This is based on https://github.com/zeromq/zeromq3-x/pull/116/files

Make OS Xframwork build optional

* Fix possible clock_t crash on iOS 9.3.2 and 9.3.5 + macOS < 10.12

* No need to read return value

* Problem: zmq_z85_decode does not validate its input
Solution: added checks for invalid characters and overflows

* Added tests, added further check for range overflow, removed (multiple) calls to strlen

* Problem: gcc fails to build
Solution: added missing include directive

* Added VS2015 test_utils_z85 project

* Fixed indentation and copyright notice

* Resolved garbage from merge

* Revert "Added VS2015 test_utils_z85 project"

This reverts commit c58b3c664c144326e77135aa1184b6f0dee11143.

* Problem: test calls zmq_z85_decode with a NULL dest
Solution: call zmq_z85_decode with a properly sized buffer

* Problem: tests for zmq_z85_* scattered over two files
Solution: merged files

* Removed reference to removed test file from CMakeLists.txt

* Problem: Missing include directive to stdint.h
Solution: Added include directive

* Define __STDC_LIMIT_MACROS before including stdint.h

* Problem: Wrong variable is checked for invalid character marker
Solution: Use correct variable

CMake static library output does not match ZeroMQ Config.

Configuration used in ZeroMQConfig.cmake.in does not match to
explicic "libzmq" output of static build.

Problem: cmake artefacts different from make generated artefacts

Problem: typo in the zmq_has doc

Solution: fix it

add test and documentation for "draft" capability

Problem: cannot check if libzmq was build with draft api

As discussed in #2307

cmake: Set output name to libzmq for the libzmq-static target

Solution: Change the target output name from libzmq-static to libzmq.

Updated buildbase and buildall to work with visual studio 2017

cmake install on linux fixed

Problem: Thread-safe solution for modify hwm of pipe

Solution: where change pipe hwm, send a command (new type pipe_hwm) to peer, so peer pipe can modify hwm thread-safely

Problems: modifying pipe from different thread is not safe and HWM always boosted by 1

Solution: initialise *hwmboost to -1 instead of 1, and use it only if
it's >= 0. The socket option code checks anyway that the
user-provided value is >= 0 so there is no risk of clashing. The
documentation also specifies that it has to be >= 0.

Solution: revert change from #2299 to avoid bugs while we look for an
alternative and thread-safe solution

Problem: set hwm after connect lead to infinite hwm

Problem: spec file still depends on libuuid

Solution: remove the dependency as it's not been needed for a while

Problem: debian packages do not build manpages

Solution: build-depend on asciidoc and xmlto

Cmake win install

Windows cmake build fixed

Problem: peer can close connection before SO_NOSIGPIPE is set

Solution: setsockopt returns EINVAL if the connection was closed by
the peer after the accept returned a valid socket. This is a valid
network error and should not cause an assert.
To handle this we have to extract the setsockopt from the stream
engine, as there's no clean way to return an error from the
constructor. Instead, try to set this option before creating the
engine in the callers, and return immediately as if the accept
had failed to avoid churn. Do the same for the connect calls by
setting the option in open_socket, so that the option for that
case is set even before connecting, so there's no possible race
condition.
Since this has to be done in 4 places (tcp/ipc listener, socks
connecter and open_socket) add an utility function in ip.cpp.
Fixes #1442