Specifically:
* zmq_event_t should not be used internally in libzmq, it was
meant to be an outward facing structure.
* In 4.x, zmq_event_t does not correspond to monitor events, so
I removed the structure entirely.
* man page for zmq_socket_monitor is incomplete and the example
code was particularly nasty.
* test_monitor.cpp needed rewriting, it was not clean.
This bug is caused by fq which can terminate pipe before sending all
frames of a message first. So sometimes two messages were mixed and this
confused ZAP handler.
This patch just modifies one pipe parameter so that socket consumes all
messages before dropping the pipe.
On Solaris 8, 9, 10/SPARC, iov_base is of type caddr_t which is char *. The Sun C++ compiler errors with "Cannot assign void* to char*". Using a static case to override this. On Solaris 11, HP-UX, AIX, and RHEL, iov_base is void * so no issues there. This seems a rather hackish solution so open to something better.
zmq.h requires definitions from platform.hpp so include platform.hpp before zmq.h. This is only an intermediate fix though. zmq.h should not require definitions from platform.hpp because platform.hpp is not installed. So, when zmq.h wants to include <inttypes.h> on ZMQ_HAVE_SOLARIS || ZMQ_HAVE_OPENVMS, it will not and fall back to #include <stdint.h> which will fail when building something like pyzmq on Solaris 8, 9/SPARC which do not have <stdint.h>
Mingw64 provides mstcpip.h and the build fails (redefinition) if the struct tcp_keepalive is redefined. Do not define the struct if __MINGW64__ is defined. Note that I had to manually pass the compile definition to cmake: -D__MINGW64__=1
V547 Expression 'file_desc >= 0' is always true. Unsigned type value is always >= 0. socket_base.cpp 845
V547 Expression 'file_desc >= 0' is always true. Unsigned type value is always >= 0. socket_base.cpp 863
V547 Expression 'file_desc >= 0' is always true. Unsigned type value is always >= 0. socket_base.cpp 897
http://www.viva64.com/en/d/0137/print/
The decision about the poller mechanism to use (select, poll, ...)
was done twice: once by the build system and once by the code in
poller.hpp. As the build-system can actually detect the mechanisms
available, prefer that result to the hard coded defaults in
poller.hpp.
At the same time, remove the duplicate detection of select() vs.
poll()-variant from proxy.cpp, signaler.cpp and zmq.cpp.
This patch has not been tested on many build platforms: especially
the cmake build needs testing / patching. For the other builds,
hard code the result as these these are all Windows platforms.
The new options allows querying the maximum allowed number of sockets.
This is system dependent and cannot be encoded in the include file as a
preprocessor macro: for ZMQ_USE_SELECT, this depends on the FD_SETSIZE
macro at time of library compilation, not at time of include file use.
The ${libdir} was getting replaced/removed by configure_file() making pkg-config give bad flags: -L -lzmq
My fix was to add @ONLY to configure_file() so ${} style pkg-config substitutions are left alone.
In addition, I put the other typical ${} substitutions back into the libzmq.pc, since its now safe.
- renamed to ZMQ_CONNECT_RID
- fixed whitespace malformating around previous patch
- renamamed next_peer_id to next_rid in preparation for
larger rename of IDENTITY to ROUTING_ID
Note: ZMQ_CONNECT_RID has no test case and no entry in the man
page, as yet.
This change simply provides the user with a socket option that sets a user defined name of the next outbound connection:
zmq_setsockopt(routerSock,ZMQ_NEXT_IDENTITY,"myname",6);
if(0 > zmq_connect(routerSock,"tcp://127.0.0.1:1234")) return 1;
ret = zmq_send(routerSock,"myname",6,ZMQ_SNDMORE);
zmq_send(routerSock,b.mem,b.used,0);
In this example, the socket is immediately given the name "myname", and is capable of immediately sending traffic.
This approach is more effective in three ways:
1) It prevents all sorts of malicious peer naming attacks that can cause undefined behavior in existing ROUTER connections. (Two connections are made that both transmit the same name to the ROUTER, the ROUTER behavior is undefined)
2) It allows immediate control of connections made to external parties for STREAM sockets. Something that is not possible right now. Before an outbound connection had no name for STREAM or ROUTER sockets because outbound connections cannot be sent to without first receiving traffic.
3) It is simpler and more general than expecting two ROUTER sockets to handshake on assigned connection names. Plus it allows inline sending to new connections on ROUTER.
When a ZMQ_STREAM socket connection is broken (intentionally, via `shutdown()`
or accidentally via client crash or network failure), there is no way for the
application to dertermine that it should drop per-connection data (such as
buffers).
This contribution makes sure the application receives a 0-length message to
notify it that the connection has been broken. This is symmetric with the
process of closing the connection from within the application (where the
application sends a 0-length message to tell ZeroMQ to close the connection).
Conflicts:
CMakeLists.txt
