Solution:
- call zmq_msg_close() if there is an error
- call free() to release resources if there is an error (CWE-404)

Problem: check_protocol() logic duplicated twice and needs cleanup

Problem:
Conditional logic in check_protocol() that checks if a protocol is supported,
is duplicated twice. Moreover, the first set of checks to ascertain if a
protocol is supported is done regardless of whether the particular protocol
will be built into the library or not.

Solution:
* Simplify/collapse all supported protocol checks into one in check_protocol()
* Enclose pgm/epgm/norm socket+protocol match checks with requisite macros

Problem: errno not set if UDP is NOT used with ZMQ_RADIO or ZMQ_DISH

Solution: set errno to ENOCOMPATPROTO in check_protocol()

Problem: Running Linux-only tests results in 10% of tests failing on non-Linux systems (OSX, et. al.)

Move tests specific to Linux under a platform conditional thereby eliminating
unnecessary builds and fixing "make test" on Mac OS X and possibly other
non-Linux systems.

Tests specific to Linux:
    - abstract namespace support for AF_UNIX sockets
    - TIPC support (AF_TIPC)

Test success rate jumps from 90% to 100% on Mac OS X after this change.

Fix typos similar to #1894

Fix typo

Solution: make the same change for the ZMQ_POLL_BASED_ON_SELECT case

Solution: correct it to spaces

Solution: disable pedantic and Werror, as libsodium headers use
pragma diagnostic which are not available in gcc 3.4.

Solution: pass a struct addrinfo hint to getaddrinfo with a hint
about the address family to avoid a failure.

Solution: check for GCC version before using pragma diagnostic
in src/tweetnacl.c to avoid an additional warning.

Solution: add Wnosign-compare to CPPFLAGS. GCC version is too old on
Solaris 10 to use pragma diagnostic (min 4.2 needed, but 3.4 is
available)

Fix return value of zmq_poller_wait when used on empty poller

Solution: return -1 (no event) instead of 0 (event)

For some reason, this just returns 0 if there are no sockets registered
on the poller. Usually this would mean there has been an event. So the
caller would have to check the return value AND the event, or write code
that takes the number of registered sockets into consideration.

By returning -1 and setting errno = ETIMEDOUT like in the usual timeout
cases, it's more consistent and convenient.

Test case included.

Solution: use zmq_assert to ensure it's not a nullpointer

Solution: if options.use_fd do not create temporary random
directory for ipc://*, since the socket is already created and
passed to the library by the user.

Solution: check for availability in autoconf and cmake, and if not
available fall back to random file name rather than random directory.

Solution: use the less nice but correct int constant 1000000000
instead of the shorter 1E9 to avoid a compiler warning when assigning
to timespec.tv_nsec, which is a long int.

Problem: zmq_bind IPv4 fallback still tries IPv6

Solution: in the Windows-specific ifdef in tcp_listener set_address,
check for error and set errno only after the IPv4 fallback has failed
too, to avoid setting errno when the socket creation succeeds through
the fallback.

Solution: if opening an IPv6 TCP socket fails because IPv6 is not
available, try to open an IPv4 socket instead when creating and
connecting a TCP endpoint.

Solution: if opening an IPv6 TCP socket fails because IPv6 is not
available, try to open an IPv4 socket instead when creating and
binding a TCP endpoint.

Problem: including recent "zmq_utils.h" crashes gcc -pedantic jobs

Problem: Recent deprecation of the "zmq_utils.h" header file caused pedantic compilations (including czmq) to fail because non-portable #warning is used.
Solution: Limit the deprecation warnings to compilers known or assumed to support the "#pragma message" (GCC, MSVC, CLANG) and wrap with GCC directives to not treat these warnings as errors on paranoid builds.

It says that ZMQ_SERVER never blocks; whereas it does block on
sending to a client whose pipe is full.

Solution: fix it.

Fix NetBSD thread scheduling problem.

Defining thread priority for SCHED_OTHER is implementation defined.
Some platforms like NetBSD cannot reassign it as they are dynamic.

<http://pubs.opengroup.org/onlinepubs/9699919799/functions/pthread_getschedparam.html>

Credit goes to <kamil@netbsd.org> and <riastradh@netbsd.org> for finding this
solution.

Fixed ZMQ_REQ_CORRELATE (see pull request #1730)

Problem: Since pull request #1730 was merged, protocol for REQ socket is
checked at the session level and this check does not take into account
the possibility of a request_id being part of the message. Thus the option
ZMQ_REQ_CORRELATE would no longer work.
This is now fixed: the possiblity of a 4 bytes integer being present
before the delimiter frame is taken into account (whether or not this
breaks the REQ/REP RFC is another issue).

Fixes #1419 segfault when using a XSUB/XPUB proxy.