The allocator consists of a common slab cache, region and slab
allocators.
The patch provides unit tests.

It also modifies rlist_shift to be more in line with the
rest of rlist functions.

Conflicts:
	third_party/luajit

Before introducing a stronger row handler API,
eradicate tbuf from log_io/recovery, and pass around
a pointer to row and row length instead.

Extract the remaining places which access tuple->data
directly into tuple_to_port.cc file, which will have full
access to all tuple formats and all destinations to which
a tuple may need to be converted.

In this file it'll be possible to perform a conversion
efficiently, at the same time, it will be the only place
of cross-dependency between all tuple formats and all
conversion destinations.

Conflicts:
	cmake/compiler.cmake
	include/sio.h
	src/box/bitset_index.cc
	src/box/box_lua.cc
	src/lua/init.cc
	src/lua/session.cc
	src/sio.cc
	test/unit/objc_catchcxx.m
	third_party/luajit

Conflicts:
	third_party/luajit

* 	Implementing LUA printer (https://blueprints.launchpad.net/tarantool/+spec/tc-lua-printer)
* 	Moving xlog printer code into tc_print_xlog.{c,h}, leaving basic printing code in tc_print.{c,h}.
* 	Fix TC bug (https://bugs.launchpad.net/tarantool/+bug/1185746), Added support for snap print formatters.
* 	Added temporary workaround for option (-B or --bin), that allow you to print STR instead of INT and INT64 in lua formatter.
	Works in "keys" and UPDATE (ASSIGN or INSERT).
*	Fix of stdin: add custom function for pipe instead of stdin (very slow with readline)
*	Fix for realloc problems
	1) 	if size is 0, then behaviour is undefined
	2) 	in C99 if realloc returns NULL, and size is not zero (errno == ENOMEM) input buffer of memory is needed to be freed manually.
		Whitepaper of POSIX standart:
		http://pubs.opengroup.org/onlinepubs/9699919799/functions/realloc.html
*	Add support of -D (--delim) argument, that sets delimiter on start of tarantool client.
	Also, if you "cat" .snap or .xlog files in LUA format it prints delim in end of every coomand.

Derive some tests for multipart HASH indexes
from the original implementation by A. Demakov
before dropping the branch.

- simplify the comparison function, only shortuct the single-part
  NUM key case.
- tweak mh_find() declaration
- remove wrappers around hash values
  and keys

Don't drop entire vardir on test suite start, it may have
symlinks, configuration files, etc. - useful for debugging.