Significantly reworked the monitoring infrastructure with a more granular per socket API and to play well with monitoring endpoints in application threads

* zmq_device is now a wrapper that calls zmq_proxy
* zmq_proxy adds capture socket

Include a note about potential interactions with reactors, and update
reference to old FAIL_UNROUTABLE name.

The previous note in connect introduced but did not elaborate on the
requirement to bind before connect in inproc. As that discussion is in
detail in the inproc docs link to there.

Based on discussion with Steve O on the list, make the difference
between bind and connect usage more clear, and add a note reflecting the
fact that inproc must have bind before connect to reinforce the
information in zmq_inproc.

This patch, salvaged from a trainwreck accidental merge earlier, adds a
new sockopt, ZMQ_DELAY_ATTACH_ON_CONNECT which prevents a end point
being available to push messages to until it has fully connected, making
connect work more like bind. This also applies to reconnecting sockets,
which may cause message loss of in-queue messages, so it is sensible to
use this in conjunction with a low HWM and potentially an alternative
acknowledgement path.

Notes on most of the individual commits can be found the repository log.

Revert "After speaking with Ben Gray and the discussion on the mailing list, this is an attempt to create a sockopt to allow connecting pipes to not immediately be available for traffic. The problem is in a PUSH to many PULL situation, where there is a connect to a PULL which is not there. This connect will immediately create a pipe (unlike bind), and traffic will be load balanced to that pipe. This means if there is a persistently unavailable end point then the traffic will queue until HWM is hit, and older messages will be lost."

This reverts commit fe3fb419.

After speaking with Ben Gray and the discussion on the mailing list, this is an attempt to create a sockopt to allow connecting pipes to not immediately be available for traffic. The problem is in a PUSH to many PULL situation, where there is a connect to a PULL which is not there. This connect will immediately create a pipe (unlike bind), and traffic will be load balanced to that pipe. This means if there is a persistently unavailable end point then the traffic will queue until HWM is hit, and older messages will be lost.

This patch adds a sockopt ZMQ_DELAY_ATTACH_ON_CONNECT, which if set to 1 will attempt to preempt this behavior. It does this by extending the use of the session_base to include in the outbound as well as the inbound pipe, and only associates the pipe with the socket once it receives the connected callback via a process_attach message. This works, and a test has been added to show so, but may introduce unexpected complications. The shutdown logic in this class has become marginally more awkward because of this, requiring the session to serve as the sink for both pipes if shutdown occurs with a still-connecting pipe in place. It is also possible there could be issues around flushing the messages, but as I could not directly think how to create such an issue I have not written any code with regards to that.

The documentation has been updated to reflect the change, but please do check over the code and test and review.

The PGM transport supports the omission of the network interface to
select the default one like:

  announce.connect("epgm://eth0;239.255.128.46:64646"); // Use eth0
  announce.connect("epgm://239.255.128.46:64646"); // Use the default

Also, mention C++ in the additional community bindings of 0MQ in
zmq.txt.

Rename type zmq_monitor_fn -> zmq_monitor for a more natural callback definition API (zmq_monitor type, monitor.function callback)

Change zmq_monitor_fn type to cast between pointer-to-object and pointer-to-function in a more standards compliant way

The current ZMQ_MONITOR code does not compile in gcc 4.7, as -pedantic
and -Werror are enabled, and ISO C++ doesn't allow casting between
normal pointers (void*) and function pointers, as pedantically their
size could be different. This caused the library not compilable. This
commit workaround the problem by introducing one more indirection, i.e.
instead of calling

    (void *)listener

which is an error, we have to use

    *(void **)&listener

which is an undefined behavior :) but works on most platforms

Also, `optval_ = monitor` will not set the parameter in getsockopt(),
and the extra casting caused the LHS to be an rvalue which again makes
the code not compilable. The proper way is to pass a pointer of function
pointer and assign with indirection, i.e. `*optval_ = monitor`.

Also, fixed an asciidoc error in zmq_getsockopt.txt because the `~~~~`
is too long.