Spelling fixes in the error message file.

Add more information to the error when key cardinality
doesn't match index cardinality in exact match.

Avoid server start/stop in box_big/hash.test (ugly,
but until we have named/dynamic spaces, which are already in the
pipeline, we should not make our test procedure slower).

Rename multipart-pk.test to tree_pk_multipar.test:
consistently use underscores, not dashes, in file
names.

(Unrelated) Recycle more error codes.

Make keyValue an inline function to save a few CPU cycles.

Avoid creation of an extra table in box.lua just to find out
variable list array cardinality.

Revert changes to box.update():
a) they break backward compatibility

This reason alone should be sufficient. There must be no changes
which break backward compatibility in a stable release, unless
it's a fix for a crashing/security bug.

box.update() changes got to be done in a separate task,
and should not break backward compatibility.

Leave update_ol for now, to test multi-part primary key updates,
even though the function itself is quite heavy, hardly something
we can recommend while running Lua procedures in production.

This reverts commit 8dc70aac.

Make sure we do not try to destroy uninitialized
mutexes in the snapshot thread.

Implement FIFO policy for the wal writer queue.

Code review fixes:
     - Key validation procedure was moved to Index classes;
     - Additional hash tests was added.

This patch introduces wal_mode configuration variable,
which can take either "fsync" or "fsync_delay" values.

In "fsync" mode, we open the write ahead log in O_SYNC
mode, and thus sync every write to disk.

Th fsync_delay mode is the backward-compatible mode:
in this mode, we only do fsync once every fsync_delay
seconds. If fsync_delay is 0, there is no fsync.

Otherwise, this patch is only adding auxiliary headers,
object files, etc.

Actually, make the feature work.
Add syntax to Lua to be able to delete fields.
Fix a bug when we delete all fields in a tuple and happily
crash thereafter.

Multipart primary key implementation was added. DELETE and UPDATE commands was
updated for multipart key supporting.

(https://bugs.launchpad.net/tarantool/+bug/955226)

A fix for Bug#928222 Memory leak when running a long
stored procedure.

An original patch was prepared done by Alexey Demakov.

When running a long stored procedure, the server could
allocate a lot of memory in the fiber garbage collected
pool, when performing requests in a loop.

To fix the bug, garbage collect memory after every call to
rw_callback() from stored procedures module.

Add ptruncate() method to palloc, which rolls back
used memory to a saved position and use it in box_lua.m.

Since VALGRIND MEMPOOL API doesn't support ptruncate(),
don't use it, instead consistently use
VALGRIND_MAKE_MEM_DEFINED/VALGRIND_MAKE_MEM_UNDEFINED.

No test case since the problem is detected by a longevity
test, not a functional test.

- removed SIGUSR2 forwarding
- removed logger and pid command line arguments
- moved man pages generation to doc/man/
- added logger_pid to show info output
- changed instance config preformating scheme

- release_disabled test suite ini file option support
- debug build only error-injection test, cmake build cleaner output
- tarantool server version output with build flags

A fix and a test case for Bu#929654 "
Regression with secondary HASH indexes"

Secodnary HASH indexes where skipped by build_indexes()
when starting from a snapshot.

Add a way to reserve space for as many values as necessary
and enable HASH indexes in build_indexes().

Add a test.

Implement blueprint 'add-delete-fields':
https://blueprints.launchpad.net/tarantool/+spec/add-and-delete-fields-in-tuple

Make it possible to add new fields with UPDATE, and
delete existing fields.

Rewrite the UPDATE algorithm to pre-compute new tuple
length and do most of the work in place, without
allocating temporary buffer for fields.

Add Connector/C support for new UPDATE options.

Add and update tests.

Joint effort by Konstantin Shulgin (initial implementation),
Dmitry Simonenko (Connector/C) and
Konstantin Osipov (spec/code review).

(https://blueprints.launchpad.net/tarantool/+spec/compile-with-external-luajit)

lua:
box.fiber.status function was added. The function returns current fiber
status, like a cocoroutine.status() function in lua standard library.

core:
fiber_is_caller function was added. The function returns true if
checking fiber is caller of current working fiber.

ENUM0 macro.

* SIGUSR2 handler sends SIGHUP to logger
  (default extra/logger could already handle SIGHUP).
* Add some options to cmd-line (to override config's options).