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.

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.

Fixed a pair of minor bugs:
1)Fixed multidim_test, it could occasionally crash
2)Fixed rtree_misc.test.lua, snapshot test part
3)Reverted another rtree key mode -
  ind:select{0, 0, 10, 10} and
  ind:select{{0, 0, 10, 10}} - both works now

* make sure long-run tests pass
* more systematic approach to downcasts to MemtxIndex in
alter/system spaces code

Support multi-part keys. Allow key duplicate
compaction. Reduce storage cost, encode metadata,
do not store msgpack (except value part).
Reconstruct tuple on read using index schema.

[replication]

JOIN support temporary put to not working
condition: there is no way to access possibly
deleted space schema during the operation
for a tuple reconstruction
(multi-thread).

recreate space key_def during join; enable join

sophia: make every operation asynchronous

sophia: fix join cursor and tests

sophia: implement delayed update operation

sophia: make delete write-only

implement Handler::executeReplace

sophia: merge fixes

make upsert operation an engine specific

sophia: make update logic common to memtx

sophia-integration: switch to Sophia v1.2.3; implement upsert

sophia: add half_commit mode support

This mode disables conflict resolution for
'prepared' transactions and solves the issue with
concurrent write-write conflicts during wal write/yield.

sophia: do not allocate tuple during replace or update; refactoring

sophia: extend and reorganize test suite

sophia: implement box.sophia()

sophia: #681: support or produce error on space::alter

sophia: extend recover tests

sophia: add separate iterator tests

box: enable per-engine select

    * add executeSelect() into engine handler API
    * add MemtxIndex; move preallocated m_position from Index to MemtxIndex
    * remove it->close()
    * remove IteratorGuard class

    note: Sysview and Sophia engines yet share the same select code in
    engine::executeSelect() (which does iterator alloc-free)

box: remove close call from box_iterator_free()

Reason: net.box depends on src/box/lua/tuple.h, src/box/lua/schema.lua,
src/box/lua/call.cc and these dependencies impossible to fix.

box.cfg{} is not required for net.box.

* struct stat_counter keeps all statistics data (supports several instances)
* test sends 100 rps, but this methods count only 80
* stat counting bug fixed
* style fixes
* renamed as stat.c to rmean.c

* msgpack encoding with lua c api(call, insert, eval, ping)
* remove extra hash lookup in net.box
* review fixes for netbox-write
* rewrite all encoding fucntions using Lua/C
* replace channels with fiber.sleep()/fiber.wakeup()
* review fixes (remove an unnecessary include)

This patch adds four new functions to API:

  * luaL_checkuint64()
  * luaL_touint64()
  * luaL_checkint64()
  * luaL_toint64()

This functions are needed to rewrite request encoding in net.box.

With structured configuration, i.e. a Lua table for a list of
replication peers, it is difficult to pass a configuration option
to the hook function as a single C object. Don't pass
options in, but query them from the hook function with cfg_get*
API.

As a result, all "change" hooks have the same signature now,
a step closer to the pluggable configuration API.

Introduce a layer of wrappers for a number of internal box functions,
dealing with accss to spaces and indexes.

These wrappers:
    * don't throw exceptions
    * have a common prefix box_
    * are exported in the server development headers

Rewrite Lua C bindings to use the public API described above.
Rewrite Lua FFI bindings to do the same.
Add test.

Conflicts:
	src/box/alter.cc
	test/app/snapshot.test.lua

* style fixes
* fix a bug in key opts decoder (alter.cc)
* ensure dimension is visible in box.space
* rebuild bootstrap.snap to use the new format

- fix net.box to satisfy strict.lua requirements
- after review fixes
- fix buffer:checksize() in buffer.lua

Reflect in slab stats the quota size and index memory statistics.

arena_used_ratio - how much of arena is used
items_used_ratio - how much memory is actualy used for items,
given that there is some fragmentation. Doesn't reflect index pages
arena_used - how much of the quota is used (i.e. how much memory
is used for index and tuple data)
arena_size  - this is the same as box.cfg.slab_alloc_arena.
quota_size - how much of the quota is defined (box.cfg.slab_alloc_arena)
quota_used - how much of the quota is used (reflects the memory taken
by all slabs, including slabs in caches)

Incompatible change:
Extract detailed slab statistics to a separate call, box.slab.stats()

- upsert semantics is update or insert, not "always update"
- upsert doesn't return tuple
- bind upsert as tuple method in Lua (useful for QA)
- ensure upsert is properly encoded in xlog

Implement a generic inter-cord communication bus and use it
in separating network I/O and transaction processor.

Split all request processing work into network thread part
and transaction processor thread part, and make these
parts multi-threading safe.

Generalize fiber pool into an own data structure and API.

Misc cleanups.

*) refactoring of update operation
*) added tests