1)Added 'slab_alloc_maximal' config option.
2)Tuple slab cache is creating now with a range of orders depends on
  'slab_alloc_minimal' and 'slab_alloc_maximal' config options.
3)Smallocator's mempools now selects proper order of slab depends on
   size of an object.
4)after CR fixes

Set missing trailing zero in find_path().

Rewrite snapshot_daemon to use fiber.channel() instead of
undocumented API.

Explicitly mark places which are unsafe for cancel.

Remove the snapshot in abort_checkpoint(). Use coeio_remove/coeio_unlink
bindings instead of coeio_custom().

2)minor fixes of load_cfg.lua

Everything is cancellable. In fiber_cancel() use fiber_wakeup()
instead of a direct call. This make cancel safe with
regards to wakeup() and timeouts: i.e. it's safe to cancel
a fiber which is waiting on a timeout or is about to get
scheduled.

Avoid an unnecessary interlocked instruction or pipe write
in fiber_wakeup(), it's always called from the same cord
as the one to which the event loop belongs.

Use ev_feed_event() with EV_CUSTOM instead.

Rename Factory to Engine, Factory was a name we couldn't swallow.

Introduce Handler class to disambiguate Engine from Space
(hello, MySQL).

Remove the branch which saves a snapshot without a fork.