * Added joinfields to stropmodule.c (string.join calls joinfields now)
1992-11-26 08:54:07 +00:00
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* s.listen(n) method */
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_listen(PySocketSockObject *s, PyObject *arg)
Initial revision
1991-06-25 21:36:08 +00:00
{
int backlog;
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
int res;
Convert socket methods to use METH_O and METH_NOARGS where possible.
2001-10-11 16:36:35 +00:00
backlog = PyInt_AsLong(arg);
if (backlog == -1 && PyErr_Occurred())
Initial revision
1991-06-25 21:36:08 +00:00
return NULL;
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
Merge alpha100 branch back to main trunk
1994-08-01 11:34:53 +00:00
if (backlog < 1)
backlog = 1;
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
res = listen(s->sock_fd, backlog);
new style
1994-10-10 17:59:00 +00:00
Py_END_ALLOW_THREADS
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
if (res < 0)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
new style
1994-10-10 17:59:00 +00:00
Py_INCREF(Py_None);
return Py_None;
Initial revision
1991-06-25 21:36:08 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(listen_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"listen(backlog)\n\
\n\
Enable a server to accept connections. The backlog argument must be at\n\
least 1; it specifies the number of unaccepted connection that the system\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
will allow before refusing new connections.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Changes to make it work and compile on NT, by Mark Hammond. (NT changes for posixmodule.c re-invented by Guido.)
1994-08-18 15:42:46 +00:00
#ifndef NO_DUP
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* s.makefile(mode) method.
Create a new open file object referring to a dupped version of
the socket's file descriptor. (The dup() call is necessary so
that the open file and socket objects may be closed independent
of each other.)
The mode argument specifies 'r' or 'w' passed to fdopen(). */
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_makefile(PySocketSockObject *s, PyObject *args)
Initial revision
1991-06-25 21:36:08 +00:00
{
Nuke all remaining occurrences of Py_PROTO and Py_FPROTO.
2000-07-09 03:09:57 +00:00
extern int fclose(FILE *);
mode and optional bufsize for makefile()
1995-03-14 15:05:13 +00:00
char *mode = "r";
int bufsize = -1;
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
#ifdef MS_WIN32
SF bug [#456252] Python should never stomp on [u]intptr_t. pyport.h: typedef a new Py_intptr_t type. DELICATE ASSUMPTION: That HAVE_UINTPTR_T implies intptr_t is available as well as uintptr_t. If that turns out not to be true, things must get uglier (C99 wants both, so I think it's an assumption we're *likely* to get away with). thread_nt.h, PyThread_start_new_thread: MS _beginthread is documented as returning unsigned long; no idea why uintptr_t was being used. Others: Always use Py_[u]intptr_t, never [u]intptr_t directly.
2001-08-29 21:37:10 +00:00
Py_intptr_t fd;
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
#else
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
int fd;
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
#endif
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
FILE *fp;
mode and optional bufsize for makefile()
1995-03-14 15:05:13 +00:00
PyObject *f;
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
#ifdef __VMS
char *mode_r = "r";
char *mode_w = "w";
#endif
mode and optional bufsize for makefile()
1995-03-14 15:05:13 +00:00
Massive patch by Skip Montanaro to add ":name" to as many PyArg_ParseTuple() format string arguments as possible.
2000-02-29 13:59:29 +00:00
if (!PyArg_ParseTuple(args, "|si:makefile", &mode, &bufsize))
Initial revision
1991-06-25 21:36:08 +00:00
return NULL;
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
#ifdef __VMS
if (strcmp(mode,"rb") == 0) {
mode = mode_r;
}
else {
if (strcmp(mode,"wb") == 0) {
mode = mode_w;
}
}
#endif
changed windows test symbols; removed debug printf from sockets
1996-06-26 18:22:49 +00:00
#ifdef MS_WIN32
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
if (((fd = _open_osfhandle(s->sock_fd, _O_BINARY)) < 0) ||
((fd = dup(fd)) < 0) || ((fp = fdopen(fd, mode)) == NULL))
add setblocking(); NT changes; null-terminate Unix path
1995-06-14 22:28:08 +00:00
#else
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
if ((fd = dup(s->sock_fd)) < 0 || (fp = fdopen(fd, mode)) == NULL)
add setblocking(); NT changes; null-terminate Unix path
1995-06-14 22:28:08 +00:00
#endif
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
{
mode and optional bufsize for makefile()
1995-03-14 15:05:13 +00:00
if (fd >= 0)
Use an explicit macro SOCKETCLOSE which expands to closesocket (on Windows), soclose (on OS2), or to close (everywhere else). Hopefully this fixes a new compilation error that I suddenly get on Windows because the macro definition for close -> closesocket apparently was done before including io.h, which contains a prototype for close. (No idea why this wasn't an error before.)
2000-04-21 20:33:00 +00:00
SOCKETCLOSE(fd);
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
mode and optional bufsize for makefile()
1995-03-14 15:05:13 +00:00
}
f = PyFile_FromFile(fp, "<socket>", mode, fclose);
if (f != NULL)
PyFile_SetBufSize(f, bufsize);
return f;
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
}
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(makefile_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"makefile([mode[, buffersize]]) -> file object\n\
\n\
Return a regular file object corresponding to the socket.\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
The mode and buffersize arguments are as for the built-in open() function.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Changes to make it work and compile on NT, by Mark Hammond. (NT changes for posixmodule.c re-invented by Guido.)
1994-08-18 15:42:46 +00:00
#endif /* NO_DUP */
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
/*
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
* This is the guts of the recv() and recv_into() methods, which reads into a
Comment typo
2006-12-19 14:29:04 +00:00
* char buffer. If you have any inc/dec ref to do to the objects that contain
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
* the buffer, do it in the caller. This function returns the number of bytes
* succesfully read. If there was an error, it returns -1. Note that it is
* also possible that we return a number of bytes smaller than the request
* bytes.
*/
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
static ssize_t
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
sock_recv_guts(PySocketSockObject *s, char* cbuf, int len, int flags)
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
{
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
ssize_t outlen = -1;
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
int timeout;
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
#ifdef __VMS
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
int remaining;
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
char *read_buf;
#endif
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
if (!IS_SELECTABLE(s)) {
select_error();
return -1;
Test for negative buffer sizes. Fixes #482871.
2001-11-19 10:41:26 +00:00
}
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
#ifndef __VMS
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
timeout = internal_select(s, 0);
if (!timeout)
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
outlen = recv(s->sock_fd, cbuf, len, flags);
new style
1994-10-10 17:59:00 +00:00
Py_END_ALLOW_THREADS
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
if (timeout == 1) {
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
PyErr_SetString(socket_timeout, "timed out");
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
return -1;
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
}
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
if (outlen < 0) {
/* Note: the call to errorhandler() ALWAYS indirectly returned
NULL, so ignore its return value */
s->errorhandler();
return -1;
plug some leaks
1995-09-13 18:39:47 +00:00
}
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
#else
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
read_buf = cbuf;
remaining = len;
while (remaining != 0) {
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
unsigned int segment;
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
int nread = -1;
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
segment = remaining /SEGMENT_SIZE;
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
if (segment != 0) {
segment = SEGMENT_SIZE;
}
else {
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
segment = remaining;
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
}
Py_BEGIN_ALLOW_THREADS
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
timeout = internal_select(s, 0);
if (!timeout)
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
nread = recv(s->sock_fd, read_buf, segment, flags);
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
Py_END_ALLOW_THREADS
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
if (timeout == 1) {
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
PyErr_SetString(socket_timeout, "timed out");
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
return -1;
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
}
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
if (nread < 0) {
s->errorhandler();
return -1;
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
}
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
if (nread != remaining) {
read_buf += nread;
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
break;
}
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
remaining -= segment;
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
read_buf += segment;
}
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
outlen = read_buf - cbuf;
Patch #614055: Support OpenVMS.
2002-12-06 12:48:53 +00:00
#endif /* !__VMS */
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
return outlen;
}
/* s.recv(nbytes [,flags]) method */
static PyObject *
sock_recv(PySocketSockObject *s, PyObject *args)
{
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
int recvlen, flags = 0;
ssize_t outlen;
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
PyObject *buf;
if (!PyArg_ParseTuple(args, "i|i:recv", &recvlen, &flags))
return NULL;
if (recvlen < 0) {
PyErr_SetString(PyExc_ValueError,
"negative buffersize in recv");
return NULL;
}
/* Allocate a new string. */
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
buf = PyString_FromStringAndSize((char *) 0, recvlen);
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
if (buf == NULL)
return NULL;
/* Call the guts */
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
outlen = sock_recv_guts(s, PyString_AS_STRING(buf), recvlen, flags);
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
if (outlen < 0) {
Fix spelling.
2006-07-28 18:36:01 +00:00
/* An error occurred, release the string and return an
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
error. */
Py_DECREF(buf);
return NULL;
}
if (outlen != recvlen) {
/* We did not read as many bytes as we anticipated, resize the
string if possible and be succesful. */
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
if (_PyString_Resize(&buf, outlen) < 0)
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
/* Oopsy, not so succesful after all. */
return NULL;
}
Initial revision
1991-06-25 21:36:08 +00:00
return buf;
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(recv_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"recv(buffersize[, flags]) -> data\n\
\n\
Receive up to buffersize bytes from the socket. For the optional flags\n\
argument, see the Unix manual. When no data is available, block until\n\
at least one byte is available or until the remote end is closed. When\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
the remote end is closed and all data is read, return the empty string.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
/* s.recv_into(buffer, [nbytes [,flags]]) method */
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
static PyObject*
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
sock_recv_into(PySocketSockObject *s, PyObject *args, PyObject *kwds)
Initial revision
1991-06-25 21:36:08 +00:00
{
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
static char *kwlist[] = {"buffer", "nbytes", "flags", 0};
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
int recvlen = 0, flags = 0;
ssize_t readlen;
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
char *buf;
int buflen;
/* Get the buffer's memory */
Fix method names. Will backport.
2007-03-31 18:54:18 +00:00
if (!PyArg_ParseTupleAndKeywords(args, kwds, "w#|ii:recv_into", kwlist,
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
&buf, &buflen, &recvlen, &flags))
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
return NULL;
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
assert(buf != 0 && buflen > 0);
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
if (recvlen < 0) {
PyErr_SetString(PyExc_ValueError,
Bug 1688393. Adds a control of negative values in socket.recvfrom, which caused an ugly crash.
2007-03-28 03:45:20 +00:00
"negative buffersize in recv_into");
* socketmodule.c: fix long-standing bug in recvfrom() -- addrlen wasn't initialized.
1993-05-25 12:16:29 +00:00
return NULL;
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
}
if (recvlen == 0) {
/* If nbytes was not specified, use the buffer's length */
recvlen = buflen;
}
/* Check if the buffer is large enough */
if (buflen < recvlen) {
PyErr_SetString(PyExc_ValueError,
"buffer too small for requested bytes");
* socketmodule.c: fix long-standing bug in recvfrom() -- addrlen wasn't initialized.
1993-05-25 12:16:29 +00:00
return NULL;
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
}
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
/* Call the guts */
readlen = sock_recv_guts(s, buf, recvlen, flags);
if (readlen < 0) {
/* Return an error. */
return NULL;
}
/* Return the number of bytes read. Note that we do not do anything
special here in the case that readlen < recvlen. */
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
return PyInt_FromSsize_t(readlen);
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
}
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
PyDoc_STRVAR(recv_into_doc,
"recv_into(buffer, [nbytes[, flags]]) -> nbytes_read\n\
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
\n\
A version of recv() that stores its data into a buffer rather than creating \n\
a new string. Receive up to buffersize bytes from the socket. If buffersize \n\
is not specified (or 0), receive up to the size available in the given buffer.\n\
\n\
See recv() for documentation about the flags.");
/*
Issue #1645 by Alberto Bertogli. Fix a comment.
2007-12-18 20:10:42 +00:00
* This is the guts of the recvfrom() and recvfrom_into() methods, which reads
* into a char buffer. If you have any inc/def ref to do to the objects that
* contain the buffer, do it in the caller. This function returns the number
* of bytes succesfully read. If there was an error, it returns -1. Note
* that it is also possible that we return a number of bytes smaller than the
* request bytes.
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
*
* 'addr' is a return value for the address object. Note that you must decref
* it yourself.
*/
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
static ssize_t
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
sock_recvfrom_guts(PySocketSockObject *s, char* cbuf, int len, int flags,
PyObject** addr)
{
sock_addr_t addrbuf;
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
int timeout;
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
ssize_t n = -1;
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
socklen_t addrlen;
*addr = NULL;
if (!getsockaddrlen(s, &addrlen))
return -1;
if (!IS_SELECTABLE(s)) {
select_error();
return -1;
}
Bug #876637, prevent stack corruption when socket descriptor is larger than FD_SETSIZE. This can only be acheived with ulimit -n SOME_NUMBER_BIGGER_THAN_FD_SETSIZE which is typically only available to root. Since this wouldn't normally be run in a test (ie, run as root), it doesn't seem too worthwhile to add a normal test. The bug report has one version of a test. I've written another. Not sure what the best thing to do is. Do the check before calling internal_select() because we can't set an error in between Py_BEGIN_ALLOW_THREADS and Py_END_ALLOW_THREADS. This seemed the clearest solution, ie handle before calling internal_select() rather than inside. Plus there is at least one place outside of internal_select() that needed to be handled. Will backport.
2006-02-07 07:04:46 +00:00
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
- Changes donated by Elemental Security to make it work on AIX 5.3 with IBM's 64-bit compiler (SF patch #1284289). This also closes SF bug #105470: test_pwd fails on 64bit system (Opteron).
2005-09-14 18:09:42 +00:00
memset(&addrbuf, 0, addrlen);
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
timeout = internal_select(s, 0);
Patch #1191065: Fix preprocessor problems on systems where recvfrom is a macro.
2006-04-15 08:35:59 +00:00
if (!timeout) {
changed windows test symbols; removed debug printf from sockets
1996-06-26 18:22:49 +00:00
#ifndef MS_WINDOWS
OS/2 EMX port changes (Modules part of patch #450267): Modules/ socketmodule.c EMX handles sockets like Posix, rather than use native APIs
2002-03-03 03:03:52 +00:00
#if defined(PYOS_OS2) && !defined(PYCC_GCC)
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
n = recvfrom(s->sock_fd, cbuf, len, flags,
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
SAS2SA(&addrbuf), &addrlen);
Left-aligned preprocessor directives and changed C++ comments to C comments.
1997-12-02 20:37:32 +00:00
#else
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
n = recvfrom(s->sock_fd, cbuf, len, flags,
Patch #1191065: Fix preprocessor problems on systems where recvfrom is a macro.
2006-04-15 08:35:59 +00:00
(void *) &addrbuf, &addrlen);
Left-aligned preprocessor directives and changed C++ comments to C comments.
1997-12-02 20:37:32 +00:00
#endif
Changes to make it work and compile on NT, by Mark Hammond. (NT changes for posixmodule.c re-invented by Guido.)
1994-08-18 15:42:46 +00:00
#else
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
n = recvfrom(s->sock_fd, cbuf, len, flags,
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
SAS2SA(&addrbuf), &addrlen);
Changes to make it work and compile on NT, by Mark Hammond. (NT changes for posixmodule.c re-invented by Guido.)
1994-08-18 15:42:46 +00:00
#endif
Patch #1191065: Fix preprocessor problems on systems where recvfrom is a macro.
2006-04-15 08:35:59 +00:00
}
new style
1994-10-10 17:59:00 +00:00
Py_END_ALLOW_THREADS
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
if (timeout == 1) {
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
PyErr_SetString(socket_timeout, "timed out");
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
return -1;
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
}
plug some leaks
1995-09-13 18:39:47 +00:00
if (n < 0) {
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
s->errorhandler();
return -1;
plug some leaks
1995-09-13 18:39:47 +00:00
}
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
if (!(*addr = makesockaddr(s->sock_fd, SAS2SA(&addrbuf),
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
addrlen, s->sock_proto)))
return -1;
return n;
}
/* s.recvfrom(nbytes [,flags]) method */
static PyObject *
sock_recvfrom(PySocketSockObject *s, PyObject *args)
{
PyObject *buf = NULL;
PyObject *addr = NULL;
PyObject *ret = NULL;
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
int recvlen, flags = 0;
ssize_t outlen;
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
if (!PyArg_ParseTuple(args, "i|i:recvfrom", &recvlen, &flags))
return NULL;
Bug 1688393. Adds a control of negative values in socket.recvfrom, which caused an ugly crash.
2007-03-28 03:45:20 +00:00
if (recvlen < 0) {
PyErr_SetString(PyExc_ValueError,
"negative buffersize in recvfrom");
return NULL;
}
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
buf = PyString_FromStringAndSize((char *) 0, recvlen);
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
if (buf == NULL)
Initial revision
1991-06-25 21:36:08 +00:00
return NULL;
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
outlen = sock_recvfrom_guts(s, PyString_AS_STRING(buf),
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
recvlen, flags, &addr);
if (outlen < 0) {
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
goto finally;
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
}
if (outlen != recvlen) {
/* We did not read as many bytes as we anticipated, resize the
string if possible and be succesful. */
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
if (_PyString_Resize(&buf, outlen) < 0)
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
/* Oopsy, not so succesful after all. */
goto finally;
}
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
Simplify and speedup uses of Py_BuildValue(): * Py_BuildValue("(OOO)",a,b,c) --> PyTuple_Pack(3,a,b,c) * Py_BuildValue("()",a) --> PyTuple_New(0) * Py_BuildValue("O", a) --> Py_INCREF(a)
2003-10-12 19:09:37 +00:00
ret = PyTuple_Pack(2, buf, addr);
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
finally:
new style
1994-10-10 17:59:00 +00:00
Py_XDECREF(buf);
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
Py_XDECREF(addr);
* ceval.c: ifdef out the last argument passing compat hack. * Fixed memory leaks in socket, select and sv modules: mkvalue("O", v) does INCREF(v) so if v is brand new it should be XDECREF'd
1993-02-05 09:46:15 +00:00
return ret;
Initial revision
1991-06-25 21:36:08 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(recvfrom_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"recvfrom(buffersize[, flags]) -> (data, address info)\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Like recv(buffersize, flags) but also return the sender's address info.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
/* s.recvfrom_into(buffer[, nbytes [,flags]]) method */
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
static PyObject *
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
sock_recvfrom_into(PySocketSockObject *s, PyObject *args, PyObject* kwds)
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
{
static char *kwlist[] = {"buffer", "nbytes", "flags", 0};
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
int recvlen = 0, flags = 0;
ssize_t readlen;
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
char *buf;
int buflen;
PyObject *addr = NULL;
Fix method names. Will backport.
2007-03-31 18:54:18 +00:00
if (!PyArg_ParseTupleAndKeywords(args, kwds, "w#|ii:recvfrom_into",
kwlist, &buf, &buflen,
&recvlen, &flags))
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
return NULL;
assert(buf != 0 && buflen > 0);
if (recvlen < 0) {
PyErr_SetString(PyExc_ValueError,
Bug 1688393. Adds a control of negative values in socket.recvfrom, which caused an ugly crash.
2007-03-28 03:45:20 +00:00
"negative buffersize in recvfrom_into");
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
return NULL;
}
if (recvlen == 0) {
/* If nbytes was not specified, use the buffer's length */
recvlen = buflen;
}
readlen = sock_recvfrom_guts(s, buf, recvlen, flags, &addr);
if (readlen < 0) {
/* Return an error */
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
Py_XDECREF(addr);
return NULL;
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
}
/* Return the number of bytes read and the address. Note that we do
not do anything special here in the case that readlen < recvlen. */
Fix a refleak in recvfrom_into
2006-06-08 05:12:45 +00:00
return Py_BuildValue("lN", readlen, addr);
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
}
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
PyDoc_STRVAR(recvfrom_into_doc,
"recvfrom_into(buffer[, nbytes[, flags]]) -> (nbytes, address info)\n\
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
\n\
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
Like recv_into(buffer[, nbytes[, flags]]) but also return the sender's address info.");
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
* classobject.c: in instance_getattr, don't make a method out of a function found as instance data. * socketmodule.c: added 'flags' argument sendto/recvfrom, rewrite argument parsing in send/recv. * More changes related to access (terminology change: owner instead of class; allow any object as owner; local/global variables are owned by their dictionary, only class/instance data is owned by the class; "from...import *" now only imports objects with public access; etc.)
1993-05-25 09:38:27 +00:00
/* s.send(data [,flags]) method */
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_send(PySocketSockObject *s, PyObject *args)
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
{
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
char *buf;
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
int len, n = -1, flags = 0, timeout;
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
Py_buffer pbuf;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
if (!PyArg_ParseTuple(args, "s*|i:send", &pbuf, &flags))
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
return NULL;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
if (!IS_SELECTABLE(s)) {
PyBuffer_Release(&pbuf);
Bug #876637, prevent stack corruption when socket descriptor is larger than FD_SETSIZE. This can only be acheived with ulimit -n SOME_NUMBER_BIGGER_THAN_FD_SETSIZE which is typically only available to root. Since this wouldn't normally be run in a test (ie, run as root), it doesn't seem too worthwhile to add a normal test. The bug report has one version of a test. I've written another. Not sure what the best thing to do is. Do the check before calling internal_select() because we can't set an error in between Py_BEGIN_ALLOW_THREADS and Py_END_ALLOW_THREADS. This seemed the clearest solution, ie handle before calling internal_select() rather than inside. Plus there is at least one place outside of internal_select() that needed to be handled. Will backport.
2006-02-07 07:04:46 +00:00
return select_error();
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
}
buf = pbuf.buf;
len = pbuf.len;
Bug #876637, prevent stack corruption when socket descriptor is larger than FD_SETSIZE. This can only be acheived with ulimit -n SOME_NUMBER_BIGGER_THAN_FD_SETSIZE which is typically only available to root. Since this wouldn't normally be run in a test (ie, run as root), it doesn't seem too worthwhile to add a normal test. The bug report has one version of a test. I've written another. Not sure what the best thing to do is. Do the check before calling internal_select() because we can't set an error in between Py_BEGIN_ALLOW_THREADS and Py_END_ALLOW_THREADS. This seemed the clearest solution, ie handle before calling internal_select() rather than inside. Plus there is at least one place outside of internal_select() that needed to be handled. Will backport.
2006-02-07 07:04:46 +00:00
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
timeout = internal_select(s, 1);
if (!timeout)
Patch #1516912: improve Modules support for OpenVMS.
2006-07-10 01:18:57 +00:00
#ifdef __VMS
n = sendsegmented(s->sock_fd, buf, len, flags);
#else
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
n = send(s->sock_fd, buf, len, flags);
Patch #1516912: improve Modules support for OpenVMS.
2006-07-10 01:18:57 +00:00
#endif
new style
1994-10-10 17:59:00 +00:00
Py_END_ALLOW_THREADS
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
PyBuffer_Release(&pbuf);
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
if (timeout == 1) {
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
PyErr_SetString(socket_timeout, "timed out");
return NULL;
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
if (n < 0)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
new style
1994-10-10 17:59:00 +00:00
return PyInt_FromLong((long)n);
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(send_doc,
Add sendall() method, which loops until all data is written or an error occurs, and doesn't return a count. (This is my second patch from SF patch #474307, with small change to the docstring for send().) 2.1.2 "bugfix" candidate.
2001-10-26 03:25:00 +00:00
"send(data[, flags]) -> count\n\
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
\n\
Send a data string to the socket. For the optional flags\n\
Add sendall() method, which loops until all data is written or an error occurs, and doesn't return a count. (This is my second patch from SF patch #474307, with small change to the docstring for send().) 2.1.2 "bugfix" candidate.
2001-10-26 03:25:00 +00:00
argument, see the Unix manual. Return the number of bytes\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
sent; this may be less than len(data) if the network is busy.");
Add sendall() method, which loops until all data is written or an error occurs, and doesn't return a count. (This is my second patch from SF patch #474307, with small change to the docstring for send().) 2.1.2 "bugfix" candidate.
2001-10-26 03:25:00 +00:00
/* s.sendall(data [,flags]) method */
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_sendall(PySocketSockObject *s, PyObject *args)
Add sendall() method, which loops until all data is written or an error occurs, and doesn't return a count. (This is my second patch from SF patch #474307, with small change to the docstring for send().) 2.1.2 "bugfix" candidate.
2001-10-26 03:25:00 +00:00
{
char *buf;
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
int len, n = -1, flags = 0, timeout;
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
Py_buffer pbuf;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
if (!PyArg_ParseTuple(args, "s*|i:sendall", &pbuf, &flags))
Add sendall() method, which loops until all data is written or an error occurs, and doesn't return a count. (This is my second patch from SF patch #474307, with small change to the docstring for send().) 2.1.2 "bugfix" candidate.
2001-10-26 03:25:00 +00:00
return NULL;
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
buf = pbuf.buf;
len = pbuf.len;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
if (!IS_SELECTABLE(s)) {
PyBuffer_Release(&pbuf);
Bug #876637, prevent stack corruption when socket descriptor is larger than FD_SETSIZE. This can only be acheived with ulimit -n SOME_NUMBER_BIGGER_THAN_FD_SETSIZE which is typically only available to root. Since this wouldn't normally be run in a test (ie, run as root), it doesn't seem too worthwhile to add a normal test. The bug report has one version of a test. I've written another. Not sure what the best thing to do is. Do the check before calling internal_select() because we can't set an error in between Py_BEGIN_ALLOW_THREADS and Py_END_ALLOW_THREADS. This seemed the clearest solution, ie handle before calling internal_select() rather than inside. Plus there is at least one place outside of internal_select() that needed to be handled. Will backport.
2006-02-07 07:04:46 +00:00
return select_error();
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
}
Bug #876637, prevent stack corruption when socket descriptor is larger than FD_SETSIZE. This can only be acheived with ulimit -n SOME_NUMBER_BIGGER_THAN_FD_SETSIZE which is typically only available to root. Since this wouldn't normally be run in a test (ie, run as root), it doesn't seem too worthwhile to add a normal test. The bug report has one version of a test. I've written another. Not sure what the best thing to do is. Do the check before calling internal_select() because we can't set an error in between Py_BEGIN_ALLOW_THREADS and Py_END_ALLOW_THREADS. This seemed the clearest solution, ie handle before calling internal_select() rather than inside. Plus there is at least one place outside of internal_select() that needed to be handled. Will backport.
2006-02-07 07:04:46 +00:00
Add sendall() method, which loops until all data is written or an error occurs, and doesn't return a count. (This is my second patch from SF patch #474307, with small change to the docstring for send().) 2.1.2 "bugfix" candidate.
2001-10-26 03:25:00 +00:00
Py_BEGIN_ALLOW_THREADS
do {
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
timeout = internal_select(s, 1);
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
n = -1;
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
if (timeout)
break;
Patch #1516912: improve Modules support for OpenVMS.
2006-07-10 01:18:57 +00:00
#ifdef __VMS
n = sendsegmented(s->sock_fd, buf, len, flags);
#else
Add sendall() method, which loops until all data is written or an error occurs, and doesn't return a count. (This is my second patch from SF patch #474307, with small change to the docstring for send().) 2.1.2 "bugfix" candidate.
2001-10-26 03:25:00 +00:00
n = send(s->sock_fd, buf, len, flags);
Patch #1516912: improve Modules support for OpenVMS.
2006-07-10 01:18:57 +00:00
#endif
Fix issue1628205: Socket file objects returned by socket.socket.makefile() now properly handles EINTR within the read, readline, write & flush methods. The socket.sendall() method now properly handles interrupted system calls.
2009-08-13 18:54:50 +00:00
if (n < 0) {
#ifdef EINTR
/* We must handle EINTR here as there is no way for
* the caller to know how much was sent otherwise. */
if (errno == EINTR) {
/* Run signal handlers. If an exception was
* raised, abort and leave this socket in
* an unknown state. */
if (PyErr_CheckSignals())
return NULL;
continue;
}
#endif
Add sendall() method, which loops until all data is written or an error occurs, and doesn't return a count. (This is my second patch from SF patch #474307, with small change to the docstring for send().) 2.1.2 "bugfix" candidate.
2001-10-26 03:25:00 +00:00
break;
Fix issue1628205: Socket file objects returned by socket.socket.makefile() now properly handles EINTR within the read, readline, write & flush methods. The socket.sendall() method now properly handles interrupted system calls.
2009-08-13 18:54:50 +00:00
}
Add sendall() method, which loops until all data is written or an error occurs, and doesn't return a count. (This is my second patch from SF patch #474307, with small change to the docstring for send().) 2.1.2 "bugfix" candidate.
2001-10-26 03:25:00 +00:00
buf += n;
len -= n;
} while (len > 0);
Py_END_ALLOW_THREADS
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
PyBuffer_Release(&pbuf);
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
if (timeout == 1) {
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
PyErr_SetString(socket_timeout, "timed out");
return NULL;
}
Add sendall() method, which loops until all data is written or an error occurs, and doesn't return a count. (This is my second patch from SF patch #474307, with small change to the docstring for send().) 2.1.2 "bugfix" candidate.
2001-10-26 03:25:00 +00:00
if (n < 0)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Add sendall() method, which loops until all data is written or an error occurs, and doesn't return a count. (This is my second patch from SF patch #474307, with small change to the docstring for send().) 2.1.2 "bugfix" candidate.
2001-10-26 03:25:00 +00:00
Py_INCREF(Py_None);
return Py_None;
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(sendall_doc,
Add sendall() method, which loops until all data is written or an error occurs, and doesn't return a count. (This is my second patch from SF patch #474307, with small change to the docstring for send().) 2.1.2 "bugfix" candidate.
2001-10-26 03:25:00 +00:00
"sendall(data[, flags])\n\
\n\
Send a data string to the socket. For the optional flags\n\
argument, see the Unix manual. This calls send() repeatedly\n\
until all data is sent. If an error occurs, it's impossible\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
to tell how much data has been sent.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
* classobject.c: in instance_getattr, don't make a method out of a function found as instance data. * socketmodule.c: added 'flags' argument sendto/recvfrom, rewrite argument parsing in send/recv. * More changes related to access (terminology change: owner instead of class; allow any object as owner; local/global variables are owned by their dictionary, only class/instance data is owned by the class; "from...import *" now only imports objects with public access; etc.)
1993-05-25 09:38:27 +00:00
/* s.sendto(data, [flags,] sockaddr) method */
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_sendto(PySocketSockObject *s, PyObject *args)
Initial revision
1991-06-25 21:36:08 +00:00
{
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
Py_buffer pbuf;
new style
1994-10-10 17:59:00 +00:00
PyObject *addro;
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
char *buf;
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
Py_ssize_t len;
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
sock_addr_t addrbuf;
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
int addrlen, n = -1, flags, timeout;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
* classobject.c: in instance_getattr, don't make a method out of a function found as instance data. * socketmodule.c: added 'flags' argument sendto/recvfrom, rewrite argument parsing in send/recv. * More changes related to access (terminology change: owner instead of class; allow any object as owner; local/global variables are owned by their dictionary, only class/instance data is owned by the class; "from...import *" now only imports objects with public access; etc.)
1993-05-25 09:38:27 +00:00
flags = 0;
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
if (!PyArg_ParseTuple(args, "s*O:sendto", &pbuf, &addro)) {
new style
1994-10-10 17:59:00 +00:00
PyErr_Clear();
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
if (!PyArg_ParseTuple(args, "s*iO:sendto",
&pbuf, &flags, &addro))
* classobject.c: in instance_getattr, don't make a method out of a function found as instance data. * socketmodule.c: added 'flags' argument sendto/recvfrom, rewrite argument parsing in send/recv. * More changes related to access (terminology change: owner instead of class; allow any object as owner; local/global variables are owned by their dictionary, only class/instance data is owned by the class; "from...import *" now only imports objects with public access; etc.)
1993-05-25 09:38:27 +00:00
return NULL;
Initial revision
1991-06-25 21:36:08 +00:00
}
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
buf = pbuf.buf;
len = pbuf.len;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
if (!IS_SELECTABLE(s)) {
PyBuffer_Release(&pbuf);
Bug #876637, prevent stack corruption when socket descriptor is larger than FD_SETSIZE. This can only be acheived with ulimit -n SOME_NUMBER_BIGGER_THAN_FD_SETSIZE which is typically only available to root. Since this wouldn't normally be run in a test (ie, run as root), it doesn't seem too worthwhile to add a normal test. The bug report has one version of a test. I've written another. Not sure what the best thing to do is. Do the check before calling internal_select() because we can't set an error in between Py_BEGIN_ALLOW_THREADS and Py_END_ALLOW_THREADS. This seemed the clearest solution, ie handle before calling internal_select() rather than inside. Plus there is at least one place outside of internal_select() that needed to be handled. Will backport.
2006-02-07 07:04:46 +00:00
return select_error();
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
}
Bug #876637, prevent stack corruption when socket descriptor is larger than FD_SETSIZE. This can only be acheived with ulimit -n SOME_NUMBER_BIGGER_THAN_FD_SETSIZE which is typically only available to root. Since this wouldn't normally be run in a test (ie, run as root), it doesn't seem too worthwhile to add a normal test. The bug report has one version of a test. I've written another. Not sure what the best thing to do is. Do the check before calling internal_select() because we can't set an error in between Py_BEGIN_ALLOW_THREADS and Py_END_ALLOW_THREADS. This seemed the clearest solution, ie handle before calling internal_select() rather than inside. Plus there is at least one place outside of internal_select() that needed to be handled. Will backport.
2006-02-07 07:04:46 +00:00
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
if (!getsockaddrarg(s, addro, SAS2SA(&addrbuf), &addrlen)) {
PyBuffer_Release(&pbuf);
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
return NULL;
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
}
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
timeout = internal_select(s, 1);
if (!timeout)
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
n = sendto(s->sock_fd, buf, len, flags, SAS2SA(&addrbuf), addrlen);
new style
1994-10-10 17:59:00 +00:00
Py_END_ALLOW_THREADS
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Issue #3139: Make buffer-interface thread-safe wrt. PyArg_ParseTuple, by denying s# to parse objects that have a releasebuffer procedure, and introducing s*. More module might need to get converted to use s*.
2008-08-12 14:49:50 +00:00
PyBuffer_Release(&pbuf);
Patch #1519025 and bug #926423: If a KeyboardInterrupt occurs during a socket operation on a socket with a timeout, the exception will be caught correctly. Previously, the exception was not caught.
2006-08-02 06:46:21 +00:00
if (timeout == 1) {
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
PyErr_SetString(socket_timeout, "timed out");
return NULL;
}
Initial revision
1991-06-25 21:36:08 +00:00
if (n < 0)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
new style
1994-10-10 17:59:00 +00:00
return PyInt_FromLong((long)n);
Initial revision
1991-06-25 21:36:08 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(sendto_doc,
Clean up some docstrings. Some docstrings didn't show their return value; others were inconsistent in what to name the argument or return value; a few module-global functions had "socket." in front of their name, against convention.
2002-08-08 20:37:08 +00:00
"sendto(data[, flags], address) -> count\n\
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
\n\
Like send(data, flags) but allows specifying the destination address.\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
For IP sockets, the address is a pair (hostaddr, port).");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* s.shutdown(how) method */
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_shutdown(PySocketSockObject *s, PyObject *arg)
Initial revision
1991-06-25 21:36:08 +00:00
{
int how;
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
int res;
Convert socket methods to use METH_O and METH_NOARGS where possible.
2001-10-11 16:36:35 +00:00
how = PyInt_AsLong(arg);
if (how == -1 && PyErr_Occurred())
Initial revision
1991-06-25 21:36:08 +00:00
return NULL;
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
res = shutdown(s->sock_fd, how);
new style
1994-10-10 17:59:00 +00:00
Py_END_ALLOW_THREADS
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
if (res < 0)
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
return s->errorhandler();
new style
1994-10-10 17:59:00 +00:00
Py_INCREF(Py_None);
return Py_None;
Initial revision
1991-06-25 21:36:08 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(shutdown_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"shutdown(flag)\n\
\n\
Patch #849595: Add socket.shutdown() constants.
2003-11-27 19:40:22 +00:00
Shut down the reading side of the socket (flag == SHUT_RD), the writing side\n\
of the socket (flag == SHUT_WR), or both ends (flag == SHUT_RDWR).");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Restore support for Microsoft VC6 compiler. Some functions in the msvcrt module are skipped, and socket.ioctl is enabled only when using a more recent Platform SDK. (and yes, there are still companies that use a 10-years old compiler)
2008-06-13 00:42:22 +00:00
#if defined(MS_WINDOWS) && defined(SIO_RCVALL)
Added interface to Windows' WSAIoctl and a simple example for a network sniffer.
2008-01-04 15:23:30 +00:00
static PyObject*
sock_ioctl(PySocketSockObject *s, PyObject *arg)
{
unsigned long cmd = SIO_RCVALL;
unsigned int option = RCVALL_ON;
DWORD recv;
if (!PyArg_ParseTuple(arg, "kI:ioctl", &cmd, &option))
return NULL;
if (WSAIoctl(s->sock_fd, cmd, &option, sizeof(option),
NULL, 0, &recv, NULL, NULL) == SOCKET_ERROR) {
return set_error();
}
return PyLong_FromUnsignedLong(recv);
}
PyDoc_STRVAR(sock_ioctl_doc,
"ioctl(cmd, option) -> long\n\
\n\
Control the socket with WSAIoctl syscall. Currently only socket.SIO_RCVALL\n\
is supported as control. Options must be one of the socket.RCVALL_*\n\
constants.");
#endif
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* List of methods for socket objects */
Initial revision
1991-06-25 21:36:08 +00:00
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
static PyMethodDef sock_methods[] = {
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
{"accept", (PyCFunction)sock_accept, METH_NOARGS,
accept_doc},
{"bind", (PyCFunction)sock_bind, METH_O,
bind_doc},
{"close", (PyCFunction)sock_close, METH_NOARGS,
close_doc},
{"connect", (PyCFunction)sock_connect, METH_O,
connect_doc},
{"connect_ex", (PyCFunction)sock_connect_ex, METH_O,
connect_ex_doc},
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
#ifndef NO_DUP
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
{"dup", (PyCFunction)sock_dup, METH_NOARGS,
dup_doc},
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
#endif
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
{"fileno", (PyCFunction)sock_fileno, METH_NOARGS,
fileno_doc},
Merge alpha100 branch back to main trunk
1994-08-01 11:34:53 +00:00
#ifdef HAVE_GETPEERNAME
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
{"getpeername", (PyCFunction)sock_getpeername,
METH_NOARGS, getpeername_doc},
#endif
{"getsockname", (PyCFunction)sock_getsockname,
METH_NOARGS, getsockname_doc},
{"getsockopt", (PyCFunction)sock_getsockopt, METH_VARARGS,
getsockopt_doc},
Restore support for Microsoft VC6 compiler. Some functions in the msvcrt module are skipped, and socket.ioctl is enabled only when using a more recent Platform SDK. (and yes, there are still companies that use a 10-years old compiler)
2008-06-13 00:42:22 +00:00
#if defined(MS_WINDOWS) && defined(SIO_RCVALL)
Added interface to Windows' WSAIoctl and a simple example for a network sniffer.
2008-01-04 15:23:30 +00:00
{"ioctl", (PyCFunction)sock_ioctl, METH_VARARGS,
sock_ioctl_doc},
#endif
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
{"listen", (PyCFunction)sock_listen, METH_O,
listen_doc},
Changes to make it work and compile on NT, by Mark Hammond. (NT changes for posixmodule.c re-invented by Guido.)
1994-08-18 15:42:46 +00:00
#ifndef NO_DUP
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
{"makefile", (PyCFunction)sock_makefile, METH_VARARGS,
makefile_doc},
#endif
{"recv", (PyCFunction)sock_recv, METH_VARARGS,
recv_doc},
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
{"recv_into", (PyCFunction)sock_recv_into, METH_VARARGS | METH_KEYWORDS,
recv_into_doc},
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
{"recvfrom", (PyCFunction)sock_recvfrom, METH_VARARGS,
recvfrom_doc},
Fixes in struct and socket from merge reviews. - Following Guido's comments, renamed * pack_to -> pack_into * recv_buf -> recv_into * recvfrom_buf -> recvfrom_into - Made fixes to _struct.c according to Neal Norwitz comments on the checkins list. - Converted some ints into the appropriate -- I hope -- ssize_t and size_t.
2006-06-04 13:49:49 +00:00
{"recvfrom_into", (PyCFunction)sock_recvfrom_into, METH_VARARGS | METH_KEYWORDS,
recvfrom_into_doc},
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
{"send", (PyCFunction)sock_send, METH_VARARGS,
send_doc},
{"sendall", (PyCFunction)sock_sendall, METH_VARARGS,
sendall_doc},
{"sendto", (PyCFunction)sock_sendto, METH_VARARGS,
sendto_doc},
{"setblocking", (PyCFunction)sock_setblocking, METH_O,
setblocking_doc},
{"settimeout", (PyCFunction)sock_settimeout, METH_O,
settimeout_doc},
{"gettimeout", (PyCFunction)sock_gettimeout, METH_NOARGS,
gettimeout_doc},
{"setsockopt", (PyCFunction)sock_setsockopt, METH_VARARGS,
setsockopt_doc},
{"shutdown", (PyCFunction)sock_shutdown, METH_O,
shutdown_doc},
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
#ifdef RISCOS
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
{"sleeptaskw", (PyCFunction)sock_sleeptaskw, METH_O,
sleeptaskw_doc},
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
#endif
Merge alpha100 branch back to main trunk
1994-08-01 11:34:53 +00:00
{NULL, NULL} /* sentinel */
Initial revision
1991-06-25 21:36:08 +00:00
};
RFE #567972: Socket objects' family, type and proto properties are now exposed via new get...() methods.
2006-03-17 19:17:34 +00:00
/* SockObject members */
static PyMemberDef sock_memberlist[] = {
{"family", T_INT, offsetof(PySocketSockObject, sock_family), READONLY, "the socket family"},
{"type", T_INT, offsetof(PySocketSockObject, sock_type), READONLY, "the socket type"},
{"proto", T_INT, offsetof(PySocketSockObject, sock_proto), READONLY, "the socket protocol"},
{"timeout", T_DOUBLE, offsetof(PySocketSockObject, sock_timeout), READONLY, "the socket timeout"},
{0},
};
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
new style
1994-10-10 17:59:00 +00:00
/* Deallocate a socket object in response to the last Py_DECREF().
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
First close the file description. */
Initial revision
1991-06-25 21:36:08 +00:00
static void
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_dealloc(PySocketSockObject *s)
Initial revision
1991-06-25 21:36:08 +00:00
{
In dealloc(), only close the socket if fd != -1.
2000-04-10 12:45:45 +00:00
if (s->sock_fd != -1)
Use an explicit macro SOCKETCLOSE which expands to closesocket (on Windows), soclose (on OS2), or to close (everywhere else). Hopefully this fixes a new compilation error that I suddenly get on Windows because the macro definition for close -> closesocket apparently was done before including io.h, which contains a prototype for close. (No idea why this wasn't an error before.)
2000-04-21 20:33:00 +00:00
(void) SOCKETCLOSE(s->sock_fd);
#1629: Renamed Py_Size, Py_Type and Py_Refcnt to Py_SIZE, Py_TYPE and Py_REFCNT. Macros for b/w compatibility are available.
2007-12-19 02:37:44 +00:00
Py_TYPE(s)->tp_free((PyObject *)s);
Initial revision
1991-06-25 21:36:08 +00:00
}
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_repr(PySocketSockObject *s)
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
{
char buf[512];
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
#if SIZEOF_SOCKET_T > SIZEOF_LONG
if (s->sock_fd > LONG_MAX) {
/* this can occur on Win64, and actually there is a special
ugly printf formatter for decimal pointer length integer
printing, only bother if necessary*/
PyErr_SetString(PyExc_OverflowError,
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
"no printf formatter to display "
"the socket descriptor in decimal");
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
return NULL;
}
#endif
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
PyOS_snprintf(
buf, sizeof(buf),
"<socket object, fd=%ld, family=%d, type=%d, protocol=%d>",
(long)s->sock_fd, s->sock_family,
s->sock_type,
s->sock_proto);
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromString(buf);
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
}
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
/* Create a new, uninitialized socket object. */
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
{
PyObject *new;
new = type->tp_alloc(type, 0);
The tp_new implementation should initialize the errorhandler field, otherwise this code could segfault: from socket import socket s = socket.__new__(socket) s.recv(100)
2002-06-06 20:08:25 +00:00
if (new != NULL) {
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
((PySocketSockObject *)new)->sock_fd = -1;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
((PySocketSockObject *)new)->sock_timeout = -1.0;
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
((PySocketSockObject *)new)->errorhandler = &set_error;
The tp_new implementation should initialize the errorhandler field, otherwise this code could segfault: from socket import socket s = socket.__new__(socket) s.recv(100)
2002-06-06 20:08:25 +00:00
}
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
return new;
}
/* Initialize a new socket object. */
/*ARGSUSED*/
static int
work around name clash with OS/2 TCPIP routine sock_init()
2002-06-13 11:53:52 +00:00
sock_initobj(PyObject *self, PyObject *args, PyObject *kwds)
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
{
PySocketSockObject *s = (PySocketSockObject *)self;
SOCKET_T fd;
int family = AF_INET, type = SOCK_STREAM, proto = 0;
More unconsting.
2006-02-27 17:20:04 +00:00
static char *keywords[] = {"family", "type", "proto", 0};
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
if (!PyArg_ParseTupleAndKeywords(args, kwds,
"|iii:socket", keywords,
&family, &type, &proto))
return -1;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
Py_BEGIN_ALLOW_THREADS
fd = socket(family, type, proto);
Py_END_ALLOW_THREADS
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
#ifdef MS_WINDOWS
if (fd == INVALID_SOCKET)
#else
if (fd < 0)
#endif
{
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
set_error();
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
return -1;
}
init_sockobject(s, fd, family, type, proto);
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
return 0;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
}
Merge alpha100 branch back to main trunk
1994-08-01 11:34:53 +00:00
/* Type object for socket objects. */
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
static PyTypeObject sock_type = {
PEP 3123: Provide forward compatibility with Python 3.0, while keeping backwards compatibility. Add Py_Refcnt, Py_Type, Py_Size, and PyVarObject_HEAD_INIT.
2007-07-21 06:55:02 +00:00
PyVarObject_HEAD_INIT(0, 0) /* Must fill in type value later */
Oops. In the tp_name field, the name should be "_socket.socket", not "socket.socket" -- on Windows, "socket.socket" is the wrapper class. Also added the module name to the SSL type (which is not a new-style class -- I don't want to mess with it yet).
2001-10-28 12:31:33 +00:00
"_socket.socket", /* tp_name */
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
sizeof(PySocketSockObject), /* tp_basicsize */
0, /* tp_itemsize */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
(destructor)sock_dealloc, /* tp_dealloc */
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
0, /* tp_print */
0, /* tp_getattr */
0, /* tp_setattr */
0, /* tp_compare */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
(reprfunc)sock_repr, /* tp_repr */
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
0, /* tp_as_number */
0, /* tp_as_sequence */
0, /* tp_as_mapping */
0, /* tp_hash */
0, /* tp_call */
0, /* tp_str */
Patch #661760: Cygwin auto-import module patch The attached patch enables shared extension modules to build cleanly under Cygwin without moving the static initialization of certain function pointers (i.e., ones exported from the Python DLL core) to a module initialization function. Additionally, this patch fixes the modules that have been changed in the past to accommodate Cygwin.
2003-01-06 12:41:26 +00:00
PyObject_GenericGetAttr, /* tp_getattro */
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
0, /* tp_setattro */
0, /* tp_as_buffer */
Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE, /* tp_flags */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_doc, /* tp_doc */
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
0, /* tp_traverse */
0, /* tp_clear */
0, /* tp_richcompare */
0, /* tp_weaklistoffset */
0, /* tp_iter */
0, /* tp_iternext */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_methods, /* tp_methods */
RFE #567972: Socket objects' family, type and proto properties are now exposed via new get...() methods.
2006-03-17 19:17:34 +00:00
sock_memberlist, /* tp_members */
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
0, /* tp_getset */
0, /* tp_base */
0, /* tp_dict */
0, /* tp_descr_get */
0, /* tp_descr_set */
0, /* tp_dictoffset */
work around name clash with OS/2 TCPIP routine sock_init()
2002-06-13 11:53:52 +00:00
sock_initobj, /* tp_init */
Patch #661760: Cygwin auto-import module patch The attached patch enables shared extension modules to build cleanly under Cygwin without moving the static initialization of certain function pointers (i.e., ones exported from the Python DLL core) to a module initialization function. Additionally, this patch fixes the modules that have been changed in the past to accommodate Cygwin.
2003-01-06 12:41:26 +00:00
PyType_GenericAlloc, /* tp_alloc */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
sock_new, /* tp_new */
Patch #661760: Cygwin auto-import module patch The attached patch enables shared extension modules to build cleanly under Cygwin without moving the static initialization of certain function pointers (i.e., ones exported from the Python DLL core) to a module initialization function. Additionally, this patch fixes the modules that have been changed in the past to accommodate Cygwin.
2003-01-06 12:41:26 +00:00
PyObject_Del, /* tp_free */
Initial revision
1991-06-25 21:36:08 +00:00
};
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Add gethostname() and s.avail().
1991-07-27 21:42:02 +00:00
/* Python interface to gethostname(). */
/*ARGSUSED*/
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Make use of METH_O and METH_NOARGS where possible. Use Py_UnpackTuple instead of PyArg_ParseTuple where possible.
2006-05-29 21:04:52 +00:00
socket_gethostname(PyObject *self, PyObject *unused)
Add gethostname() and s.avail().
1991-07-27 21:42:02 +00:00
{
char buf[1024];
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
int res;
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
res = gethostname(buf, (int) sizeof buf - 1);
new style
1994-10-10 17:59:00 +00:00
Py_END_ALLOW_THREADS
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
if (res < 0)
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
return set_error();
Add gethostname() and s.avail().
1991-07-27 21:42:02 +00:00
buf[sizeof buf - 1] = '\0';
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromString(buf);
Add gethostname() and s.avail().
1991-07-27 21:42:02 +00:00
}
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(gethostname_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"gethostname() -> string\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Return the current host name.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* Python interface to gethostbyname(name). */
/*ARGSUSED*/
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_gethostbyname(PyObject *self, PyObject *args)
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
{
Changes to make it work and compile on NT, by Mark Hammond. (NT changes for posixmodule.c re-invented by Guido.)
1994-08-18 15:42:46 +00:00
char *name;
- Changes donated by Elemental Security to make it work on AIX 5.3 with IBM's 64-bit compiler (SF patch #1284289). This also closes SF bug #105470: test_pwd fails on 64bit system (Opteron).
2005-09-14 18:09:42 +00:00
sock_addr_t addrbuf;
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
Fix all routines to use PyArg_ParseTuple(), and add ":name" to the argument format strings. THIS WILL PROBABLY BREAK LOTS OF CODE!!! Also fixed a bogus string in an error message in getsockaddrlen().
2000-03-24 20:56:56 +00:00
if (!PyArg_ParseTuple(args, "s:gethostbyname", &name))
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
return NULL;
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
if (setipaddr(name, SAS2SA(&addrbuf), sizeof(addrbuf), AF_INET) < 0)
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
return NULL;
Patch #1544279: Improve thread-safety of the socket module by moving the sock_addr_t storage out of the socket object. Will backport to 2.5.
2006-12-03 11:23:45 +00:00
return makeipaddr(SAS2SA(&addrbuf), sizeof(struct sockaddr_in));
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(gethostbyname_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"gethostbyname(host) -> address\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Return the IP address (a string of the form '255.255.255.255') for a host.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
/* Convenience function common to gethostbyname_ex and gethostbyaddr */
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
static PyObject *
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
gethost_common(struct hostent *h, struct sockaddr *addr, int alen, int af)
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
{
remove unused variables
1996-02-27 15:21:19 +00:00
char **pch;
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
PyObject *rtn_tuple = (PyObject *)NULL;
PyObject *name_list = (PyObject *)NULL;
PyObject *addr_list = (PyObject *)NULL;
PyObject *tmp;
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
if (h == NULL) {
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
/* Let's get real error message to return */
SF patch #474590 -- RISC OS support
2001-10-24 20:42:55 +00:00
#ifndef RISCOS
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
set_herror(h_errno);
SF patch #474590 -- RISC OS support
2001-10-24 20:42:55 +00:00
#else
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error, "host not found");
SF patch #474590 -- RISC OS support
2001-10-24 20:42:55 +00:00
#endif
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return NULL;
}
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
if (h->h_addrtype != af) {
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
/* Let's get real error message to return */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error,
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
(char *)strerror(EAFNOSUPPORT));
Remove all traces of HAVE_STRERROR. The removal of strerror.c led to the function check being removed from configure.in.
2008-03-18 15:35:58 +00:00
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
return NULL;
}
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
switch (af) {
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
case AF_INET:
if (alen < sizeof(struct sockaddr_in))
return NULL;
break;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Remove INET6 define. Use ENABLE_IPV6 instead.
2001-12-02 10:15:37 +00:00
#ifdef ENABLE_IPV6
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
case AF_INET6:
if (alen < sizeof(struct sockaddr_in6))
return NULL;
break;
#endif
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
}
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
if ((name_list = PyList_New(0)) == NULL)
goto err;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
if ((addr_list = PyList_New(0)) == NULL)
goto err;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #1511317: don't crash on invalid hostname info
2006-08-14 22:10:24 +00:00
/* SF #1511317: h_aliases can be NULL */
if (h->h_aliases) {
for (pch = h->h_aliases; *pch != NULL; pch++) {
int status;
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
tmp = PyString_FromString(*pch);
Patch #1511317: don't crash on invalid hostname info
2006-08-14 22:10:24 +00:00
if (tmp == NULL)
goto err;
status = PyList_Append(name_list, tmp);
Py_DECREF(tmp);
if (status)
goto err;
}
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
}
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
for (pch = h->h_addr_list; *pch != NULL; pch++) {
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
int status;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
switch (af) {
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
case AF_INET:
{
struct sockaddr_in sin;
memset(&sin, 0, sizeof(sin));
sin.sin_family = af;
#ifdef HAVE_SOCKADDR_SA_LEN
sin.sin_len = sizeof(sin);
#endif
memcpy(&sin.sin_addr, *pch, sizeof(sin.sin_addr));
tmp = makeipaddr((struct sockaddr *)&sin, sizeof(sin));
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
if (pch == h->h_addr_list && alen >= sizeof(sin))
memcpy((char *) addr, &sin, sizeof(sin));
break;
}
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Remove INET6 define. Use ENABLE_IPV6 instead.
2001-12-02 10:15:37 +00:00
#ifdef ENABLE_IPV6
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
case AF_INET6:
{
struct sockaddr_in6 sin6;
memset(&sin6, 0, sizeof(sin6));
sin6.sin6_family = af;
#ifdef HAVE_SOCKADDR_SA_LEN
sin6.sin6_len = sizeof(sin6);
#endif
memcpy(&sin6.sin6_addr, *pch, sizeof(sin6.sin6_addr));
tmp = makeipaddr((struct sockaddr *)&sin6,
sizeof(sin6));
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
if (pch == h->h_addr_list && alen >= sizeof(sin6))
memcpy((char *) addr, &sin6, sizeof(sin6));
break;
}
#endif
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
default: /* can't happen */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error,
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
"unsupported address family");
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return NULL;
}
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
if (tmp == NULL)
goto err;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
status = PyList_Append(addr_list, tmp);
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
Py_DECREF(tmp);
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Check of return values and proper error handling.
1997-01-03 17:18:10 +00:00
if (status)
goto err;
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
}
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
rtn_tuple = Py_BuildValue("sOO", h->h_name, name_list, addr_list);
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
Another bulky set of minor changes. Note addition of gethostbyaddr() and improved repr() for sockets, renaming of md5.md5() to md5.new(), and fixing of leaks in threads.
1995-01-02 19:30:30 +00:00
err:
Py_XDECREF(name_list);
Py_XDECREF(addr_list);
return rtn_tuple;
}
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
/* Python interface to gethostbyname_ex(name). */
/*ARGSUSED*/
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_gethostbyname_ex(PyObject *self, PyObject *args)
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
{
char *name;
struct hostent *h;
Patch #708495: Port more stuff to OpenVMS.
2003-05-03 09:14:54 +00:00
#ifdef ENABLE_IPV6
struct sockaddr_storage addr;
#else
struct sockaddr_in addr;
#endif
Instead of accessing ss_family, cast sockaddr_storage to sockaddr and access sa_family.
2001-07-23 01:30:10 +00:00
struct sockaddr *sa;
Jonathan Giddy writes: Here's a patch to fix the race condition, which wasn't fixed by Rob's patch. It holds the gethostbyname lock until the results are copied out, which means that this lock and the Python global lock are held at the same time. This shouldn't be a problem as long as the gethostbyname lock is always acquired when the global lock is not held.
1999-03-22 20:14:53 +00:00
PyObject *ret;
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
#ifdef HAVE_GETHOSTBYNAME_R
struct hostent hp_allocated;
Vladimir Marangozov implements the AIX 3-arg gethostbyname_r code.
1999-03-24 17:20:40 +00:00
#ifdef HAVE_GETHOSTBYNAME_R_3_ARG
struct hostent_data data;
#else
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
char buf[16384];
int buf_len = (sizeof buf) - 1;
int errnop;
Vladimir Marangozov implements the AIX 3-arg gethostbyname_r code.
1999-03-24 17:20:40 +00:00
#endif
#if defined(HAVE_GETHOSTBYNAME_R_3_ARG) || defined(HAVE_GETHOSTBYNAME_R_6_ARG)
Patch by Rob Riggs for Linux -- glibc2 has a different argument converntion for gethostbyname_r() etc. than Solaris!
1999-03-15 21:40:14 +00:00
int result;
#endif
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
#endif /* HAVE_GETHOSTBYNAME_R */
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
Fix all routines to use PyArg_ParseTuple(), and add ":name" to the argument format strings. THIS WILL PROBABLY BREAK LOTS OF CODE!!! Also fixed a bogus string in an error message in getsockaddrlen().
2000-03-24 20:56:56 +00:00
if (!PyArg_ParseTuple(args, "s:gethostbyname_ex", &name))
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
return NULL;
Patch #650422: Use Posix AF_ constants instead of PF_ ones.
2002-12-11 13:10:57 +00:00
if (setipaddr(name, (struct sockaddr *)&addr, sizeof(addr), AF_INET) < 0)
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
return NULL;
Py_BEGIN_ALLOW_THREADS
#ifdef HAVE_GETHOSTBYNAME_R
Vladimir Marangozov implements the AIX 3-arg gethostbyname_r code.
1999-03-24 17:20:40 +00:00
#if defined(HAVE_GETHOSTBYNAME_R_6_ARG)
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
result = gethostbyname_r(name, &hp_allocated, buf, buf_len,
&h, &errnop);
Vladimir Marangozov implements the AIX 3-arg gethostbyname_r code.
1999-03-24 17:20:40 +00:00
#elif defined(HAVE_GETHOSTBYNAME_R_5_ARG)
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
h = gethostbyname_r(name, &hp_allocated, buf, buf_len, &errnop);
Vladimir Marangozov implements the AIX 3-arg gethostbyname_r code.
1999-03-24 17:20:40 +00:00
#else /* HAVE_GETHOSTBYNAME_R_3_ARG */
Implement two suggestions by Jonathan Giddy: (1) in AIX, clear the data struct before calling gethostby{name,addr}_r(); (2) ignore the 3/5/6 args determinations made by the configure script and switch on platform identifiers instead: AIX, OSF have 3 args Sun, SGI have 5 args Linux has 6 args On all other platforms, undef HAVE_GETHOSTBYNAME_R altogether.
1999-03-24 17:24:33 +00:00
memset((void *) &data, '\0', sizeof(data));
Vladimir Marangozov implements the AIX 3-arg gethostbyname_r code.
1999-03-24 17:20:40 +00:00
result = gethostbyname_r(name, &hp_allocated, &data);
h = (result != 0) ? NULL : &hp_allocated;
Patch by Rob Riggs for Linux -- glibc2 has a different argument converntion for gethostbyname_r() etc. than Solaris!
1999-03-15 21:40:14 +00:00
#endif
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
#else /* not HAVE_GETHOSTBYNAME_R */
Clean up pass for the previous patches. - Use HAVE_GETHOSTBYNAME_R_6_ARG instead of testing for Linux and glibc2. - If gethostbyname takes 3 args, undefine HAVE_GETHOSTBYNAME_R -- don't know what code should be used. - New symbol USE_GETHOSTBYNAME_LOCK defined iff the lock should be used. - Modify the gethostbyaddr() code to also hold on to the lock until after it is safe to release, overlapping with the Python lock. (Note: I think that it could in theory be possible that Python code executed while gethostbyname_lock is held could attempt to reacquire the lock -- e.g. in a signal handler or destructor. I will simply say "don't do that then.")
1999-03-22 21:44:51 +00:00
#ifdef USE_GETHOSTBYNAME_LOCK
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
PyThread_acquire_lock(netdb_lock, 1);
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
#endif
h = gethostbyname(name);
#endif /* HAVE_GETHOSTBYNAME_R */
Py_END_ALLOW_THREADS
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
/* Some C libraries would require addr.__ss_family instead of
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
addr.ss_family.
Therefore, we cast the sockaddr_storage into sockaddr to
access sa_family. */
Instead of accessing ss_family, cast sockaddr_storage to sockaddr and access sa_family.
2001-07-23 01:30:10 +00:00
sa = (struct sockaddr*)&addr;
SF patch 555085 (timeout socket implementation) by Michael Gilfix. I've made considerable changes to Michael's code, specifically to use the select() system call directly and to store the timeout as a C double instead of a Python object; internally, -1.0 (or anything negative) represents the None from the API. I'm not 100% sure that all corner cases are covered correctly, so please keep an eye on this. Next I'm going to try it Windows before Tim complains. No way is this a bugfix candidate. :-)
2002-06-06 21:08:16 +00:00
ret = gethost_common(h, (struct sockaddr *)&addr, sizeof(addr),
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
sa->sa_family);
Clean up pass for the previous patches. - Use HAVE_GETHOSTBYNAME_R_6_ARG instead of testing for Linux and glibc2. - If gethostbyname takes 3 args, undefine HAVE_GETHOSTBYNAME_R -- don't know what code should be used. - New symbol USE_GETHOSTBYNAME_LOCK defined iff the lock should be used. - Modify the gethostbyaddr() code to also hold on to the lock until after it is safe to release, overlapping with the Python lock. (Note: I think that it could in theory be possible that Python code executed while gethostbyname_lock is held could attempt to reacquire the lock -- e.g. in a signal handler or destructor. I will simply say "don't do that then.")
1999-03-22 21:44:51 +00:00
#ifdef USE_GETHOSTBYNAME_LOCK
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
PyThread_release_lock(netdb_lock);
Jonathan Giddy writes: Here's a patch to fix the race condition, which wasn't fixed by Rob's patch. It holds the gethostbyname lock until the results are copied out, which means that this lock and the Python global lock are held at the same time. This shouldn't be a problem as long as the gethostbyname lock is always acquired when the global lock is not held.
1999-03-22 20:14:53 +00:00
#endif
return ret;
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(ghbn_ex_doc,
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
"gethostbyname_ex(host) -> (name, aliaslist, addresslist)\n\
\n\
Return the true host name, a list of aliases, and a list of IP addresses,\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
for a host. The host argument is a string giving a host name or IP number.");
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
/* Python interface to gethostbyaddr(IP). */
/*ARGSUSED*/
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_gethostbyaddr(PyObject *self, PyObject *args)
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
{
Remove INET6 define. Use ENABLE_IPV6 instead.
2001-12-02 10:15:37 +00:00
#ifdef ENABLE_IPV6
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
struct sockaddr_storage addr;
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#else
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
struct sockaddr_in addr;
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
struct sockaddr *sa = (struct sockaddr *)&addr;
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
char *ip_num;
struct hostent *h;
Clean up pass for the previous patches. - Use HAVE_GETHOSTBYNAME_R_6_ARG instead of testing for Linux and glibc2. - If gethostbyname takes 3 args, undefine HAVE_GETHOSTBYNAME_R -- don't know what code should be used. - New symbol USE_GETHOSTBYNAME_LOCK defined iff the lock should be used. - Modify the gethostbyaddr() code to also hold on to the lock until after it is safe to release, overlapping with the Python lock. (Note: I think that it could in theory be possible that Python code executed while gethostbyname_lock is held could attempt to reacquire the lock -- e.g. in a signal handler or destructor. I will simply say "don't do that then.")
1999-03-22 21:44:51 +00:00
PyObject *ret;
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
#ifdef HAVE_GETHOSTBYNAME_R
struct hostent hp_allocated;
Vladimir Marangozov implements the AIX 3-arg gethostbyname_r code.
1999-03-24 17:20:40 +00:00
#ifdef HAVE_GETHOSTBYNAME_R_3_ARG
struct hostent_data data;
#else
Fix issue 4884, preventing a crash in the socket code when python is compiled with llvm-gcc and run with a glibc <2.10.
2009-01-09 16:47:07 +00:00
/* glibcs up to 2.10 assume that the buf argument to
gethostbyaddr_r is 8-byte aligned, which at least llvm-gcc
does not ensure. The attribute below instructs the compiler
to maintain this alignment. */
char buf[16384] Py_ALIGNED(8);
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
int buf_len = (sizeof buf) - 1;
int errnop;
Vladimir Marangozov implements the AIX 3-arg gethostbyname_r code.
1999-03-24 17:20:40 +00:00
#endif
#if defined(HAVE_GETHOSTBYNAME_R_3_ARG) || defined(HAVE_GETHOSTBYNAME_R_6_ARG)
Patch by Rob Riggs for Linux -- glibc2 has a different argument converntion for gethostbyname_r() etc. than Solaris!
1999-03-15 21:40:14 +00:00
int result;
#endif
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
#endif /* HAVE_GETHOSTBYNAME_R */
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
char *ap;
int al;
int af;
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
Fix all routines to use PyArg_ParseTuple(), and add ":name" to the argument format strings. THIS WILL PROBABLY BREAK LOTS OF CODE!!! Also fixed a bogus string in an error message in getsockaddrlen().
2000-03-24 20:56:56 +00:00
if (!PyArg_ParseTuple(args, "s:gethostbyaddr", &ip_num))
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
return NULL;
Patch #650422: Use Posix AF_ constants instead of PF_ ones.
2002-12-11 13:10:57 +00:00
af = AF_UNSPEC;
Pass length of result structure into setipaddr. Fixes bug #565747. Committed to 2.2 branch.
2002-07-28 16:10:31 +00:00
if (setipaddr(ip_num, sa, sizeof(addr), af) < 0)
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return NULL;
af = sa->sa_family;
ap = NULL;
al = 0;
switch (af) {
case AF_INET:
ap = (char *)&((struct sockaddr_in *)sa)->sin_addr;
al = sizeof(((struct sockaddr_in *)sa)->sin_addr);
break;
Remove INET6 define. Use ENABLE_IPV6 instead.
2001-12-02 10:15:37 +00:00
#ifdef ENABLE_IPV6
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
case AF_INET6:
ap = (char *)&((struct sockaddr_in6 *)sa)->sin6_addr;
al = sizeof(((struct sockaddr_in6 *)sa)->sin6_addr);
break;
#endif
default:
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error, "unsupported address family");
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
return NULL;
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
}
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
Py_BEGIN_ALLOW_THREADS
#ifdef HAVE_GETHOSTBYNAME_R
Vladimir Marangozov implements the AIX 3-arg gethostbyname_r code.
1999-03-24 17:20:40 +00:00
#if defined(HAVE_GETHOSTBYNAME_R_6_ARG)
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
result = gethostbyaddr_r(ap, al, af,
&hp_allocated, buf, buf_len,
Patch by Rob Riggs for Linux -- glibc2 has a different argument converntion for gethostbyname_r() etc. than Solaris!
1999-03-15 21:40:14 +00:00
&h, &errnop);
Vladimir Marangozov implements the AIX 3-arg gethostbyname_r code.
1999-03-24 17:20:40 +00:00
#elif defined(HAVE_GETHOSTBYNAME_R_5_ARG)
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
h = gethostbyaddr_r(ap, al, af,
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
&hp_allocated, buf, buf_len, &errnop);
Vladimir Marangozov implements the AIX 3-arg gethostbyname_r code.
1999-03-24 17:20:40 +00:00
#else /* HAVE_GETHOSTBYNAME_R_3_ARG */
Implement two suggestions by Jonathan Giddy: (1) in AIX, clear the data struct before calling gethostby{name,addr}_r(); (2) ignore the 3/5/6 args determinations made by the configure script and switch on platform identifiers instead: AIX, OSF have 3 args Sun, SGI have 5 args Linux has 6 args On all other platforms, undef HAVE_GETHOSTBYNAME_R altogether.
1999-03-24 17:24:33 +00:00
memset((void *) &data, '\0', sizeof(data));
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
result = gethostbyaddr_r(ap, al, af, &hp_allocated, &data);
Vladimir Marangozov implements the AIX 3-arg gethostbyname_r code.
1999-03-24 17:20:40 +00:00
h = (result != 0) ? NULL : &hp_allocated;
Patch by Rob Riggs for Linux -- glibc2 has a different argument converntion for gethostbyname_r() etc. than Solaris!
1999-03-15 21:40:14 +00:00
#endif
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
#else /* not HAVE_GETHOSTBYNAME_R */
Clean up pass for the previous patches. - Use HAVE_GETHOSTBYNAME_R_6_ARG instead of testing for Linux and glibc2. - If gethostbyname takes 3 args, undefine HAVE_GETHOSTBYNAME_R -- don't know what code should be used. - New symbol USE_GETHOSTBYNAME_LOCK defined iff the lock should be used. - Modify the gethostbyaddr() code to also hold on to the lock until after it is safe to release, overlapping with the Python lock. (Note: I think that it could in theory be possible that Python code executed while gethostbyname_lock is held could attempt to reacquire the lock -- e.g. in a signal handler or destructor. I will simply say "don't do that then.")
1999-03-22 21:44:51 +00:00
#ifdef USE_GETHOSTBYNAME_LOCK
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
PyThread_acquire_lock(netdb_lock, 1);
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
#endif
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
h = gethostbyaddr(ap, al, af);
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
#endif /* HAVE_GETHOSTBYNAME_R */
Py_END_ALLOW_THREADS
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
ret = gethost_common(h, (struct sockaddr *)&addr, sizeof(addr), af);
Clean up pass for the previous patches. - Use HAVE_GETHOSTBYNAME_R_6_ARG instead of testing for Linux and glibc2. - If gethostbyname takes 3 args, undefine HAVE_GETHOSTBYNAME_R -- don't know what code should be used. - New symbol USE_GETHOSTBYNAME_LOCK defined iff the lock should be used. - Modify the gethostbyaddr() code to also hold on to the lock until after it is safe to release, overlapping with the Python lock. (Note: I think that it could in theory be possible that Python code executed while gethostbyname_lock is held could attempt to reacquire the lock -- e.g. in a signal handler or destructor. I will simply say "don't do that then.")
1999-03-22 21:44:51 +00:00
#ifdef USE_GETHOSTBYNAME_LOCK
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
PyThread_release_lock(netdb_lock);
Clean up pass for the previous patches. - Use HAVE_GETHOSTBYNAME_R_6_ARG instead of testing for Linux and glibc2. - If gethostbyname takes 3 args, undefine HAVE_GETHOSTBYNAME_R -- don't know what code should be used. - New symbol USE_GETHOSTBYNAME_LOCK defined iff the lock should be used. - Modify the gethostbyaddr() code to also hold on to the lock until after it is safe to release, overlapping with the Python lock. (Note: I think that it could in theory be possible that Python code executed while gethostbyname_lock is held could attempt to reacquire the lock -- e.g. in a signal handler or destructor. I will simply say "don't do that then.")
1999-03-22 21:44:51 +00:00
#endif
return ret;
Added gethostbyname_ex(), which returns the same kind of data as gethostbyaddr(). (Plain gethostbyname() returns only the IP address.) This moves the code shared by gethostbyaddr() and gethostbyname_ex() to a subroutine. Original patch by Dan Stromberg; some tweaks by GvR.
1998-08-04 22:16:43 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(gethostbyaddr_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"gethostbyaddr(host) -> (name, aliaslist, addresslist)\n\
\n\
Return the true host name, a list of aliases, and a list of IP addresses,\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
for a host. The host argument is a string giving a host name or IP number.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* Python interface to getservbyname(name).
This only returns the port number, since the other info is already
known or not useful (like the list of aliases). */
/*ARGSUSED*/
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_getservbyname(PyObject *self, PyObject *args)
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
{
Added socket.getservbyport(), and make its second argument and that of getservbyname() optional. Update the tests and the docs.
2004-06-28 00:50:43 +00:00
char *name, *proto=NULL;
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
struct servent *sp;
Added socket.getservbyport(), and make its second argument and that of getservbyname() optional. Update the tests and the docs.
2004-06-28 00:50:43 +00:00
if (!PyArg_ParseTuple(args, "s|s:getservbyname", &name, &proto))
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
return NULL;
new style
1994-10-10 17:59:00 +00:00
Py_BEGIN_ALLOW_THREADS
* Makefile: cosmetics * socketmodule.c: get rid of makepair(); fix makesocketaddr to fix broken recvfrom() * socketmodule: get rid of getStrarg() * ceval.h: move eval_code() to new file eval.h, so compile.h is no longer needed. * ceval.c: move thread comments to ceval.h; always make save/restore thread functions available (for dynloaded modules) * cdmodule.c, listobject.c: don't include compile.h * flmodule.c: include ceval.h * import.c: include eval.h instead of ceval.h * cgen.py: add forground(); noport(); winopen(""); to initgl(). * bltinmodule.c, socketmodule.c, fileobject.c, posixmodule.c, selectmodule.c: adapt to threads (add BGN/END SAVE macros) * stdwinmodule.c: adapt to threads and use a special stdwin lock. * pythonmain.c: don't include getpythonpath(). * pythonrun.c: use BGN/END SAVE instead of direct calls; also more BGN/END SAVE calls etc. * thread.c: bigger stack size for sun; change exit() to _exit() * threadmodule.c: use BGN/END SAVE macros where possible * timemodule.c: adapt better to threads; use BGN/END SAVE; add longsleep internal function if BSD_TIME; cosmetics
1992-08-05 19:58:53 +00:00
sp = getservbyname(name, proto);
new style
1994-10-10 17:59:00 +00:00
Py_END_ALLOW_THREADS
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
if (sp == NULL) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error, "service/proto not found");
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
return NULL;
}
new style
1994-10-10 17:59:00 +00:00
return PyInt_FromLong((long) ntohs(sp->s_port));
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(getservbyname_doc,
Added socket.getservbyport(), and make its second argument and that of getservbyname() optional. Update the tests and the docs.
2004-06-28 00:50:43 +00:00
"getservbyname(servicename[, protocolname]) -> integer\n\
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
\n\
Return a port number from a service name and protocol name.\n\
Added socket.getservbyport(), and make its second argument and that of getservbyname() optional. Update the tests and the docs.
2004-06-28 00:50:43 +00:00
The optional protocol name, if given, should be 'tcp' or 'udp',\n\
otherwise any protocol will match.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Added socket.getservbyport(), and make its second argument and that of getservbyname() optional. Update the tests and the docs.
2004-06-28 00:50:43 +00:00
/* Python interface to getservbyport(port).
This only returns the service name, since the other info is already
known or not useful (like the list of aliases). */
/*ARGSUSED*/
static PyObject *
socket_getservbyport(PyObject *self, PyObject *args)
{
- Issue #5104: The socket module now raises OverflowError when 16-bit port and protocol numbers are supplied outside the allowed 0-65536 range on bind() and getservbyport().
2009-01-31 22:57:30 +00:00
int port;
Added socket.getservbyport(), and make its second argument and that of getservbyname() optional. Update the tests and the docs.
2004-06-28 00:50:43 +00:00
char *proto=NULL;
struct servent *sp;
- Issue #5104: The socket module now raises OverflowError when 16-bit port and protocol numbers are supplied outside the allowed 0-65536 range on bind() and getservbyport().
2009-01-31 22:57:30 +00:00
if (!PyArg_ParseTuple(args, "i|s:getservbyport", &port, &proto))
Added socket.getservbyport(), and make its second argument and that of getservbyname() optional. Update the tests and the docs.
2004-06-28 00:50:43 +00:00
return NULL;
- Issue #5104: The socket module now raises OverflowError when 16-bit port and protocol numbers are supplied outside the allowed 0-65536 range on bind() and getservbyport().
2009-01-31 22:57:30 +00:00
if (port < 0 || port > 0xffff) {
PyErr_SetString(
PyExc_OverflowError,
"getservbyport: port must be 0-65535.");
return NULL;
}
Added socket.getservbyport(), and make its second argument and that of getservbyname() optional. Update the tests and the docs.
2004-06-28 00:50:43 +00:00
Py_BEGIN_ALLOW_THREADS
- Issue #5104: The socket module now raises OverflowError when 16-bit port and protocol numbers are supplied outside the allowed 0-65536 range on bind() and getservbyport().
2009-01-31 22:57:30 +00:00
sp = getservbyport(htons((short)port), proto);
Added socket.getservbyport(), and make its second argument and that of getservbyname() optional. Update the tests and the docs.
2004-06-28 00:50:43 +00:00
Py_END_ALLOW_THREADS
if (sp == NULL) {
PyErr_SetString(socket_error, "port/proto not found");
return NULL;
}
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromString(sp->s_name);
Added socket.getservbyport(), and make its second argument and that of getservbyname() optional. Update the tests and the docs.
2004-06-28 00:50:43 +00:00
}
PyDoc_STRVAR(getservbyport_doc,
"getservbyport(port[, protocolname]) -> string\n\
\n\
Return the service name from a port number and protocol name.\n\
The optional protocol name, if given, should be 'tcp' or 'udp',\n\
otherwise any protocol will match.");
Added getprotobyname() interface.
1996-12-19 16:35:04 +00:00
/* Python interface to getprotobyname(name).
This only returns the protocol number, since the other info is
already known or not useful (like the list of aliases). */
/*ARGSUSED*/
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_getprotobyname(PyObject *self, PyObject *args)
Added getprotobyname() interface.
1996-12-19 16:35:04 +00:00
{
char *name;
struct protoent *sp;
Changes for BeOS, QNX and long long, by Chris Herborth.
1998-08-04 22:53:56 +00:00
#ifdef __BEOS__
/* Not available in BeOS yet. - [cjh] */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error, "getprotobyname not supported");
Changes for BeOS, QNX and long long, by Chris Herborth.
1998-08-04 22:53:56 +00:00
return NULL;
#else
Fix all routines to use PyArg_ParseTuple(), and add ":name" to the argument format strings. THIS WILL PROBABLY BREAK LOTS OF CODE!!! Also fixed a bogus string in an error message in getsockaddrlen().
2000-03-24 20:56:56 +00:00
if (!PyArg_ParseTuple(args, "s:getprotobyname", &name))
Added getprotobyname() interface.
1996-12-19 16:35:04 +00:00
return NULL;
Py_BEGIN_ALLOW_THREADS
sp = getprotobyname(name);
Py_END_ALLOW_THREADS
if (sp == NULL) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error, "protocol not found");
Added getprotobyname() interface.
1996-12-19 16:35:04 +00:00
return NULL;
}
return PyInt_FromLong((long) sp->p_proto);
Changes for BeOS, QNX and long long, by Chris Herborth.
1998-08-04 22:53:56 +00:00
#endif
Added getprotobyname() interface.
1996-12-19 16:35:04 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(getprotobyname_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"getprotobyname(name) -> integer\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Return the protocol number for the named protocol. (Rarely used.)");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Added getprotobyname() interface.
1996-12-19 16:35:04 +00:00
Patch #1003700: Add socketpair function to socket module.
2004-08-09 04:51:41 +00:00
#ifdef HAVE_SOCKETPAIR
/* Create a pair of sockets using the socketpair() function.
Updated the socketpair() docstring and documentation to explain that the default famility is AF_UNIX if defined for the platform, otherwise the default is AF_INET.
2004-08-23 05:16:23 +00:00
Arguments as for socket() except the default family is AF_UNIX if
Patch #1015012. Improve markup and punctuation in libsocket.tex
2004-08-26 00:51:16 +00:00
defined on the platform; otherwise, the default is AF_INET. */
Patch #1003700: Add socketpair function to socket module.
2004-08-09 04:51:41 +00:00
/*ARGSUSED*/
static PyObject *
socket_socketpair(PyObject *self, PyObject *args)
{
PySocketSockObject *s0 = NULL, *s1 = NULL;
SOCKET_T sv[2];
int family, type = SOCK_STREAM, proto = 0;
PyObject *res = NULL;
#if defined(AF_UNIX)
family = AF_UNIX;
#else
family = AF_INET;
#endif
if (!PyArg_ParseTuple(args, "|iii:socketpair",
&family, &type, &proto))
return NULL;
/* Create a pair of socket fds */
if (socketpair(family, type, proto, sv) < 0)
return set_error();
s0 = new_sockobject(sv[0], family, type, proto);
if (s0 == NULL)
goto finally;
s1 = new_sockobject(sv[1], family, type, proto);
if (s1 == NULL)
goto finally;
res = PyTuple_Pack(2, s0, s1);
finally:
if (res == NULL) {
if (s0 == NULL)
SOCKETCLOSE(sv[0]);
if (s1 == NULL)
SOCKETCLOSE(sv[1]);
}
Py_XDECREF(s0);
Py_XDECREF(s1);
return res;
}
PyDoc_STRVAR(socketpair_doc,
"socketpair([family[, type[, proto]]]) -> (socket object, socket object)\n\
\n\
Create a pair of socket objects from the sockets returned by the platform\n\
socketpair() function.\n\
Updated the socketpair() docstring and documentation to explain that the default famility is AF_UNIX if defined for the platform, otherwise the default is AF_INET.
2004-08-23 05:16:23 +00:00
The arguments are the same as for socket() except the default family is\n\
Patch #1015012. Improve markup and punctuation in libsocket.tex
2004-08-26 00:51:16 +00:00
AF_UNIX if defined on the platform; otherwise, the default is AF_INET.");
Patch #1003700: Add socketpair function to socket module.
2004-08-09 04:51:41 +00:00
#endif /* HAVE_SOCKETPAIR */
Changes to make it work and compile on NT, by Mark Hammond. (NT changes for posixmodule.c re-invented by Guido.)
1994-08-18 15:42:46 +00:00
#ifndef NO_DUP
socketmodule.c: added socket.fromfd(fd, family, type, [proto]); converted socket() to use of getargs().
1992-12-08 13:38:24 +00:00
/* Create a socket object from a numeric file description.
Useful e.g. if stdin is a socket.
Additional arguments as for socket(). */
/*ARGSUSED*/
new style
1994-10-10 17:59:00 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_fromfd(PyObject *self, PyObject *args)
socketmodule.c: added socket.fromfd(fd, family, type, [proto]); converted socket() to use of getargs().
1992-12-08 13:38:24 +00:00
{
new style
1994-10-10 17:59:00 +00:00
PySocketSockObject *s;
Trent Mick <trentm@activestate.com>: This patch fixes possible overflows in the socket module for 64-bit platforms (mainly Win64). The changes are: - abstract the socket type to SOCKET_T (this is SOCKET on Windows, int on Un*x), this is necessary because sizeof(SOCKET) > sizeof(int) on Win64 - use INVALID_SOCKET on Win32/64 for an error return value for accept() - ensure no overflow of the socket variable for: (1) a PyObject return value (use PyLong_FromLongLong if necessary); and (2) printf formatting in repr(). Closes SourceForge patch #100516.
2000-06-30 02:46:07 +00:00
SOCKET_T fd;
int family, type, proto = 0;
Fix all routines to use PyArg_ParseTuple(), and add ":name" to the argument format strings. THIS WILL PROBABLY BREAK LOTS OF CODE!!! Also fixed a bogus string in an error message in getsockaddrlen().
2000-03-24 20:56:56 +00:00
if (!PyArg_ParseTuple(args, "iii|i:fromfd",
&fd, &family, &type, &proto))
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
return NULL;
* mymalloc.h: always allocate one extra byte, since some malloc's return NULL for malloc(0) or realloc(p, 0). (This should be done differently than wasting one byte, but alas...) * Moved "add'l libraries" option in Makefile to an earlier place. * Remove argument compatibility hacks (b) and (c). * Add grey2mono, dither2mono and mono2grey to imageop. * Dup the fd in socket.fromfd(). * Added new modules mpz, md5 (by JH, requiring GNU MP 1.2). Affects Makefile and config.c. * socketmodule.c: added socket.fromfd(fd, family, type, [proto]), converted socket() to use of getargs().
1992-12-14 16:59:51 +00:00
/* Dup the fd so it and the socket can be closed independently */
fd = dup(fd);
if (fd < 0)
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
return set_error();
s = new_sockobject(fd, family, type, proto);
new style
1994-10-10 17:59:00 +00:00
return (PyObject *) s;
socketmodule.c: added socket.fromfd(fd, family, type, [proto]); converted socket() to use of getargs().
1992-12-08 13:38:24 +00:00
}
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(fromfd_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"fromfd(fd, family, type[, proto]) -> socket object\n\
\n\
Bug #1460564: document that socket.fromfd() duplicates the given file descriptor.
2006-04-01 07:33:08 +00:00
Create a socket object from a duplicate of the given\n\
file descriptor.\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
The remaining arguments are the same as for socket().");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Changes to make it work and compile on NT, by Mark Hammond. (NT changes for posixmodule.c re-invented by Guido.)
1994-08-18 15:42:46 +00:00
#endif /* NO_DUP */
socketmodule.c: added socket.fromfd(fd, family, type, [proto]); converted socket() to use of getargs().
1992-12-08 13:38:24 +00:00
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_ntohs(PyObject *self, PyObject *args)
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
{
int x1, x2;
Fix all routines to use PyArg_ParseTuple(), and add ":name" to the argument format strings. THIS WILL PROBABLY BREAK LOTS OF CODE!!! Also fixed a bogus string in an error message in getsockaddrlen().
2000-03-24 20:56:56 +00:00
if (!PyArg_ParseTuple(args, "i:ntohs", &x1)) {
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
return NULL;
}
Patch #1635058 by Mark Roberts: ensure that htonl and friends never accept or return negative numbers, per the underlying C implementation.
2007-01-14 17:03:32 +00:00
if (x1 < 0) {
PyErr_SetString(PyExc_OverflowError,
"can't convert negative number to unsigned long");
return NULL;
}
x2 = (unsigned int)ntohs((unsigned short)x1);
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
return PyInt_FromLong(x2);
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(ntohs_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"ntohs(integer) -> integer\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Convert a 16-bit integer from network to host byte order.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
static PyObject *
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
socket_ntohl(PyObject *self, PyObject *arg)
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
{
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
unsigned long x;
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
if (PyInt_Check(arg)) {
x = PyInt_AS_LONG(arg);
Put checks for error returns in the right place.
2002-07-25 16:37:51 +00:00
if (x == (unsigned long) -1 && PyErr_Occurred())
return NULL;
Patch #1635058 by Mark Roberts: ensure that htonl and friends never accept or return negative numbers, per the underlying C implementation.
2007-01-14 17:03:32 +00:00
if ((long)x < 0) {
PyErr_SetString(PyExc_OverflowError,
"can't convert negative number to unsigned long");
return NULL;
}
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
}
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
else if (PyLong_Check(arg)) {
x = PyLong_AsUnsignedLong(arg);
Put checks for error returns in the right place.
2002-07-25 16:37:51 +00:00
if (x == (unsigned long) -1 && PyErr_Occurred())
return NULL;
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
#if SIZEOF_LONG > 4
{
unsigned long y;
/* only want the trailing 32 bits */
y = x & 0xFFFFFFFFUL;
if (y ^ x)
return PyErr_Format(PyExc_OverflowError,
"long int larger than 32 bits");
x = y;
internal_connect(): Windows. When sock_timeout > 0 and connect() yields WSAEWOULDBLOCK, the second connect() attempt appears to yield WSAEISCONN on Win98 but WSAEINVAL on Win2K. So accept either as meaning "yawn, fine". This allows test_socket to succeed on my Win2K box (which it already did on my Win98SE box).
2002-08-06 22:25:02 +00:00
}
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
#endif
}
else
internal_connect(): Windows. When sock_timeout > 0 and connect() yields WSAEWOULDBLOCK, the second connect() attempt appears to yield WSAEISCONN on Win98 but WSAEINVAL on Win2K. So accept either as meaning "yawn, fine". This allows test_socket to succeed on my Win2K box (which it already did on my Win98SE box).
2002-08-06 22:25:02 +00:00
return PyErr_Format(PyExc_TypeError,
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
"expected int/long, %s found",
#1629: Renamed Py_Size, Py_Type and Py_Refcnt to Py_SIZE, Py_TYPE and Py_REFCNT. Macros for b/w compatibility are available.
2007-12-19 02:37:44 +00:00
Py_TYPE(arg)->tp_name);
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
if (x == (unsigned long) -1 && PyErr_Occurred())
return NULL;
Patch #1635058 by Mark Roberts: ensure that htonl and friends never accept or return negative numbers, per the underlying C implementation.
2007-01-14 17:03:32 +00:00
return PyLong_FromUnsignedLong(ntohl(x));
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(ntohl_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"ntohl(integer) -> integer\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Convert a 32-bit integer from network to host byte order.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_htons(PyObject *self, PyObject *args)
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
{
Fix erroneus argument parsing of socket.htons() on 64bit big endian machines.
2003-12-17 09:40:23 +00:00
int x1, x2;
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
Fix all routines to use PyArg_ParseTuple(), and add ":name" to the argument format strings. THIS WILL PROBABLY BREAK LOTS OF CODE!!! Also fixed a bogus string in an error message in getsockaddrlen().
2000-03-24 20:56:56 +00:00
if (!PyArg_ParseTuple(args, "i:htons", &x1)) {
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
return NULL;
}
Patch #1635058 by Mark Roberts: ensure that htonl and friends never accept or return negative numbers, per the underlying C implementation.
2007-01-14 17:03:32 +00:00
if (x1 < 0) {
PyErr_SetString(PyExc_OverflowError,
"can't convert negative number to unsigned long");
return NULL;
}
x2 = (unsigned int)htons((unsigned short)x1);
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
return PyInt_FromLong(x2);
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(htons_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"htons(integer) -> integer\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Convert a 16-bit integer from host to network byte order.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
static PyObject *
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
socket_htonl(PyObject *self, PyObject *arg)
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
{
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
unsigned long x;
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
if (PyInt_Check(arg)) {
x = PyInt_AS_LONG(arg);
Put checks for error returns in the right place.
2002-07-25 16:37:51 +00:00
if (x == (unsigned long) -1 && PyErr_Occurred())
return NULL;
Patch #1635058 by Mark Roberts: ensure that htonl and friends never accept or return negative numbers, per the underlying C implementation.
2007-01-14 17:03:32 +00:00
if ((long)x < 0) {
PyErr_SetString(PyExc_OverflowError,
"can't convert negative number to unsigned long");
return NULL;
}
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
}
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
else if (PyLong_Check(arg)) {
x = PyLong_AsUnsignedLong(arg);
Put checks for error returns in the right place.
2002-07-25 16:37:51 +00:00
if (x == (unsigned long) -1 && PyErr_Occurred())
return NULL;
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
#if SIZEOF_LONG > 4
{
unsigned long y;
/* only want the trailing 32 bits */
y = x & 0xFFFFFFFFUL;
if (y ^ x)
return PyErr_Format(PyExc_OverflowError,
"long int larger than 32 bits");
x = y;
internal_connect(): Windows. When sock_timeout > 0 and connect() yields WSAEWOULDBLOCK, the second connect() attempt appears to yield WSAEISCONN on Win98 but WSAEINVAL on Win2K. So accept either as meaning "yawn, fine". This allows test_socket to succeed on my Win2K box (which it already did on my Win98SE box).
2002-08-06 22:25:02 +00:00
}
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
#endif
}
else
internal_connect(): Windows. When sock_timeout > 0 and connect() yields WSAEWOULDBLOCK, the second connect() attempt appears to yield WSAEISCONN on Win98 but WSAEINVAL on Win2K. So accept either as meaning "yawn, fine". This allows test_socket to succeed on my Win2K box (which it already did on my Win98SE box).
2002-08-06 22:25:02 +00:00
return PyErr_Format(PyExc_TypeError,
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
"expected int/long, %s found",
#1629: Renamed Py_Size, Py_Type and Py_Refcnt to Py_SIZE, Py_TYPE and Py_REFCNT. Macros for b/w compatibility are available.
2007-12-19 02:37:44 +00:00
Py_TYPE(arg)->tp_name);
Patch #1635058 by Mark Roberts: ensure that htonl and friends never accept or return negative numbers, per the underlying C implementation.
2007-01-14 17:03:32 +00:00
return PyLong_FromUnsignedLong(htonl((unsigned long)x));
ntohs, ntohl, htons, htonl -- thanks to Lance
1996-06-12 04:04:55 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(htonl_doc,
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
"htonl(integer) -> integer\n\
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Convert a 32-bit integer from host to network byte order.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
/* socket.inet_aton() and socket.inet_ntoa() functions. */
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(inet_aton_doc,
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
"inet_aton(string) -> packed 32-bit IP representation\n\
\n\
Port inet_ntoa and inet_aton to Windows: - fix unescaped newline in string literal - removed unused err variable - Windows doesn't have inet_aton; use inet_addr instead
1999-08-20 19:11:27 +00:00
Convert an IP address in string format (123.45.67.89) to the 32-bit packed\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
binary format used in low-level network functions.");
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
static PyObject*
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_inet_aton(PyObject *self, PyObject *args)
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
{
It appears that inet_aton() doesn't really exist except in libresolv; the proper function to call is inet_addr(). Since we already had code to do that (for MS-Windows), this simplifies things a lot!
1999-09-09 15:42:59 +00:00
#ifndef INADDR_NONE
#define INADDR_NONE (-1)
#endif
Ummm, try to get it right this time
2003-02-13 02:15:42 +00:00
#ifdef HAVE_INET_ATON
Addressing SF bug #643005, implement socket.inet_aton() using inet_aton() rather than inet_addr() -- the latter is obsolete because it has a problem: "255.255.255.255" is a valid address but indistinguishable from an error. (I'm not sure if inet_aton() exists everywhere -- in case it doesn't, I've left the old code in with an #ifdef.)
2003-02-12 23:08:22 +00:00
struct in_addr buf;
Patch 1471925 - Weak linking support for OSX This patch causes several symbols in the socket and posix module to be weakly linked on OSX and disables usage of ftime on OSX. These changes make it possible to use a binary build on OSX 10.4 on a 10.3 system.
2006-04-23 11:59:25 +00:00
#endif
#if !defined(HAVE_INET_ATON) || defined(USE_INET_ATON_WEAKLINK)
Issue #1008086: Fixes socket.inet_aton() to always return 4 bytes even on LP64 platforms (most 64-bit Linux, bsd, unix systems).
2009-02-11 23:45:25 +00:00
#if (SIZEOF_INT != 4)
#error "Not sure if in_addr_t exists and int is not 32-bits."
#endif
socket_inet_aton(): ip_addr was left undefined before use in the !HAVE_INET_ATON case. Repaired that, and tried to repair what looked like out-of-date comments.
2003-02-13 03:13:40 +00:00
/* Have to use inet_addr() instead */
Issue #1008086: Fixes socket.inet_aton() to always return 4 bytes even on LP64 platforms (most 64-bit Linux, bsd, unix systems).
2009-02-11 23:45:25 +00:00
unsigned int packed_addr;
socket_inet_aton(): ip_addr was left undefined before use in the !HAVE_INET_ATON case. Repaired that, and tried to repair what looked like out-of-date comments.
2003-02-13 03:13:40 +00:00
#endif
char *ip_addr;
Port inet_ntoa and inet_aton to Windows: - fix unescaped newline in string literal - removed unused err variable - Windows doesn't have inet_aton; use inet_addr instead
1999-08-20 19:11:27 +00:00
socket_inet_aton(): ip_addr was left undefined before use in the !HAVE_INET_ATON case. Repaired that, and tried to repair what looked like out-of-date comments.
2003-02-13 03:13:40 +00:00
if (!PyArg_ParseTuple(args, "s:inet_aton", &ip_addr))
Port inet_ntoa and inet_aton to Windows: - fix unescaped newline in string literal - removed unused err variable - Windows doesn't have inet_aton; use inet_addr instead
1999-08-20 19:11:27 +00:00
return NULL;
Addressing SF bug #643005, implement socket.inet_aton() using inet_aton() rather than inet_addr() -- the latter is obsolete because it has a problem: "255.255.255.255" is a valid address but indistinguishable from an error. (I'm not sure if inet_aton() exists everywhere -- in case it doesn't, I've left the old code in with an #ifdef.)
2003-02-12 23:08:22 +00:00
socket_inet_aton(): ip_addr was left undefined before use in the !HAVE_INET_ATON case. Repaired that, and tried to repair what looked like out-of-date comments.
2003-02-13 03:13:40 +00:00
#ifdef HAVE_INET_ATON
Patch 1471925 - Weak linking support for OSX This patch causes several symbols in the socket and posix module to be weakly linked on OSX and disables usage of ftime on OSX. These changes make it possible to use a binary build on OSX 10.4 on a 10.3 system.
2006-04-23 11:59:25 +00:00
#ifdef USE_INET_ATON_WEAKLINK
if (inet_aton != NULL) {
#endif
Addressing SF bug #643005, implement socket.inet_aton() using inet_aton() rather than inet_addr() -- the latter is obsolete because it has a problem: "255.255.255.255" is a valid address but indistinguishable from an error. (I'm not sure if inet_aton() exists everywhere -- in case it doesn't, I've left the old code in with an #ifdef.)
2003-02-12 23:08:22 +00:00
if (inet_aton(ip_addr, &buf))
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromStringAndSize((char *)(&buf),
Addressing SF bug #643005, implement socket.inet_aton() using inet_aton() rather than inet_addr() -- the latter is obsolete because it has a problem: "255.255.255.255" is a valid address but indistinguishable from an error. (I'm not sure if inet_aton() exists everywhere -- in case it doesn't, I've left the old code in with an #ifdef.)
2003-02-12 23:08:22 +00:00
sizeof(buf));
PyErr_SetString(socket_error,
"illegal IP address string passed to inet_aton");
return NULL;
Patch 1471925 - Weak linking support for OSX This patch causes several symbols in the socket and posix module to be weakly linked on OSX and disables usage of ftime on OSX. These changes make it possible to use a binary build on OSX 10.4 on a 10.3 system.
2006-04-23 11:59:25 +00:00
#ifdef USE_INET_ATON_WEAKLINK
} else {
#endif
#endif
#if !defined(HAVE_INET_ATON) || defined(USE_INET_ATON_WEAKLINK)
patch [ 756021 ] Allow socket.inet_aton("255.255.255.255") on Windows
2005-08-26 08:34:00 +00:00
/* special-case this address as inet_addr might return INADDR_NONE
* for this */
if (strcmp(ip_addr, "255.255.255.255") == 0) {
packed_addr = 0xFFFFFFFF;
} else {
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
patch [ 756021 ] Allow socket.inet_aton("255.255.255.255") on Windows
2005-08-26 08:34:00 +00:00
packed_addr = inet_addr(ip_addr);
Port inet_ntoa and inet_aton to Windows: - fix unescaped newline in string literal - removed unused err variable - Windows doesn't have inet_aton; use inet_addr instead
1999-08-20 19:11:27 +00:00
patch [ 756021 ] Allow socket.inet_aton("255.255.255.255") on Windows
2005-08-26 08:34:00 +00:00
if (packed_addr == INADDR_NONE) { /* invalid address */
PyErr_SetString(socket_error,
"illegal IP address string passed to inet_aton");
return NULL;
}
Port inet_ntoa and inet_aton to Windows: - fix unescaped newline in string literal - removed unused err variable - Windows doesn't have inet_aton; use inet_addr instead
1999-08-20 19:11:27 +00:00
}
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromStringAndSize((char *) &packed_addr,
Port inet_ntoa and inet_aton to Windows: - fix unescaped newline in string literal - removed unused err variable - Windows doesn't have inet_aton; use inet_addr instead
1999-08-20 19:11:27 +00:00
sizeof(packed_addr));
Patch 1471925 - Weak linking support for OSX This patch causes several symbols in the socket and posix module to be weakly linked on OSX and disables usage of ftime on OSX. These changes make it possible to use a binary build on OSX 10.4 on a 10.3 system.
2006-04-23 11:59:25 +00:00
#ifdef USE_INET_ATON_WEAKLINK
}
#endif
Addressing SF bug #643005, implement socket.inet_aton() using inet_aton() rather than inet_addr() -- the latter is obsolete because it has a problem: "255.255.255.255" is a valid address but indistinguishable from an error. (I'm not sure if inet_aton() exists everywhere -- in case it doesn't, I've left the old code in with an #ifdef.)
2003-02-12 23:08:22 +00:00
#endif
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(inet_ntoa_doc,
Fixed inet_ntoa() docstring.
2000-03-07 14:05:16 +00:00
"inet_ntoa(packed_ip) -> ip_address_string\n\
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Convert an IP address from 32-bit packed binary format to string format");
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
static PyObject*
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_inet_ntoa(PyObject *self, PyObject *args)
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
{
char *packed_str;
Port inet_ntoa and inet_aton to Windows: - fix unescaped newline in string literal - removed unused err variable - Windows doesn't have inet_aton; use inet_addr instead
1999-08-20 19:11:27 +00:00
int addr_len;
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
struct in_addr packed_addr;
Fix all routines to use PyArg_ParseTuple(), and add ":name" to the argument format strings. THIS WILL PROBABLY BREAK LOTS OF CODE!!! Also fixed a bogus string in an error message in getsockaddrlen().
2000-03-24 20:56:56 +00:00
if (!PyArg_ParseTuple(args, "s#:inet_ntoa", &packed_str, &addr_len)) {
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
return NULL;
}
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
if (addr_len != sizeof(packed_addr)) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error,
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
"packed IP wrong length for inet_ntoa");
return NULL;
}
memcpy(&packed_addr, packed_str, addr_len);
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromString(inet_ntoa(packed_addr));
Patch by Ben Gertzfield to add inet_aton() and inet_ntoa() calls.
1999-08-20 18:21:51 +00:00
}
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
#ifdef HAVE_INET_PTON
PyDoc_STRVAR(inet_pton_doc,
"inet_pton(af, ip) -> packed IP address string\n\
\n\
Convert an IP address from string format to a packed string suitable\n\
for use with low-level network functions.");
static PyObject *
socket_inet_pton(PyObject *self, PyObject *args)
{
int af;
char* ip;
int retval;
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
#ifdef ENABLE_IPV6
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
char packed[MAX(sizeof(struct in_addr), sizeof(struct in6_addr))];
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
#else
char packed[sizeof(struct in_addr)];
#endif
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
if (!PyArg_ParseTuple(args, "is:inet_pton", &af, &ip)) {
return NULL;
}
Patch #924294: Do not check for AF_INET6 if it is not defined. Will backport to 2.3.
2004-06-02 12:35:29 +00:00
#if !defined(ENABLE_IPV6) && defined(AF_INET6)
Patch #781722: Reject AF_INET6 if IPv6 is disabled. Will backport to 2.3.
2003-08-05 06:25:06 +00:00
if(af == AF_INET6) {
PyErr_SetString(socket_error,
"can't use AF_INET6, IPv6 is disabled");
return NULL;
}
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
#endif
Patch #781722: Reject AF_INET6 if IPv6 is disabled. Will backport to 2.3.
2003-08-05 06:25:06 +00:00
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
retval = inet_pton(af, ip, packed);
if (retval < 0) {
PyErr_SetFromErrno(socket_error);
return NULL;
} else if (retval == 0) {
PyErr_SetString(socket_error,
"illegal IP address string passed to inet_pton");
return NULL;
} else if (af == AF_INET) {
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromStringAndSize(packed,
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
sizeof(struct in_addr));
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
#ifdef ENABLE_IPV6
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
} else if (af == AF_INET6) {
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromStringAndSize(packed,
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
sizeof(struct in6_addr));
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
#endif
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
} else {
PyErr_SetString(socket_error, "unknown address family");
return NULL;
}
}
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
PyDoc_STRVAR(inet_ntop_doc,
"inet_ntop(af, packed_ip) -> string formatted IP address\n\
\n\
Convert a packed IP address of the given family to string format.");
static PyObject *
socket_inet_ntop(PyObject *self, PyObject *args)
{
int af;
char* packed;
int len;
const char* retval;
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
#ifdef ENABLE_IPV6
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
char ip[MAX(INET_ADDRSTRLEN, INET6_ADDRSTRLEN) + 1];
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
#else
char ip[INET_ADDRSTRLEN + 1];
#endif
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
/* Guarantee NUL-termination for PyString_FromString() below */
Fix apparently trivial buffer overflow (SF bug 1060396). memset() wrote one past the end of the buffer, which was likely to be unused padding or a yet-to-be-initialized local variable. This routine is already tested by test_socket.
2004-11-07 14:24:25 +00:00
memset((void *) &ip[0], '\0', sizeof(ip));
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
if (!PyArg_ParseTuple(args, "is#:inet_ntop", &af, &packed, &len)) {
return NULL;
}
if (af == AF_INET) {
if (len != sizeof(struct in_addr)) {
PyErr_SetString(PyExc_ValueError,
"invalid length of packed IP address string");
return NULL;
}
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
#ifdef ENABLE_IPV6
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
} else if (af == AF_INET6) {
if (len != sizeof(struct in6_addr)) {
PyErr_SetString(PyExc_ValueError,
"invalid length of packed IP address string");
return NULL;
}
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
#endif
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
} else {
PyErr_Format(PyExc_ValueError,
"unknown address family %d", af);
return NULL;
}
retval = inet_ntop(af, packed, ip, sizeof(ip));
if (!retval) {
PyErr_SetFromErrno(socket_error);
return NULL;
} else {
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
return PyString_FromString(retval);
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
}
/* NOTREACHED */
PyErr_SetString(PyExc_RuntimeError, "invalid handling of inet_ntop");
return NULL;
}
#endif /* HAVE_INET_PTON */
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
/* Python interface to getaddrinfo(host, port). */
/*ARGSUSED*/
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_getaddrinfo(PyObject *self, PyObject *args)
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
{
Fix memory leaks detecting in bug report #478003.
2001-11-07 08:31:03 +00:00
struct addrinfo hints, *res;
struct addrinfo *res0 = NULL;
Implement IDNA (Internationalized Domain Names in Applications).
2003-04-18 10:39:54 +00:00
PyObject *hobj = NULL;
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
PyObject *pobj = (PyObject *)NULL;
Bump size of sprintf buffer. Suggested by Alex Coventry.
2001-08-12 09:28:40 +00:00
char pbuf[30];
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
char *hptr, *pptr;
int family, socktype, protocol, flags;
int error;
PyObject *all = (PyObject *)NULL;
PyObject *single = (PyObject *)NULL;
Implement IDNA (Internationalized Domain Names in Applications).
2003-04-18 10:39:54 +00:00
PyObject *idna = NULL;
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
family = socktype = protocol = flags = 0;
Patch #650422: Use Posix AF_ constants instead of PF_ ones.
2002-12-11 13:10:57 +00:00
family = AF_UNSPEC;
Implement IDNA (Internationalized Domain Names in Applications).
2003-04-18 10:39:54 +00:00
if (!PyArg_ParseTuple(args, "OO|iiii:getaddrinfo",
&hobj, &pobj, &family, &socktype,
&protocol, &flags)) {
return NULL;
}
if (hobj == Py_None) {
hptr = NULL;
} else if (PyUnicode_Check(hobj)) {
idna = PyObject_CallMethod(hobj, "encode", "s", "idna");
if (!idna)
return NULL;
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
hptr = PyString_AsString(idna);
} else if (PyString_Check(hobj)) {
hptr = PyString_AsString(hobj);
Implement IDNA (Internationalized Domain Names in Applications).
2003-04-18 10:39:54 +00:00
} else {
Support for buffer protocol for socket and struct. * Added socket.recv_buf() and socket.recvfrom_buf() methods, that use the buffer protocol (send and sendto already did). * Added struct.pack_to(), that is the corresponding buffer compatible method to unpack_from(). * Fixed minor typos in arraymodule.
2006-05-26 12:03:27 +00:00
PyErr_SetString(PyExc_TypeError,
Implement IDNA (Internationalized Domain Names in Applications).
2003-04-18 10:39:54 +00:00
"getaddrinfo() argument 1 must be string or None");
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return NULL;
}
if (PyInt_Check(pobj)) {
Enable PyOS_snprintf() et al. during alpha phase of 2.2.0 and add another use case to the socketmodule.
2001-07-31 18:05:33 +00:00
PyOS_snprintf(pbuf, sizeof(pbuf), "%ld", PyInt_AsLong(pobj));
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
pptr = pbuf;
This reverts r63675 based on the discussion in this thread: http://mail.python.org/pipermail/python-dev/2008-June/079988.html Python 2.6 should stick with PyString_* in its codebase. The PyBytes_* names in the spirit of 3.0 are available via a #define only. See the email thread.
2008-06-09 04:58:54 +00:00
} else if (PyString_Check(pobj)) {
pptr = PyString_AsString(pobj);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
} else if (pobj == Py_None) {
pptr = (char *)NULL;
} else {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error, "Int or String expected");
Release host name memory. Fixes #783312. Will backport to 2.3.
2003-08-07 11:55:15 +00:00
goto err;
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
}
memset(&hints, 0, sizeof(hints));
hints.ai_family = family;
hints.ai_socktype = socktype;
hints.ai_protocol = protocol;
hints.ai_flags = flags;
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
Py_BEGIN_ALLOW_THREADS
ACQUIRE_GETADDRINFO_LOCK
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
error = getaddrinfo(hptr, pptr, &hints, &res0);
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
Py_END_ALLOW_THREADS
RELEASE_GETADDRINFO_LOCK /* see comment in setipaddr() */
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
if (error) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
set_gaierror(error);
Release host name memory. Fixes #783312. Will backport to 2.3.
2003-08-07 11:55:15 +00:00
goto err;
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
}
if ((all = PyList_New(0)) == NULL)
goto err;
for (res = res0; res; res = res->ai_next) {
PySocket_getaddrinfo(): fix two refcount bugs, both having to do with a misunderstanding of the refcont behavior of the 'O' format code in PyArg_ParseTuple() and Py_BuildValue(), respectively. - pobj is only a borrowed reference, so should *not* be DECREF'ed at the end. This was the cause of SF bug #470635. - The Py_BuildValue() call would leak the object produced by makesockaddr(). (I found this by eyeballing the code.)
2001-10-12 18:59:27 +00:00
PyObject *addr =
Patch #929192: Improvements to bluetooth support. Added setbdaddr and makebdaddr. Extended makesockaddr to understand Bluetooth addresses. Changed getsockaddr to expect the Bluetooth addresses as a string, not a six element tuple. Reformatted some of the Bluetooth code to be more consistent with PEP 7.
2004-06-03 09:24:42 +00:00
makesockaddr(-1, res->ai_addr, res->ai_addrlen, protocol);
PySocket_getaddrinfo(): fix two refcount bugs, both having to do with a misunderstanding of the refcont behavior of the 'O' format code in PyArg_ParseTuple() and Py_BuildValue(), respectively. - pobj is only a borrowed reference, so should *not* be DECREF'ed at the end. This was the cause of SF bug #470635. - The Py_BuildValue() call would leak the object produced by makesockaddr(). (I found this by eyeballing the code.)
2001-10-12 18:59:27 +00:00
if (addr == NULL)
goto err;
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
single = Py_BuildValue("iiisO", res->ai_family,
res->ai_socktype, res->ai_protocol,
res->ai_canonname ? res->ai_canonname : "",
PySocket_getaddrinfo(): fix two refcount bugs, both having to do with a misunderstanding of the refcont behavior of the 'O' format code in PyArg_ParseTuple() and Py_BuildValue(), respectively. - pobj is only a borrowed reference, so should *not* be DECREF'ed at the end. This was the cause of SF bug #470635. - The Py_BuildValue() call would leak the object produced by makesockaddr(). (I found this by eyeballing the code.)
2001-10-12 18:59:27 +00:00
addr);
Py_DECREF(addr);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
if (single == NULL)
goto err;
if (PyList_Append(all, single))
goto err;
Py_XDECREF(single);
}
Implement IDNA (Internationalized Domain Names in Applications).
2003-04-18 10:39:54 +00:00
Py_XDECREF(idna);
SF bug #592645 fix memory leak in socket.getaddrinfo
2002-08-09 03:37:42 +00:00
if (res0)
freeaddrinfo(res0);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return all;
err:
Py_XDECREF(single);
Py_XDECREF(all);
Implement IDNA (Internationalized Domain Names in Applications).
2003-04-18 10:39:54 +00:00
Py_XDECREF(idna);
Fix memory leaks detecting in bug report #478003.
2001-11-07 08:31:03 +00:00
if (res0)
freeaddrinfo(res0);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return (PyObject *)NULL;
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(getaddrinfo_doc,
Clean up some docstrings. Some docstrings didn't show their return value; others were inconsistent in what to name the argument or return value; a few module-global functions had "socket." in front of their name, against convention.
2002-08-08 20:37:08 +00:00
"getaddrinfo(host, port [, family, socktype, proto, flags])\n\
-> list of (family, socktype, proto, canonname, sockaddr)\n\
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Resolve host and port into addrinfo struct.");
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
/* Python interface to getnameinfo(sa, flags). */
/*ARGSUSED*/
static PyObject *
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_getnameinfo(PyObject *self, PyObject *args)
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
{
PyObject *sa = (PyObject *)NULL;
int flags;
char *hostp;
Correct getnameinfo refcounting and tuple parsing. Fixes #476648.
2001-11-02 23:34:52 +00:00
int port, flowinfo, scope_id;
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
char hbuf[NI_MAXHOST], pbuf[NI_MAXSERV];
struct addrinfo hints, *res = NULL;
int error;
PyObject *ret = (PyObject *)NULL;
flags = flowinfo = scope_id = 0;
Correct getnameinfo refcounting and tuple parsing. Fixes #476648.
2001-11-02 23:34:52 +00:00
if (!PyArg_ParseTuple(args, "Oi:getnameinfo", &sa, &flags))
return NULL;
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
if (!PyArg_ParseTuple(sa, "si|ii",
&hostp, &port, &flowinfo, &scope_id))
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return NULL;
Enable PyOS_snprintf() et al. during alpha phase of 2.2.0 and add another use case to the socketmodule.
2001-07-31 18:05:33 +00:00
PyOS_snprintf(pbuf, sizeof(pbuf), "%d", port);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
memset(&hints, 0, sizeof(hints));
Patch #650422: Use Posix AF_ constants instead of PF_ ones.
2002-12-11 13:10:57 +00:00
hints.ai_family = AF_UNSPEC;
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
hints.ai_socktype = SOCK_DGRAM; /* make numeric port happy */
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
Py_BEGIN_ALLOW_THREADS
ACQUIRE_GETADDRINFO_LOCK
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
error = getaddrinfo(hostp, pbuf, &hints, &res);
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
Py_END_ALLOW_THREADS
RELEASE_GETADDRINFO_LOCK /* see comment in setipaddr() */
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
if (error) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
set_gaierror(error);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
goto fail;
}
if (res->ai_next) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error,
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
"sockaddr resolved to multiple addresses");
goto fail;
}
switch (res->ai_family) {
case AF_INET:
{
char *t1;
int t2;
if (PyArg_ParseTuple(sa, "si", &t1, &t2) == 0) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
PyErr_SetString(socket_error,
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
"IPv4 sockaddr must be 2 tuple");
goto fail;
}
break;
}
Remove INET6 define. Use ENABLE_IPV6 instead.
2001-12-02 10:15:37 +00:00
#ifdef ENABLE_IPV6
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
case AF_INET6:
{
struct sockaddr_in6 *sin6;
sin6 = (struct sockaddr_in6 *)res->ai_addr;
sin6->sin6_flowinfo = flowinfo;
sin6->sin6_scope_id = scope_id;
break;
}
#endif
}
error = getnameinfo(res->ai_addr, res->ai_addrlen,
hbuf, sizeof(hbuf), pbuf, sizeof(pbuf), flags);
if (error) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
set_gaierror(error);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
goto fail;
}
ret = Py_BuildValue("ss", hbuf, pbuf);
fail:
if (res)
freeaddrinfo(res);
return ret;
}
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(getnameinfo_doc,
Clean up some docstrings. Some docstrings didn't show their return value; others were inconsistent in what to name the argument or return value; a few module-global functions had "socket." in front of their name, against convention.
2002-08-08 20:37:08 +00:00
"getnameinfo(sockaddr, flags) --> (host, port)\n\
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
\n\
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
Get host and port for a sockaddr.");
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
Add default timeout functionality. This adds setdefaulttimeout() and getdefaulttimeout() functions to the socket and _socket modules, and appropriate tests.
2002-07-18 17:08:35 +00:00
/* Python API to getting and setting the default timeout value. */
static PyObject *
socket_getdefaulttimeout(PyObject *self)
{
if (defaulttimeout < 0.0) {
Py_INCREF(Py_None);
return Py_None;
}
else
return PyFloat_FromDouble(defaulttimeout);
}
PyDoc_STRVAR(getdefaulttimeout_doc,
Clean up some docstrings. Some docstrings didn't show their return value; others were inconsistent in what to name the argument or return value; a few module-global functions had "socket." in front of their name, against convention.
2002-08-08 20:37:08 +00:00
"getdefaulttimeout() -> timeout\n\
Add default timeout functionality. This adds setdefaulttimeout() and getdefaulttimeout() functions to the socket and _socket modules, and appropriate tests.
2002-07-18 17:08:35 +00:00
\n\
Returns the default timeout in floating seconds for new socket objects.\n\
A value of None indicates that new socket objects have no timeout.\n\
When the socket module is first imported, the default is None.");
static PyObject *
socket_setdefaulttimeout(PyObject *self, PyObject *arg)
{
double timeout;
if (arg == Py_None)
timeout = -1.0;
else {
timeout = PyFloat_AsDouble(arg);
if (timeout < 0.0) {
if (!PyErr_Occurred())
PyErr_SetString(PyExc_ValueError,
"Timeout value out of range");
return NULL;
}
}
defaulttimeout = timeout;
Py_INCREF(Py_None);
return Py_None;
}
PyDoc_STRVAR(setdefaulttimeout_doc,
Clean up some docstrings. Some docstrings didn't show their return value; others were inconsistent in what to name the argument or return value; a few module-global functions had "socket." in front of their name, against convention.
2002-08-08 20:37:08 +00:00
"setdefaulttimeout(timeout)\n\
Add default timeout functionality. This adds setdefaulttimeout() and getdefaulttimeout() functions to the socket and _socket modules, and appropriate tests.
2002-07-18 17:08:35 +00:00
\n\
Set the default timeout in floating seconds for new socket objects.\n\
A value of None indicates that new socket objects have no timeout.\n\
When the socket module is first imported, the default is None.");
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
/* List of functions exported by this module. */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
static PyMethodDef socket_methods[] = {
{"gethostbyname", socket_gethostbyname,
Use METH_VARARGS instead of numeric constant 1 in method def. tables
2000-08-03 02:06:16 +00:00
METH_VARARGS, gethostbyname_doc},
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"gethostbyname_ex", socket_gethostbyname_ex,
Use METH_VARARGS instead of numeric constant 1 in method def. tables
2000-08-03 02:06:16 +00:00
METH_VARARGS, ghbn_ex_doc},
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"gethostbyaddr", socket_gethostbyaddr,
Use METH_VARARGS instead of numeric constant 1 in method def. tables
2000-08-03 02:06:16 +00:00
METH_VARARGS, gethostbyaddr_doc},
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"gethostname", socket_gethostname,
Make use of METH_O and METH_NOARGS where possible. Use Py_UnpackTuple instead of PyArg_ParseTuple where possible.
2006-05-29 21:04:52 +00:00
METH_NOARGS, gethostname_doc},
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"getservbyname", socket_getservbyname,
Use METH_VARARGS instead of numeric constant 1 in method def. tables
2000-08-03 02:06:16 +00:00
METH_VARARGS, getservbyname_doc},
Added socket.getservbyport(), and make its second argument and that of getservbyname() optional. Update the tests and the docs.
2004-06-28 00:50:43 +00:00
{"getservbyport", socket_getservbyport,
METH_VARARGS, getservbyport_doc},
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"getprotobyname", socket_getprotobyname,
Make use of METH_O and METH_NOARGS where possible. Use Py_UnpackTuple instead of PyArg_ParseTuple where possible.
2006-05-29 21:04:52 +00:00
METH_VARARGS, getprotobyname_doc},
Changes to make it work and compile on NT, by Mark Hammond. (NT changes for posixmodule.c re-invented by Guido.)
1994-08-18 15:42:46 +00:00
#ifndef NO_DUP
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"fromfd", socket_fromfd,
Use METH_VARARGS instead of numeric constant 1 in method def. tables
2000-08-03 02:06:16 +00:00
METH_VARARGS, fromfd_doc},
Patch #1003700: Add socketpair function to socket module.
2004-08-09 04:51:41 +00:00
#endif
#ifdef HAVE_SOCKETPAIR
{"socketpair", socket_socketpair,
METH_VARARGS, socketpair_doc},
Changes to make it work and compile on NT, by Mark Hammond. (NT changes for posixmodule.c re-invented by Guido.)
1994-08-18 15:42:46 +00:00
#endif
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"ntohs", socket_ntohs,
Use METH_VARARGS instead of numeric constant 1 in method def. tables
2000-08-03 02:06:16 +00:00
METH_VARARGS, ntohs_doc},
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"ntohl", socket_ntohl,
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
METH_O, ntohl_doc},
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"htons", socket_htons,
Use METH_VARARGS instead of numeric constant 1 in method def. tables
2000-08-03 02:06:16 +00:00
METH_VARARGS, htons_doc},
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"htonl", socket_htonl,
Extended socket.htonl and ntohl to accept longs. Fixes SF bug #568322. The code should raise an OverflowError if the long is > 32 bits, even on platforms where sizeof(long) > 4.
2002-07-25 16:01:12 +00:00
METH_O, htonl_doc},
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"inet_aton", socket_inet_aton,
Use METH_VARARGS instead of numeric constant 1 in method def. tables
2000-08-03 02:06:16 +00:00
METH_VARARGS, inet_aton_doc},
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"inet_ntoa", socket_inet_ntoa,
Use METH_VARARGS instead of numeric constant 1 in method def. tables
2000-08-03 02:06:16 +00:00
METH_VARARGS, inet_ntoa_doc},
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
#ifdef HAVE_INET_PTON
{"inet_pton", socket_inet_pton,
METH_VARARGS, inet_pton_doc},
{"inet_ntop", socket_inet_ntop,
METH_VARARGS, inet_ntop_doc},
#endif
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"getaddrinfo", socket_getaddrinfo,
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
METH_VARARGS, getaddrinfo_doc},
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
{"getnameinfo", socket_getnameinfo,
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
METH_VARARGS, getnameinfo_doc},
Silence warning about getdefaulttimeout in PyMethodDef.
2002-07-18 21:11:26 +00:00
{"getdefaulttimeout", (PyCFunction)socket_getdefaulttimeout,
Add default timeout functionality. This adds setdefaulttimeout() and getdefaulttimeout() functions to the socket and _socket modules, and appropriate tests.
2002-07-18 17:08:35 +00:00
METH_NOARGS, getdefaulttimeout_doc},
{"setdefaulttimeout", socket_setdefaulttimeout,
METH_O, setdefaulttimeout_doc},
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
{NULL, NULL} /* Sentinel */
Initial revision
1991-06-25 21:36:08 +00:00
};
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
#ifdef RISCOS
#define OS_INIT_DEFINED
static int
os_init(void)
{
_kernel_swi_regs r;
r.r[0] = 0;
_kernel_swi(0x43380, &r, &r);
taskwindow = r.r[0];
Patch #734231: Update RiscOS support. In particular, correct riscospath.extsep, and use os.extsep throughout.
2003-05-10 07:36:56 +00:00
return 1;
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
}
#endif /* RISCOS */
changed windows test symbols; removed debug printf from sockets
1996-06-26 18:22:49 +00:00
#ifdef MS_WINDOWS
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
#define OS_INIT_DEFINED
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
/* Additional initialization and cleanup for Windows */
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
static void
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
os_cleanup(void)
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
{
WSACleanup();
}
static int
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
os_init(void)
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
{
WSADATA WSAData;
int ret;
char buf[100];
ret = WSAStartup(0x0101, &WSAData);
switch (ret) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
case 0: /* No error */
Fix [ 766669 ] Consistent GPF on exit Use Py_AtExit instead of atexit so we are called during Py_Finalize() rather than during DLL teardown.
2003-07-09 04:57:46 +00:00
Py_AtExit(os_cleanup);
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
return 1; /* Success */
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
case WSASYSNOTREADY:
PyErr_SetString(PyExc_ImportError,
"WSAStartup failed: network not ready");
break;
case WSAVERNOTSUPPORTED:
case WSAEINVAL:
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
PyErr_SetString(
PyExc_ImportError,
"WSAStartup failed: requested version not supported");
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
break;
default:
sprintf -> PyOS_snprintf in some "obviously safe" cases. Also changed <>-style #includes to ""-style in some places where the former didn't make sense.
2001-11-28 20:27:42 +00:00
PyOS_snprintf(buf, sizeof(buf),
"WSAStartup failed: error code %d", ret);
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
PyErr_SetString(PyExc_ImportError, buf);
break;
}
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
return 0; /* Failure */
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
}
changed windows test symbols; removed debug printf from sockets
1996-06-26 18:22:49 +00:00
#endif /* MS_WINDOWS */
More rational support for Windows (tested on Roj's NT 3.5 machine only). Use PyArg_ParseTuple() where it makes sense.
1996-06-20 16:25:29 +00:00
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
#ifdef PYOS_OS2
#define OS_INIT_DEFINED
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
/* Additional initialization for OS/2 */
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
static int
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
os_init(void)
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
{
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
#ifndef PYCC_GCC
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
char reason[64];
int rc = sock_init();
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
if (rc == 0) {
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
return 1; /* Success */
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
}
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
PyOS_snprintf(reason, sizeof(reason),
"OS/2 TCP/IP Error# %d", sock_errno());
PyErr_SetString(PyExc_ImportError, reason);
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
return 0; /* Failure */
OS/2 EMX port changes (Modules part of patch #450267): Modules/ socketmodule.c EMX handles sockets like Posix, rather than use native APIs
2002-03-03 03:03:52 +00:00
#else
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
/* No need to initialise sockets with GCC/EMX */
return 1; /* Success */
OS/2 EMX port changes (Modules part of patch #450267): Modules/ socketmodule.c EMX handles sockets like Posix, rather than use native APIs
2002-03-03 03:03:52 +00:00
#endif
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
}
#endif /* PYOS_OS2 */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
#ifndef OS_INIT_DEFINED
static int
os_init(void)
{
return 1; /* Success */
}
#endif
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
/* C API table - always add new things to the end for binary
compatibility. */
static
PySocketModule_APIObject PySocketModuleAPI =
{
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
&sock_type,
Make socket.sslerror a subclass of socket.error . Added socket.error to the socket module's C API.
2004-03-23 23:16:54 +00:00
NULL
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
};
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
/* Initialize the _socket module.
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
This module is actually called "_socket", and there's a wrapper
"socket.py" which implements some additional functionality. On some
platforms (e.g. Windows and OS/2), socket.py also implements a
wrapper for the socket type that provides missing functionality such
as makefile(), dup() and fromfd(). The import of "_socket" may fail
with an ImportError exception if os-specific initialization fails.
On Windows, this does WINSOCK initialization. When WINSOCK is
initialized succesfully, a call to WSACleanup() is scheduled to be
made at exit time.
*/
Many changes to the interface, and added comments.
1991-06-27 15:51:29 +00:00
Patch #568124: Add doc string macros.
2002-06-13 20:33:02 +00:00
PyDoc_STRVAR(socket_doc,
Clean up some docstrings. Some docstrings didn't show their return value; others were inconsistent in what to name the argument or return value; a few module-global functions had "socket." in front of their name, against convention.
2002-08-08 20:37:08 +00:00
"Implementation module for socket operations.\n\
\n\
See the socket module for documentation.");
Added copious docstrings, plus two minor layout tweaks.
1998-07-07 20:45:43 +00:00
Excise DL_EXPORT/DL_IMPORT from Modules/*. Required adding a prototype for Py_Main(). Thanks to Kalle Svensson and Skip Montanaro for the patches.
2002-08-02 02:27:13 +00:00
PyMODINIT_FUNC
Create a new section of pyport.h to hold all external function declarations for systems that are missing those declarations from system include files. Start by moving a pointy-haired ones from their previous locations to the new section. (The gethostname() one, for instance, breaks on several systems, because some define it as (char *, size_t) and some as (char *, int).) I purposely decided not to include the summary of used #defines like Tim did in the first section of pyport.h. In my opinion, the number of #defines likedly to be used by this section would make such an overview unwieldy. I would suggest documenting the non-obvious ones, though.
2000-07-24 16:06:23 +00:00
init_socket(void)
Initial revision
1991-06-25 21:36:08 +00:00
{
Removed unused variable
2008-01-11 15:42:29 +00:00
PyObject *m, *has_ipv6;
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
if (!os_init())
os2 patch by Jeff Rush
1997-11-22 21:53:48 +00:00
return;
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
#1629: Renamed Py_Size, Py_Type and Py_Refcnt to Py_SIZE, Py_TYPE and Py_REFCNT. Macros for b/w compatibility are available.
2007-12-19 02:37:44 +00:00
Py_TYPE(&sock_type) = &PyType_Type;
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
m = Py_InitModule3(PySocket_MODULE_NAME,
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
socket_methods,
socket_doc);
Check return result from Py_InitModule*(). This API can fail. Probably should be backported.
2006-01-19 06:09:39 +00:00
if (m == NULL)
return;
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
Change socket.error to inherit from IOError rather than being a stand alone class. This addresses the primary concern in http://bugs.python.org/issue1706815 python-dev discussion here: http://mail.python.org/pipermail/python-dev/2007-July/073749.html I chose IOError rather than EnvironmentError as the base class since socket objects are often used as transparent duck typed file objects in code already prepared to deal with IOError exceptions. also a minor fix: urllib2 - fix a couple places where IOError was raised rather than URLError. for better or worse, URLError already inherits from IOError so this won't break any existing code. test_urllib2net - replace bad ftp urls.
2007-09-09 23:36:46 +00:00
socket_error = PyErr_NewException("socket.error",
PyExc_IOError, NULL);
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
if (socket_error == NULL)
Apply two changes, systematically: (1) Use PyErr_NewException("module.class", NULL, NULL) to create the exception object. (2) Remove all calls to Py_FatalError(); instead, return or ignore the errors -- the import code now checks PyErr_Occurred() after calling a module's init function, so it's no longer a fatal error for the initialization to fail. Also did some small cleanups, e.g. removed unnecessary test for "already initialized" from initfpectl(), and unified initposix()/initnt(). I haven't checked this very thoroughly, so while the changes are pretty trivial -- beware of untested code!
1997-10-01 04:29:29 +00:00
return;
Make socket.sslerror a subclass of socket.error . Added socket.error to the socket module's C API.
2004-03-23 23:16:54 +00:00
PySocketModuleAPI.error = socket_error;
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
Py_INCREF(socket_error);
PyModule_AddObject(m, "error", socket_error);
socket_herror = PyErr_NewException("socket.herror",
socket_error, NULL);
if (socket_herror == NULL)
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return;
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
Py_INCREF(socket_herror);
PyModule_AddObject(m, "herror", socket_herror);
socket_gaierror = PyErr_NewException("socket.gaierror", socket_error,
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
NULL);
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
if (socket_gaierror == NULL)
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
return;
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
Py_INCREF(socket_gaierror);
PyModule_AddObject(m, "gaierror", socket_gaierror);
SF patch #760257: add socket.timeout exception (Contributed by Bob Halley) Added a new exception, socket.timeout so that timeouts can be differentiated from other socket exceptions. Docs, more tests, and newsitem to follow.
2003-06-29 03:08:05 +00:00
socket_timeout = PyErr_NewException("socket.timeout",
socket_error, NULL);
if (socket_timeout == NULL)
return;
Py_INCREF(socket_timeout);
PyModule_AddObject(m, "timeout", socket_timeout);
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
Py_INCREF((PyObject *)&sock_type);
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
if (PyModule_AddObject(m, "SocketType",
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
(PyObject *)&sock_type) != 0)
Apply two changes, systematically: (1) Use PyErr_NewException("module.class", NULL, NULL) to create the exception object. (2) Remove all calls to Py_FatalError(); instead, return or ignore the errors -- the import code now checks PyErr_Occurred() after calling a module's init function, so it's no longer a fatal error for the initialization to fail. Also did some small cleanups, e.g. removed unnecessary test for "already initialized" from initfpectl(), and unified initposix()/initnt(). I haven't checked this very thoroughly, so while the changes are pretty trivial -- beware of untested code!
1997-10-01 04:29:29 +00:00
return;
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
Py_INCREF((PyObject *)&sock_type);
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
if (PyModule_AddObject(m, "socket",
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
(PyObject *)&sock_type) != 0)
Made SocketType and socket the same thing: a subclassable type whose constructor acts just like socket() before. All three arguments have a sensible default now; socket() is equivalent to socket(AF_INET, SOCK_STREAM). One minor issue: the socket() function and the SocketType had different doc strings; socket.__doc__ gave the signature, SocketType.__doc__ gave the methods. I've merged these for now, but maybe the list of methods is no longer necessary since it can easily be recovered through socket.__dict__.keys(). The problem with keeping it is that the total doc string is a bit long (34 lines -- it scrolls of a standard tty screen). Another general issue with the socket module is that it's a big mess. There's pages and pages of random platform #ifdefs, and the naming conventions are totally wrong: it uses Py prefixes and CapWords for static functions. That's a cleanup for another day... (Also I think the big starting comment that summarizes the API can go -- it's a repeat of the docstring.)
2001-10-27 22:20:47 +00:00
return;
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
#ifdef ENABLE_IPV6
has_ipv6 = Py_True;
#else
has_ipv6 = Py_False;
#endif
Py_INCREF(has_ipv6);
PyModule_AddObject(m, "has_ipv6", has_ipv6);
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
/* Export C API */
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
if (PyModule_AddObject(m, PySocket_CAPI_NAME,
Break SSL support out of _socket module and place it into a new helper module _ssl. The support for the RAND_* APIs in _ssl is now only enabled for OpenSSL 0.9.5 and up since they were added in that release. Note that socketmodule.* should really be renamed to _socket.* -- unfortunately, this seems to lose the CVS history of the file. Please review and test... I was only able to test the header file chaos in socketmodule.c/h on Linux. The test run through fine and compiles don't give errors or warnings. WARNING: This patch does *not* include changes to the various non-Unix build process files.
2002-02-16 18:23:30 +00:00
PyCObject_FromVoidPtr((void *)&PySocketModuleAPI, NULL)
) != 0)
return;
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
/* Address families (we only support AF_INET and AF_UNIX) */
#ifdef AF_UNSPEC
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AF_UNSPEC", AF_UNSPEC);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AF_INET", AF_INET);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#ifdef AF_INET6
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AF_INET6", AF_INET6);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif /* AF_INET6 */
OS/2's TCP/IP stack supports AF_UNIX sockets, with the limitation that the socket name must start with /socket32/. Unlike Unix systems, this file never exists in the file system.
2004-04-04 07:13:49 +00:00
#if defined(AF_UNIX)
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AF_UNIX", AF_UNIX);
Merge alpha100 branch back to main trunk
1994-08-01 11:34:53 +00:00
#endif /* AF_UNIX */
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#ifdef AF_AX25
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
/* Amateur Radio AX.25 */
PyModule_AddIntConstant(m, "AF_AX25", AF_AX25);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef AF_IPX
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AF_IPX", AF_IPX); /* Novell IPX */
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef AF_APPLETALK
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
/* Appletalk DDP */
PyModule_AddIntConstant(m, "AF_APPLETALK", AF_APPLETALK);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef AF_NETROM
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
/* Amateur radio NetROM */
PyModule_AddIntConstant(m, "AF_NETROM", AF_NETROM);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef AF_BRIDGE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
/* Multiprotocol bridge */
PyModule_AddIntConstant(m, "AF_BRIDGE", AF_BRIDGE);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
Patch #984654: Add more address family constants.
2004-07-19 17:01:20 +00:00
#ifdef AF_ATMPVC
/* ATM PVCs */
PyModule_AddIntConstant(m, "AF_ATMPVC", AF_ATMPVC);
#endif
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#ifdef AF_AAL5
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
/* Reserved for Werner's ATM */
PyModule_AddIntConstant(m, "AF_AAL5", AF_AAL5);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef AF_X25
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
/* Reserved for X.25 project */
PyModule_AddIntConstant(m, "AF_X25", AF_X25);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef AF_INET6
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AF_INET6", AF_INET6); /* IP version 6 */
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef AF_ROSE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
/* Amateur Radio X.25 PLP */
PyModule_AddIntConstant(m, "AF_ROSE", AF_ROSE);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
Patch #984654: Add more address family constants.
2004-07-19 17:01:20 +00:00
#ifdef AF_DECnet
/* Reserved for DECnet project */
PyModule_AddIntConstant(m, "AF_DECnet", AF_DECnet);
#endif
#ifdef AF_NETBEUI
/* Reserved for 802.2LLC project */
PyModule_AddIntConstant(m, "AF_NETBEUI", AF_NETBEUI);
#endif
#ifdef AF_SECURITY
/* Security callback pseudo AF */
PyModule_AddIntConstant(m, "AF_SECURITY", AF_SECURITY);
#endif
#ifdef AF_KEY
/* PF_KEY key management API */
PyModule_AddIntConstant(m, "AF_KEY", AF_KEY);
#endif
#ifdef AF_NETLINK
/* */
PyModule_AddIntConstant(m, "AF_NETLINK", AF_NETLINK);
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
PyModule_AddIntConstant(m, "NETLINK_ROUTE", NETLINK_ROUTE);
Patch #1462080: Conditionalize some NETLINK defines.
2006-04-06 22:29:33 +00:00
#ifdef NETLINK_SKIP
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
PyModule_AddIntConstant(m, "NETLINK_SKIP", NETLINK_SKIP);
Patch #1462080: Conditionalize some NETLINK defines.
2006-04-06 22:29:33 +00:00
#endif
#ifdef NETLINK_W1
PyModule_AddIntConstant(m, "NETLINK_W1", NETLINK_W1);
#endif
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
PyModule_AddIntConstant(m, "NETLINK_USERSOCK", NETLINK_USERSOCK);
PyModule_AddIntConstant(m, "NETLINK_FIREWALL", NETLINK_FIREWALL);
NETLINK_TCPDIAG and NETLINK_NFLOG aren't defined on older Linux systems; define these conditionally.
2006-02-21 01:07:27 +00:00
#ifdef NETLINK_TCPDIAG
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
PyModule_AddIntConstant(m, "NETLINK_TCPDIAG", NETLINK_TCPDIAG);
NETLINK_TCPDIAG and NETLINK_NFLOG aren't defined on older Linux systems; define these conditionally.
2006-02-21 01:07:27 +00:00
#endif
#ifdef NETLINK_NFLOG
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
PyModule_AddIntConstant(m, "NETLINK_NFLOG", NETLINK_NFLOG);
NETLINK_TCPDIAG and NETLINK_NFLOG aren't defined on older Linux systems; define these conditionally.
2006-02-21 01:07:27 +00:00
#endif
Get socketmodule to build after adding netlink support.
2006-01-16 04:31:40 +00:00
#ifdef NETLINK_XFRM
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
PyModule_AddIntConstant(m, "NETLINK_XFRM", NETLINK_XFRM);
Get socketmodule to build after adding netlink support.
2006-01-16 04:31:40 +00:00
#endif
Patch #1462080: Conditionalize some NETLINK defines.
2006-04-06 22:29:33 +00:00
#ifdef NETLINK_ARPD
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
PyModule_AddIntConstant(m, "NETLINK_ARPD", NETLINK_ARPD);
Patch #1462080: Conditionalize some NETLINK defines.
2006-04-06 22:29:33 +00:00
#endif
#ifdef NETLINK_ROUTE6
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
PyModule_AddIntConstant(m, "NETLINK_ROUTE6", NETLINK_ROUTE6);
Patch #1462080: Conditionalize some NETLINK defines.
2006-04-06 22:29:33 +00:00
#endif
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
PyModule_AddIntConstant(m, "NETLINK_IP6_FW", NETLINK_IP6_FW);
Patch #1657276: Make NETLINK_DNRTMSG conditional. Will backport.
2007-02-13 12:14:19 +00:00
#ifdef NETLINK_DNRTMSG
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
PyModule_AddIntConstant(m, "NETLINK_DNRTMSG", NETLINK_DNRTMSG);
Patch #1657276: Make NETLINK_DNRTMSG conditional. Will backport.
2007-02-13 12:14:19 +00:00
#endif
Patch #1462080: Conditionalize some NETLINK defines.
2006-04-06 22:29:33 +00:00
#ifdef NETLINK_TAPBASE
Patch #1103116: AF_NETLINK sockets basic support.
2006-01-14 18:12:57 +00:00
PyModule_AddIntConstant(m, "NETLINK_TAPBASE", NETLINK_TAPBASE);
Patch #1462080: Conditionalize some NETLINK defines.
2006-04-06 22:29:33 +00:00
#endif
Bug #1460564: document that socket.fromfd() duplicates the given file descriptor.
2006-04-01 07:33:08 +00:00
#endif /* AF_NETLINK */
Patch #984654: Add more address family constants.
2004-07-19 17:01:20 +00:00
#ifdef AF_ROUTE
/* Alias to emulate 4.4BSD */
PyModule_AddIntConstant(m, "AF_ROUTE", AF_ROUTE);
#endif
#ifdef AF_ASH
/* Ash */
PyModule_AddIntConstant(m, "AF_ASH", AF_ASH);
#endif
#ifdef AF_ECONET
/* Acorn Econet */
PyModule_AddIntConstant(m, "AF_ECONET", AF_ECONET);
#endif
#ifdef AF_ATMSVC
/* ATM SVCs */
PyModule_AddIntConstant(m, "AF_ATMSVC", AF_ATMSVC);
#endif
#ifdef AF_SNA
/* Linux SNA Project (nutters!) */
PyModule_AddIntConstant(m, "AF_SNA", AF_SNA);
#endif
#ifdef AF_IRDA
/* IRDA sockets */
PyModule_AddIntConstant(m, "AF_IRDA", AF_IRDA);
#endif
#ifdef AF_PPPOX
/* PPPoX sockets */
PyModule_AddIntConstant(m, "AF_PPPOX", AF_PPPOX);
#endif
#ifdef AF_WANPIPE
/* Wanpipe API Sockets */
PyModule_AddIntConstant(m, "AF_WANPIPE", AF_WANPIPE);
#endif
#ifdef AF_LLC
/* Linux LLC */
PyModule_AddIntConstant(m, "AF_LLC", AF_LLC);
#endif
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
Add FreeBSD support for bluetooth sockets. (SF Patch #888148, reviewed by loewis)
2004-02-02 06:05:24 +00:00
#ifdef USE_BLUETOOTH
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
PyModule_AddIntConstant(m, "AF_BLUETOOTH", AF_BLUETOOTH);
PyModule_AddIntConstant(m, "BTPROTO_L2CAP", BTPROTO_L2CAP);
Patch #1432399: Add HCI sockets.
2007-02-14 10:07:37 +00:00
PyModule_AddIntConstant(m, "BTPROTO_HCI", BTPROTO_HCI);
PyModule_AddIntConstant(m, "SOL_HCI", SOL_HCI);
PyModule_AddIntConstant(m, "HCI_FILTER", HCI_FILTER);
Add FreeBSD support for bluetooth sockets. (SF Patch #888148, reviewed by loewis)
2004-02-02 06:05:24 +00:00
#if !defined(__FreeBSD__)
Fix build on FreeBSD. Bluetooth HCI API in FreeBSD is quite different from Linux's. Just fix the build for now but the code doesn't support the complete capability of HCI on FreeBSD yet.
2007-06-05 18:16:52 +00:00
PyModule_AddIntConstant(m, "HCI_TIME_STAMP", HCI_TIME_STAMP);
PyModule_AddIntConstant(m, "HCI_DATA_DIR", HCI_DATA_DIR);
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
PyModule_AddIntConstant(m, "BTPROTO_SCO", BTPROTO_SCO);
Add FreeBSD support for bluetooth sockets. (SF Patch #888148, reviewed by loewis)
2004-02-02 06:05:24 +00:00
#endif
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
PyModule_AddIntConstant(m, "BTPROTO_RFCOMM", BTPROTO_RFCOMM);
Simplify further by using AddStringConstant
2006-05-30 03:18:50 +00:00
PyModule_AddStringConstant(m, "BDADDR_ANY", "00:00:00:00:00:00");
PyModule_AddStringConstant(m, "BDADDR_LOCAL", "00:00:00:FF:FF:FF");
Patch #874083: Bluetooth support for socket module.
2004-01-31 12:34:17 +00:00
#endif
Autotest for netpacket/packet.h, as it is not available on all Linux versions. Depend AF_PACKET on HAVE_NETPACKET_PACKET_H. Fixes #449157
2001-08-10 20:29:40 +00:00
#ifdef HAVE_NETPACKET_PACKET_H
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AF_PACKET", AF_PACKET);
PyModule_AddIntConstant(m, "PF_PACKET", PF_PACKET);
PyModule_AddIntConstant(m, "PACKET_HOST", PACKET_HOST);
PyModule_AddIntConstant(m, "PACKET_BROADCAST", PACKET_BROADCAST);
PyModule_AddIntConstant(m, "PACKET_MULTICAST", PACKET_MULTICAST);
PyModule_AddIntConstant(m, "PACKET_OTHERHOST", PACKET_OTHERHOST);
PyModule_AddIntConstant(m, "PACKET_OUTGOING", PACKET_OUTGOING);
PyModule_AddIntConstant(m, "PACKET_LOOPBACK", PACKET_LOOPBACK);
PyModule_AddIntConstant(m, "PACKET_FASTROUTE", PACKET_FASTROUTE);
RISCOS changes by dschwertberger.
2001-03-02 06:34:14 +00:00
#endif
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
Issue #1646: Make socket support TIPC. The socket module now has support for TIPC under Linux, see http://tipc.sf.net/ for more information. Thanks to Alberto Bertogli for the patch
2008-01-07 16:12:44 +00:00
#ifdef HAVE_LINUX_TIPC_H
PyModule_AddIntConstant(m, "AF_TIPC", AF_TIPC);
/* for addresses */
PyModule_AddIntConstant(m, "TIPC_ADDR_NAMESEQ", TIPC_ADDR_NAMESEQ);
PyModule_AddIntConstant(m, "TIPC_ADDR_NAME", TIPC_ADDR_NAME);
PyModule_AddIntConstant(m, "TIPC_ADDR_ID", TIPC_ADDR_ID);
PyModule_AddIntConstant(m, "TIPC_ZONE_SCOPE", TIPC_ZONE_SCOPE);
PyModule_AddIntConstant(m, "TIPC_CLUSTER_SCOPE", TIPC_CLUSTER_SCOPE);
PyModule_AddIntConstant(m, "TIPC_NODE_SCOPE", TIPC_NODE_SCOPE);
/* for setsockopt() */
PyModule_AddIntConstant(m, "SOL_TIPC", SOL_TIPC);
PyModule_AddIntConstant(m, "TIPC_IMPORTANCE", TIPC_IMPORTANCE);
PyModule_AddIntConstant(m, "TIPC_SRC_DROPPABLE", TIPC_SRC_DROPPABLE);
PyModule_AddIntConstant(m, "TIPC_DEST_DROPPABLE",
TIPC_DEST_DROPPABLE);
PyModule_AddIntConstant(m, "TIPC_CONN_TIMEOUT", TIPC_CONN_TIMEOUT);
PyModule_AddIntConstant(m, "TIPC_LOW_IMPORTANCE",
TIPC_LOW_IMPORTANCE);
PyModule_AddIntConstant(m, "TIPC_MEDIUM_IMPORTANCE",
TIPC_MEDIUM_IMPORTANCE);
PyModule_AddIntConstant(m, "TIPC_HIGH_IMPORTANCE",
TIPC_HIGH_IMPORTANCE);
PyModule_AddIntConstant(m, "TIPC_CRITICAL_IMPORTANCE",
TIPC_CRITICAL_IMPORTANCE);
/* for subscriptions */
PyModule_AddIntConstant(m, "TIPC_SUB_PORTS", TIPC_SUB_PORTS);
PyModule_AddIntConstant(m, "TIPC_SUB_SERVICE", TIPC_SUB_SERVICE);
Guard definition of TIPC_SUB_CANCEL with an #ifdef.
2008-01-11 09:19:11 +00:00
#ifdef TIPC_SUB_CANCEL
/* doesn't seem to be available everywhere */
PyModule_AddIntConstant(m, "TIPC_SUB_CANCEL", TIPC_SUB_CANCEL);
#endif
Issue #1646: Make socket support TIPC. The socket module now has support for TIPC under Linux, see http://tipc.sf.net/ for more information. Thanks to Alberto Bertogli for the patch
2008-01-07 16:12:44 +00:00
PyModule_AddIntConstant(m, "TIPC_WAIT_FOREVER", TIPC_WAIT_FOREVER);
PyModule_AddIntConstant(m, "TIPC_PUBLISHED", TIPC_PUBLISHED);
PyModule_AddIntConstant(m, "TIPC_WITHDRAWN", TIPC_WITHDRAWN);
PyModule_AddIntConstant(m, "TIPC_SUBSCR_TIMEOUT", TIPC_SUBSCR_TIMEOUT);
PyModule_AddIntConstant(m, "TIPC_CFG_SRV", TIPC_CFG_SRV);
PyModule_AddIntConstant(m, "TIPC_TOP_SRV", TIPC_TOP_SRV);
#endif
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
/* Socket types */
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOCK_STREAM", SOCK_STREAM);
PyModule_AddIntConstant(m, "SOCK_DGRAM", SOCK_DGRAM);
Changes for BeOS, QNX and long long, by Chris Herborth.
1998-08-04 22:53:56 +00:00
#ifndef __BEOS__
/* We have incomplete socket support. */
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOCK_RAW", SOCK_RAW);
PyModule_AddIntConstant(m, "SOCK_SEQPACKET", SOCK_SEQPACKET);
Patch #650422: Use Posix AF_ constants instead of PF_ ones.
2002-12-11 13:10:57 +00:00
#if defined(SOCK_RDM)
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOCK_RDM", SOCK_RDM);
Changes for BeOS, QNX and long long, by Chris Herborth.
1998-08-04 22:53:56 +00:00
#endif
Patch #650422: Use Posix AF_ constants instead of PF_ ones.
2002-12-11 13:10:57 +00:00
#endif
add socket related constants
1995-02-07 15:39:52 +00:00
#ifdef SO_DEBUG
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_DEBUG", SO_DEBUG);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_ACCEPTCONN
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_ACCEPTCONN", SO_ACCEPTCONN);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_REUSEADDR
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_REUSEADDR", SO_REUSEADDR);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
[Patch #982665 ] add SO_EXCLUSIVEADDRUSE constant
2004-07-10 14:19:21 +00:00
#ifdef SO_EXCLUSIVEADDRUSE
PyModule_AddIntConstant(m, "SO_EXCLUSIVEADDRUSE", SO_EXCLUSIVEADDRUSE);
#endif
add socket related constants
1995-02-07 15:39:52 +00:00
#ifdef SO_KEEPALIVE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_KEEPALIVE", SO_KEEPALIVE);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_DONTROUTE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_DONTROUTE", SO_DONTROUTE);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_BROADCAST
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_BROADCAST", SO_BROADCAST);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_USELOOPBACK
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_USELOOPBACK", SO_USELOOPBACK);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_LINGER
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_LINGER", SO_LINGER);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_OOBINLINE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_OOBINLINE", SO_OOBINLINE);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_REUSEPORT
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_REUSEPORT", SO_REUSEPORT);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_SNDBUF
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_SNDBUF", SO_SNDBUF);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_RCVBUF
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_RCVBUF", SO_RCVBUF);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_SNDLOWAT
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_SNDLOWAT", SO_SNDLOWAT);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_RCVLOWAT
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_RCVLOWAT", SO_RCVLOWAT);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_SNDTIMEO
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_SNDTIMEO", SO_SNDTIMEO);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_RCVTIMEO
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_RCVTIMEO", SO_RCVTIMEO);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_ERROR
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_ERROR", SO_ERROR);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef SO_TYPE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SO_TYPE", SO_TYPE);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
/* Maximum number of connections for "listen" */
#ifdef SOMAXCONN
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOMAXCONN", SOMAXCONN);
add socket related constants
1995-02-07 15:39:52 +00:00
#else
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
PyModule_AddIntConstant(m, "SOMAXCONN", 5); /* Common value */
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
/* Flags for send, recv */
#ifdef MSG_OOB
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "MSG_OOB", MSG_OOB);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef MSG_PEEK
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "MSG_PEEK", MSG_PEEK);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef MSG_DONTROUTE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "MSG_DONTROUTE", MSG_DONTROUTE);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
Patch by Charles G Waldman: 1) Adds MSG_DONTWAIT if defined (I needed this) 2) Spells "coreectly" correctly ;-)
2000-04-25 21:34:53 +00:00
#ifdef MSG_DONTWAIT
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "MSG_DONTWAIT", MSG_DONTWAIT);
Patch by Charles G Waldman: 1) Adds MSG_DONTWAIT if defined (I needed this) 2) Spells "coreectly" correctly ;-)
2000-04-25 21:34:53 +00:00
#endif
add socket related constants
1995-02-07 15:39:52 +00:00
#ifdef MSG_EOR
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "MSG_EOR", MSG_EOR);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef MSG_TRUNC
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "MSG_TRUNC", MSG_TRUNC);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef MSG_CTRUNC
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "MSG_CTRUNC", MSG_CTRUNC);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef MSG_WAITALL
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "MSG_WAITALL", MSG_WAITALL);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef MSG_BTAG
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "MSG_BTAG", MSG_BTAG);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef MSG_ETAG
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "MSG_ETAG", MSG_ETAG);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
/* Protocol level and numbers, usable for [gs]etsockopt */
#ifdef SOL_SOCKET
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOL_SOCKET", SOL_SOCKET);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#ifdef SOL_IP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOL_IP", SOL_IP);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOL_IP", 0);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef SOL_IPX
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOL_IPX", SOL_IPX);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef SOL_AX25
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOL_AX25", SOL_AX25);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef SOL_ATALK
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOL_ATALK", SOL_ATALK);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef SOL_NETROM
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOL_NETROM", SOL_NETROM);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef SOL_ROSE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOL_ROSE", SOL_ROSE);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef SOL_TCP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOL_TCP", SOL_TCP);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOL_TCP", 6);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef SOL_UDP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOL_UDP", SOL_UDP);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "SOL_UDP", 17);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
add socket related constants
1995-02-07 15:39:52 +00:00
#ifdef IPPROTO_IP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_IP", IPPROTO_IP);
Some systems (e.g. Linux) use enums for some symbols (like IPPROTO_IP) so that our #ifdef test has the wrong effect. Substitute hardcoded values for some important symbols (but not for the whole range -- some are pretty obscure so it's not worth it).
1998-05-28 20:18:18 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_IP", 0);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#ifdef IPPROTO_HOPOPTS
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_HOPOPTS", IPPROTO_HOPOPTS);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
add socket related constants
1995-02-07 15:39:52 +00:00
#ifdef IPPROTO_ICMP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_ICMP", IPPROTO_ICMP);
Some systems (e.g. Linux) use enums for some symbols (like IPPROTO_IP) so that our #ifdef test has the wrong effect. Substitute hardcoded values for some important symbols (but not for the whole range -- some are pretty obscure so it's not worth it).
1998-05-28 20:18:18 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_ICMP", 1);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IPPROTO_IGMP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_IGMP", IPPROTO_IGMP);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IPPROTO_GGP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_GGP", IPPROTO_GGP);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#ifdef IPPROTO_IPV4
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_IPV4", IPPROTO_IPV4);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
Patch #813445: Add missing socket.IPPROTO_IPV6. Backported to 2.3.
2003-10-03 13:56:20 +00:00
#ifdef IPPROTO_IPV6
PyModule_AddIntConstant(m, "IPPROTO_IPV6", IPPROTO_IPV6);
#endif
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#ifdef IPPROTO_IPIP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_IPIP", IPPROTO_IPIP);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
add socket related constants
1995-02-07 15:39:52 +00:00
#ifdef IPPROTO_TCP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_TCP", IPPROTO_TCP);
Some systems (e.g. Linux) use enums for some symbols (like IPPROTO_IP) so that our #ifdef test has the wrong effect. Substitute hardcoded values for some important symbols (but not for the whole range -- some are pretty obscure so it's not worth it).
1998-05-28 20:18:18 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_TCP", 6);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IPPROTO_EGP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_EGP", IPPROTO_EGP);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IPPROTO_PUP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_PUP", IPPROTO_PUP);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IPPROTO_UDP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_UDP", IPPROTO_UDP);
Some systems (e.g. Linux) use enums for some symbols (like IPPROTO_IP) so that our #ifdef test has the wrong effect. Substitute hardcoded values for some important symbols (but not for the whole range -- some are pretty obscure so it's not worth it).
1998-05-28 20:18:18 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_UDP", 17);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IPPROTO_IDP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_IDP", IPPROTO_IDP);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IPPROTO_HELLO
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_HELLO", IPPROTO_HELLO);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IPPROTO_ND
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_ND", IPPROTO_ND);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IPPROTO_TP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_TP", IPPROTO_TP);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#ifdef IPPROTO_IPV6
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_IPV6", IPPROTO_IPV6);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPPROTO_ROUTING
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_ROUTING", IPPROTO_ROUTING);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPPROTO_FRAGMENT
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_FRAGMENT", IPPROTO_FRAGMENT);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPPROTO_RSVP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_RSVP", IPPROTO_RSVP);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPPROTO_GRE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_GRE", IPPROTO_GRE);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPPROTO_ESP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_ESP", IPPROTO_ESP);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPPROTO_AH
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_AH", IPPROTO_AH);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPPROTO_MOBILE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_MOBILE", IPPROTO_MOBILE);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPPROTO_ICMPV6
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_ICMPV6", IPPROTO_ICMPV6);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPPROTO_NONE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_NONE", IPPROTO_NONE);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPPROTO_DSTOPTS
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_DSTOPTS", IPPROTO_DSTOPTS);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
add socket related constants
1995-02-07 15:39:52 +00:00
#ifdef IPPROTO_XTP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_XTP", IPPROTO_XTP);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IPPROTO_EON
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_EON", IPPROTO_EON);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#ifdef IPPROTO_PIM
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_PIM", IPPROTO_PIM);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPPROTO_IPCOMP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_IPCOMP", IPPROTO_IPCOMP);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPPROTO_VRRP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_VRRP", IPPROTO_VRRP);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
add socket related constants
1995-02-07 15:39:52 +00:00
#ifdef IPPROTO_BIP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_BIP", IPPROTO_BIP);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
/**/
#ifdef IPPROTO_RAW
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_RAW", IPPROTO_RAW);
Some systems (e.g. Linux) use enums for some symbols (like IPPROTO_IP) so that our #ifdef test has the wrong effect. Substitute hardcoded values for some important symbols (but not for the whole range -- some are pretty obscure so it's not worth it).
1998-05-28 20:18:18 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_RAW", 255);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IPPROTO_MAX
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPROTO_MAX", IPPROTO_MAX);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
/* Some port configuration */
#ifdef IPPORT_RESERVED
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPORT_RESERVED", IPPORT_RESERVED);
add socket related constants
1995-02-07 15:39:52 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPORT_RESERVED", 1024);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IPPORT_USERRESERVED
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPORT_USERRESERVED", IPPORT_USERRESERVED);
add socket related constants
1995-02-07 15:39:52 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPPORT_USERRESERVED", 5000);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
/* Some reserved IP v.4 addresses */
#ifdef INADDR_ANY
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_ANY", INADDR_ANY);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_ANY", 0x00000000);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#endif
#ifdef INADDR_BROADCAST
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_BROADCAST", INADDR_BROADCAST);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_BROADCAST", 0xffffffff);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#endif
#ifdef INADDR_LOOPBACK
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_LOOPBACK", INADDR_LOOPBACK);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_LOOPBACK", 0x7F000001);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#endif
#ifdef INADDR_UNSPEC_GROUP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_UNSPEC_GROUP", INADDR_UNSPEC_GROUP);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_UNSPEC_GROUP", 0xe0000000);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#endif
#ifdef INADDR_ALLHOSTS_GROUP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_ALLHOSTS_GROUP",
INADDR_ALLHOSTS_GROUP);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_ALLHOSTS_GROUP", 0xe0000001);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#endif
#ifdef INADDR_MAX_LOCAL_GROUP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_MAX_LOCAL_GROUP",
INADDR_MAX_LOCAL_GROUP);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_MAX_LOCAL_GROUP", 0xe00000ff);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#endif
#ifdef INADDR_NONE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_NONE", INADDR_NONE);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#else
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "INADDR_NONE", 0xffffffff);
add INADDR_* symbols
1995-02-17 15:11:07 +00:00
#endif
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
/* IPv4 [gs]etsockopt options */
add socket related constants
1995-02-07 15:39:52 +00:00
#ifdef IP_OPTIONS
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_OPTIONS", IP_OPTIONS);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IP_HDRINCL
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_HDRINCL", IP_HDRINCL);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IP_TOS
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_TOS", IP_TOS);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IP_TTL
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_TTL", IP_TTL);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IP_RECVOPTS
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_RECVOPTS", IP_RECVOPTS);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IP_RECVRETOPTS
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_RECVRETOPTS", IP_RECVRETOPTS);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IP_RECVDSTADDR
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_RECVDSTADDR", IP_RECVDSTADDR);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IP_RETOPTS
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_RETOPTS", IP_RETOPTS);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IP_MULTICAST_IF
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_MULTICAST_IF", IP_MULTICAST_IF);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IP_MULTICAST_TTL
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_MULTICAST_TTL", IP_MULTICAST_TTL);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IP_MULTICAST_LOOP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_MULTICAST_LOOP", IP_MULTICAST_LOOP);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IP_ADD_MEMBERSHIP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_ADD_MEMBERSHIP", IP_ADD_MEMBERSHIP);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
#ifdef IP_DROP_MEMBERSHIP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_DROP_MEMBERSHIP", IP_DROP_MEMBERSHIP);
add socket related constants
1995-02-07 15:39:52 +00:00
#endif
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#ifdef IP_DEFAULT_MULTICAST_TTL
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_DEFAULT_MULTICAST_TTL",
IP_DEFAULT_MULTICAST_TTL);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef IP_DEFAULT_MULTICAST_LOOP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_DEFAULT_MULTICAST_LOOP",
IP_DEFAULT_MULTICAST_LOOP);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef IP_MAX_MEMBERSHIPS
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IP_MAX_MEMBERSHIPS", IP_MAX_MEMBERSHIPS);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
/* IPv6 [gs]etsockopt options, defined in RFC2553 */
#ifdef IPV6_JOIN_GROUP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPV6_JOIN_GROUP", IPV6_JOIN_GROUP);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPV6_LEAVE_GROUP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPV6_LEAVE_GROUP", IPV6_LEAVE_GROUP);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPV6_MULTICAST_HOPS
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPV6_MULTICAST_HOPS", IPV6_MULTICAST_HOPS);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPV6_MULTICAST_IF
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPV6_MULTICAST_IF", IPV6_MULTICAST_IF);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPV6_MULTICAST_LOOP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPV6_MULTICAST_LOOP", IPV6_MULTICAST_LOOP);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
#ifdef IPV6_UNICAST_HOPS
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPV6_UNICAST_HOPS", IPV6_UNICAST_HOPS);
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
#endif
Add IPV6_ socket options from RFCs 3493 and 3542. Fixes #867012.
2003-12-30 11:14:01 +00:00
/* Additional IPV6 socket options, defined in RFC 3493 */
#ifdef IPV6_V6ONLY
PyModule_AddIntConstant(m, "IPV6_V6ONLY", IPV6_V6ONLY);
#endif
/* Advanced IPV6 socket options, from RFC 3542 */
#ifdef IPV6_CHECKSUM
PyModule_AddIntConstant(m, "IPV6_CHECKSUM", IPV6_CHECKSUM);
#endif
#ifdef IPV6_DONTFRAG
PyModule_AddIntConstant(m, "IPV6_DONTFRAG", IPV6_DONTFRAG);
#endif
#ifdef IPV6_DSTOPTS
PyModule_AddIntConstant(m, "IPV6_DSTOPTS", IPV6_DSTOPTS);
#endif
#ifdef IPV6_HOPLIMIT
PyModule_AddIntConstant(m, "IPV6_HOPLIMIT", IPV6_HOPLIMIT);
#endif
#ifdef IPV6_HOPOPTS
PyModule_AddIntConstant(m, "IPV6_HOPOPTS", IPV6_HOPOPTS);
#endif
#ifdef IPV6_NEXTHOP
PyModule_AddIntConstant(m, "IPV6_NEXTHOP", IPV6_NEXTHOP);
#endif
#ifdef IPV6_PATHMTU
PyModule_AddIntConstant(m, "IPV6_PATHMTU", IPV6_PATHMTU);
#endif
#ifdef IPV6_PKTINFO
PyModule_AddIntConstant(m, "IPV6_PKTINFO", IPV6_PKTINFO);
#endif
#ifdef IPV6_RECVDSTOPTS
PyModule_AddIntConstant(m, "IPV6_RECVDSTOPTS", IPV6_RECVDSTOPTS);
#endif
#ifdef IPV6_RECVHOPLIMIT
PyModule_AddIntConstant(m, "IPV6_RECVHOPLIMIT", IPV6_RECVHOPLIMIT);
#endif
#ifdef IPV6_RECVHOPOPTS
PyModule_AddIntConstant(m, "IPV6_RECVHOPOPTS", IPV6_RECVHOPOPTS);
#endif
#ifdef IPV6_RECVPKTINFO
PyModule_AddIntConstant(m, "IPV6_RECVPKTINFO", IPV6_RECVPKTINFO);
#endif
#ifdef IPV6_RECVRTHDR
PyModule_AddIntConstant(m, "IPV6_RECVRTHDR", IPV6_RECVRTHDR);
#endif
#ifdef IPV6_RECVTCLASS
PyModule_AddIntConstant(m, "IPV6_RECVTCLASS", IPV6_RECVTCLASS);
#endif
#ifdef IPV6_RTHDR
PyModule_AddIntConstant(m, "IPV6_RTHDR", IPV6_RTHDR);
#endif
#ifdef IPV6_RTHDRDSTOPTS
PyModule_AddIntConstant(m, "IPV6_RTHDRDSTOPTS", IPV6_RTHDRDSTOPTS);
#endif
#ifdef IPV6_RTHDR_TYPE_0
PyModule_AddIntConstant(m, "IPV6_RTHDR_TYPE_0", IPV6_RTHDR_TYPE_0);
#endif
#ifdef IPV6_RECVPATHMTU
PyModule_AddIntConstant(m, "IPV6_RECVPATHMTU", IPV6_RECVPATHMTU);
#endif
#ifdef IPV6_TCLASS
PyModule_AddIntConstant(m, "IPV6_TCLASS", IPV6_TCLASS);
#endif
#ifdef IPV6_USE_MIN_MTU
PyModule_AddIntConstant(m, "IPV6_USE_MIN_MTU", IPV6_USE_MIN_MTU);
#endif
Add more constants. Contributed by itojun.
2001-08-04 22:37:23 +00:00
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
/* TCP options */
#ifdef TCP_NODELAY
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "TCP_NODELAY", TCP_NODELAY);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
#ifdef TCP_MAXSEG
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "TCP_MAXSEG", TCP_MAXSEG);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
Add TCP socket options from glibc 2.2.4. Fixes #495680. 2.2.1 bugfix candidate.
2001-12-22 15:05:32 +00:00
#ifdef TCP_CORK
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "TCP_CORK", TCP_CORK);
Add TCP socket options from glibc 2.2.4. Fixes #495680. 2.2.1 bugfix candidate.
2001-12-22 15:05:32 +00:00
#endif
#ifdef TCP_KEEPIDLE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "TCP_KEEPIDLE", TCP_KEEPIDLE);
Add TCP socket options from glibc 2.2.4. Fixes #495680. 2.2.1 bugfix candidate.
2001-12-22 15:05:32 +00:00
#endif
#ifdef TCP_KEEPINTVL
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "TCP_KEEPINTVL", TCP_KEEPINTVL);
Add TCP socket options from glibc 2.2.4. Fixes #495680. 2.2.1 bugfix candidate.
2001-12-22 15:05:32 +00:00
#endif
#ifdef TCP_KEEPCNT
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "TCP_KEEPCNT", TCP_KEEPCNT);
Add TCP socket options from glibc 2.2.4. Fixes #495680. 2.2.1 bugfix candidate.
2001-12-22 15:05:32 +00:00
#endif
#ifdef TCP_SYNCNT
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "TCP_SYNCNT", TCP_SYNCNT);
Add TCP socket options from glibc 2.2.4. Fixes #495680. 2.2.1 bugfix candidate.
2001-12-22 15:05:32 +00:00
#endif
#ifdef TCP_LINGER2
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "TCP_LINGER2", TCP_LINGER2);
Add TCP socket options from glibc 2.2.4. Fixes #495680. 2.2.1 bugfix candidate.
2001-12-22 15:05:32 +00:00
#endif
#ifdef TCP_DEFER_ACCEPT
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "TCP_DEFER_ACCEPT", TCP_DEFER_ACCEPT);
Add TCP socket options from glibc 2.2.4. Fixes #495680. 2.2.1 bugfix candidate.
2001-12-22 15:05:32 +00:00
#endif
#ifdef TCP_WINDOW_CLAMP
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "TCP_WINDOW_CLAMP", TCP_WINDOW_CLAMP);
Add TCP socket options from glibc 2.2.4. Fixes #495680. 2.2.1 bugfix candidate.
2001-12-22 15:05:32 +00:00
#endif
#ifdef TCP_INFO
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "TCP_INFO", TCP_INFO);
Add TCP socket options from glibc 2.2.4. Fixes #495680. 2.2.1 bugfix candidate.
2001-12-22 15:05:32 +00:00
#endif
#ifdef TCP_QUICKACK
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "TCP_QUICKACK", TCP_QUICKACK);
Add TCP socket options from glibc 2.2.4. Fixes #495680. 2.2.1 bugfix candidate.
2001-12-22 15:05:32 +00:00
#endif
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
/* IPX options */
#ifdef IPX_TYPE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "IPX_TYPE", IPX_TYPE);
Patch by Marc-Andre Lemburg to add a bunch more symbols.
1999-08-09 14:40:40 +00:00
#endif
Threading speedup patches by Christopher Lindblad <cjl@infoseek.com>.
1998-04-09 20:56:35 +00:00
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
/* get{addr,name}info parameters */
#ifdef EAI_ADDRFAMILY
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_ADDRFAMILY", EAI_ADDRFAMILY);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef EAI_AGAIN
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_AGAIN", EAI_AGAIN);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef EAI_BADFLAGS
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_BADFLAGS", EAI_BADFLAGS);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef EAI_FAIL
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_FAIL", EAI_FAIL);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef EAI_FAMILY
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_FAMILY", EAI_FAMILY);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef EAI_MEMORY
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_MEMORY", EAI_MEMORY);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef EAI_NODATA
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_NODATA", EAI_NODATA);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef EAI_NONAME
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_NONAME", EAI_NONAME);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
Add IPV6_ socket options from RFCs 3493 and 3542. Fixes #867012.
2003-12-30 11:14:01 +00:00
#ifdef EAI_OVERFLOW
PyModule_AddIntConstant(m, "EAI_OVERFLOW", EAI_OVERFLOW);
#endif
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#ifdef EAI_SERVICE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_SERVICE", EAI_SERVICE);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef EAI_SOCKTYPE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_SOCKTYPE", EAI_SOCKTYPE);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef EAI_SYSTEM
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_SYSTEM", EAI_SYSTEM);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef EAI_BADHINTS
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_BADHINTS", EAI_BADHINTS);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef EAI_PROTOCOL
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_PROTOCOL", EAI_PROTOCOL);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef EAI_MAX
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "EAI_MAX", EAI_MAX);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef AI_PASSIVE
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AI_PASSIVE", AI_PASSIVE);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef AI_CANONNAME
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AI_CANONNAME", AI_CANONNAME);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef AI_NUMERICHOST
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AI_NUMERICHOST", AI_NUMERICHOST);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
Add IPV6_ socket options from RFCs 3493 and 3542. Fixes #867012.
2003-12-30 11:14:01 +00:00
#ifdef AI_NUMERICSERV
PyModule_AddIntConstant(m, "AI_NUMERICSERV", AI_NUMERICSERV);
#endif
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#ifdef AI_MASK
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AI_MASK", AI_MASK);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef AI_ALL
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AI_ALL", AI_ALL);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef AI_V4MAPPED_CFG
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AI_V4MAPPED_CFG", AI_V4MAPPED_CFG);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef AI_ADDRCONFIG
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AI_ADDRCONFIG", AI_ADDRCONFIG);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef AI_V4MAPPED
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AI_V4MAPPED", AI_V4MAPPED);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef AI_DEFAULT
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "AI_DEFAULT", AI_DEFAULT);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef NI_MAXHOST
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "NI_MAXHOST", NI_MAXHOST);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef NI_MAXSERV
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "NI_MAXSERV", NI_MAXSERV);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef NI_NOFQDN
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "NI_NOFQDN", NI_NOFQDN);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef NI_NUMERICHOST
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "NI_NUMERICHOST", NI_NUMERICHOST);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef NI_NAMEREQD
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "NI_NAMEREQD", NI_NAMEREQD);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef NI_NUMERICSERV
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "NI_NUMERICSERV", NI_NUMERICSERV);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
#ifdef NI_DGRAM
Use the PyModule_Add*() APIs instead of manipulating the module dict directly.
2002-04-01 14:53:37 +00:00
PyModule_AddIntConstant(m, "NI_DGRAM", NI_DGRAM);
Patch #401196: IPv6 extensions to the socket module. New functions getnameinfo, getaddrinfo. New exceptions socket.gaierror, socket.herror. Various new constants, in particular AF_INET6 and error codes and parameters for getaddrinfo. AF_INET6 support in setipaddr, makesockaddr, getsockaddr, getsockaddrlen, gethost_common, PySocket_gethostbyaddr.
2001-07-21 18:05:31 +00:00
#endif
Patch #849595: Add socket.shutdown() constants.
2003-11-27 19:40:22 +00:00
/* shutdown() parameters */
#ifdef SHUT_RD
PyModule_AddIntConstant(m, "SHUT_RD", SHUT_RD);
#elif defined(SD_RECEIVE)
PyModule_AddIntConstant(m, "SHUT_RD", SD_RECEIVE);
#else
PyModule_AddIntConstant(m, "SHUT_RD", 0);
#endif
#ifdef SHUT_WR
PyModule_AddIntConstant(m, "SHUT_WR", SHUT_WR);
#elif defined(SD_SEND)
PyModule_AddIntConstant(m, "SHUT_WR", SD_SEND);
#else
PyModule_AddIntConstant(m, "SHUT_WR", 1);
#endif
#ifdef SHUT_RDWR
PyModule_AddIntConstant(m, "SHUT_RDWR", SHUT_RDWR);
#elif defined(SD_BOTH)
PyModule_AddIntConstant(m, "SHUT_RDWR", SD_BOTH);
#else
PyModule_AddIntConstant(m, "SHUT_RDWR", 2);
#endif
Added interface to Windows' WSAIoctl and a simple example for a network sniffer.
2008-01-04 15:23:30 +00:00
#ifdef SIO_RCVALL
Make Modules/socketobject.c compile for Windows again.
2008-01-13 11:19:43 +00:00
{
PyObject *tmp;
tmp = PyLong_FromUnsignedLong(SIO_RCVALL);
if (tmp == NULL)
return;
PyModule_AddObject(m, "SIO_RCVALL", tmp);
}
Added interface to Windows' WSAIoctl and a simple example for a network sniffer.
2008-01-04 15:23:30 +00:00
PyModule_AddIntConstant(m, "RCVALL_OFF", RCVALL_OFF);
PyModule_AddIntConstant(m, "RCVALL_ON", RCVALL_ON);
PyModule_AddIntConstant(m, "RCVALL_SOCKETLEVELONLY", RCVALL_SOCKETLEVELONLY);
Repair compilation for Visual Studio 2005. I applied the same changes manually to VS7.1 and VC6 files; completely untested. (Christian, don't try too hard merging this change into py3k. It will be easier to do the same work again on the branch)
2008-03-28 21:55:29 +00:00
#ifdef RCVALL_IPLEVEL
Added interface to Windows' WSAIoctl and a simple example for a network sniffer.
2008-01-04 15:23:30 +00:00
PyModule_AddIntConstant(m, "RCVALL_IPLEVEL", RCVALL_IPLEVEL);
Repair compilation for Visual Studio 2005. I applied the same changes manually to VS7.1 and VC6 files; completely untested. (Christian, don't try too hard merging this change into py3k. It will be easier to do the same work again on the branch)
2008-03-28 21:55:29 +00:00
#endif
#ifdef RCVALL_MAX
Added interface to Windows' WSAIoctl and a simple example for a network sniffer.
2008-01-04 15:23:30 +00:00
PyModule_AddIntConstant(m, "RCVALL_MAX", RCVALL_MAX);
Repair compilation for Visual Studio 2005. I applied the same changes manually to VS7.1 and VC6 files; completely untested. (Christian, don't try too hard merging this change into py3k. It will be easier to do the same work again on the branch)
2008-03-28 21:55:29 +00:00
#endif
Added interface to Windows' WSAIoctl and a simple example for a network sniffer.
2008-01-04 15:23:30 +00:00
#endif /* _MSTCPIP_ */
Threading speedup patches by Christopher Lindblad <cjl@infoseek.com>.
1998-04-09 20:56:35 +00:00
/* Initialize gethostbyname lock */
[ 731644] & [ 604210 ] Release the GIL around getaddrinfo(), yet protect access with lock on those platforms that getaddrinfo() isn't (known to be) thread-safe. Thanks to MvL for mentoring this patch.
2003-05-09 07:53:18 +00:00
#if defined(USE_GETHOSTBYNAME_LOCK) || defined(USE_GETADDRINFO_LOCK)
netdb_lock = PyThread_allocate_lock();
Threading speedup patches by Christopher Lindblad <cjl@infoseek.com>.
1998-04-09 20:56:35 +00:00
#endif
Initial revision
1991-06-25 21:36:08 +00:00
}
Emulate inet_{pton,ntop} on systems that don't provide it.
2001-06-24 21:18:26 +00:00
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
Emulate inet_{pton,ntop} on systems that don't provide it.
2001-06-24 21:18:26 +00:00
#ifndef HAVE_INET_PTON
Backport of the PCbuild9 directory from the py3k branch. I've finished the last task for the PCbuild9 directory today. I don't think there is much left to do. Now you can all play around with the shiny new VS 2008 and try the PGO builds. I was able to get a speed improvement of about 10% on py3k. Have fun! :)
2007-11-22 11:21:16 +00:00
#if !defined(NTDDI_VERSION) || (NTDDI_VERSION < NTDDI_LONGHORN)
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
/* Simplistic emulation code for inet_pton that only works for IPv4 */
Patch by Jp Calderone: - The socket module now provides the functions inet_pton and inet_ntop for converting between string and packed representation of IP addresses. See SF patch #658327. This still needs a bit of work in the doc area, because it is not available on all platforms (especially not on Windows).
2003-04-25 05:48:32 +00:00
/* These are not exposed because they do not set errno properly */
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
Whitespace normalization, folding long lines, uniform comment delimiters. Also repaired some docstrings and comments.
2002-06-07 02:08:35 +00:00
int
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
inet_pton(int af, const char *src, void *dst)
Emulate inet_{pton,ntop} on systems that don't provide it.
2001-06-24 21:18:26 +00:00
{
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
if (af == AF_INET) {
Issue #1008086: Fixes socket.inet_aton() to always return 4 bytes even on LP64 platforms (most 64-bit Linux, bsd, unix systems).
2009-02-11 23:45:25 +00:00
#if (SIZEOF_INT != 4)
#error "Not sure if in_addr_t exists and int is not 32-bits."
#endif
unsigned int packed_addr;
Emulate inet_{pton,ntop} on systems that don't provide it.
2001-06-24 21:18:26 +00:00
packed_addr = inet_addr(src);
if (packed_addr == INADDR_NONE)
return 0;
memcpy(dst, &packed_addr, 4);
return 1;
}
/* Should set errno to EAFNOSUPPORT */
return -1;
}
Silence warnings in MSVC++: hide unused variables, add constness back to inet_pton/ntop, convert htons argument to u_short.
2001-07-21 09:42:15 +00:00
const char *
inet_ntop(int af, const void *src, char *dst, socklen_t size)
Emulate inet_{pton,ntop} on systems that don't provide it.
2001-06-24 21:18:26 +00:00
{
if (af == AF_INET) {
struct in_addr packed_addr;
if (size < 16)
/* Should set errno to ENOSPC. */
return NULL;
memcpy(&packed_addr, src, sizeof(packed_addr));
return strncpy(dst, inet_ntoa(packed_addr), size);
}
/* Should set errno to EAFNOSUPPORT */
return NULL;
}
Major cleanup. Renamed static methods to avoid Py prefix. Other misc cleanup as well, e.g. renamed NTinit to os_init.
2002-06-07 03:19:37 +00:00
Emulate inet_{pton,ntop} on systems that don't provide it.
2001-06-24 21:18:26 +00:00
#endif
Backport of the PCbuild9 directory from the py3k branch. I've finished the last task for the PCbuild9 directory today. I don't think there is much left to do. Now you can all play around with the shiny new VS 2008 and try the PGO builds. I was able to get a speed improvement of about 10% on py3k. Have fun! :)
2007-11-22 11:21:16 +00:00
#endif
Reference in a new issue Copy permalink