This regression introducing during replacing port_ffi with port_buf
by 2fcce449

Linux-specific option obsolete since kernel 2.2

Wait() no longer called hence waiters not initialized. Does it make
sence to have a test for :waiters (undocumented internals)?

- Waiting on a brand new socket (neither listen nor connect called)
  completes immediately with HUP status on Linux. However on FreeBSD/osx
  it never completes.

- Attempting to connect to a port bound on a loopback interface which is
  not yet listening yield ECONNREFUSED in blocking mode on Linux. On
  FreeBSD/osx blocking connect never completes ("never" means it doesn't
  complete in a reasonable time).

- Calling shutdown() on a listening socket works in Linux but in
  FreeBSD/osx it doesn't.

Note: it was incomplete, made thing unnecessary complex and it did not
align with the CMake way (c) for crosscompilation.

Pass sysroot on OSX.
Make the code in luajit.cmake a bit more sane and straightforward.

Note: v1 of this patch was reverted due to performance regression.

Update reference to the pure php client

This reverts commit 615f0985.

This reverts commit 3010b98b.

Document UPSERT in the protocol spec.

iproto thread can add quirks to thread scheduling, replace
single-call yields with yields in a loop.

Conflicts:
	debian/changelog
	debian/control

 tarantool-lts-client (1.5) and tarantool (1.6)
 have file /usr/bin/tarantool

Build system fixes on osx, +misc.

Pass sysroot.
Abort build if preprocessing fails (a stage in tarantool.h generation).

Note: it was incomplete, made thing unnecessary complex and it did not
align with the CMake way (c) for crosscompilation.

Pass sysroot on OSX.
Make the code in luajit.cmake a bit more sane and straightforward.

We were passing -g flag twice, just uncool.

gh-1017 double-run bug fixed, and fixed same bug in stat_net. But bug with probability doesn't repeated

Make code in alter.cc less error prone and less verbose.
Miscellaneous cleanups.