SQL functions UPPER and LOWER now works
with COLLATE as they should according to
ANSI Standart.

Closes #3052.

SQLITE_ENABLE_PREUPDATE_HOOK is dead macro.
Remove it.

Part of #2356

To implement new TRUNCATE operation, we have introduced a
new P2 argument for OP_Clear opcode that calles box_truncate
instead of tarantoolSqlite3ClearTable.
This operation should work faster than DELETE FROM; but have
a few restricts.

Closes #2201.

@TarantoolBot document
Title: New TRUNCATE operation
TRUNCATE is DDL operation.
Removes all rows from a table or specified partitions of a table,
without logging the individual row deletions.
TRUNCATE TABLE is similar to the DELETE statement with no WHERE
clause; however, TRUNCATE TABLE is faster and uses fewer system
resources.
It couldn't be used with system tables or with tables having FKs.
It also couldn't be called in transaction.
The triggers on table will have ignored.
Example:
TRUNCATE TABLE t1;

This check happens twice. The patch simply removes check.

Get rid of is_primkey in Column structure as it becomes
redundant. Moved the last member coll with collation pointer
to field_def structure. Finally, dropped Column.

This patch dissallows define multiple "NULL", "NOT NULL"
options per column and fixes silent implicit behavior
for invalid "NULL PRIMARY KEY" construction.

Closes #3473.

Basic structures (struct Table/Index) of legacy SQL's data
dictionary used to so-called tnum to refer to engine's
btree structures. Tarantool used this field to store composition
of space_id and index_id. Recently both structures incorporated
native space_def/index_def descriptors.
This patch finally removes tnum field. It also refactors
init_callback machinery, removing varargs from declarations.

Closes #3482

Currently, the way txn_stmt::old_tuple and new_tuple are referenced
depends on the engine. For vinyl, the rules are straightforward: if
txn_stmt::{old_tuple,new_tuple} is not NULL, then the reference to the
corresponding tuple is elevated. Hence when a transaction is committed
or rolled back, vinyl calls tuple_unref on both txn_stmt::old_tuple and
new_tuple. For memtx, things are different: the engine doesn't
explicitly increment the reference counter of the tuples - it simply
sets them to the newly inserted tuple and the replaced tuple. On commit,
the reference counter of the old tuple is decreased to delete the
replaced tuple, while on rollback the reference counter of the new tuple
is decreased to delete the new tuple.

Because of this, we can't implement the blackhole engine (aka /dev/null)
without implementing commit and rollback engine methods - even though
such an engine doesn't store anything it still has to set the new_tuple
for on_replace trigger and hence it is responsible for releasing it on
commit or rollback. Since commit/rollback are rather inappropriate for
this kind of engine, let's instead unify txn_stmt reference counting
rules and make txn.c unreference the tuples no matter what engine is.
This doesn't change vinyl, because it already conforms. For memtx, this
means that we need to increase the reference counter when we insert a
new tuple into a space - not a big deal as tuple_ref is almost free.

By now, replace function takes new tuple and old tuple as arguments, instead
of single txn_stmt. It has been done in order to avoid abusing txn_stmt:
the only usage was extracting tuples from it.
As a result, this function can be used by ephemeral tables
without any patching.

(cherry picked from commit 880712c9)

They are fairly small and closely related so let's merge them and call
the result sysview.[hc].

This should reduce maintenance burden and help us introduce a new
engine.

It will be used to check if a replica fell too much behind its peers and
so needs to be rebootstrapped.

Needed for #461

Not needed anymore as we now use the new IPROTO_VOTE command instead of
IPROTO_VOTE_DEPRECATED. Let's remove it altogether and reuse its code
for IPROTO_BALLOT (they are never decoded together so no conflict should
happen). Worst that can happen is we choose a read-only master when
bootstrapping an older version of tarantool.

This patch contains some follow-up fixes for fe8ae607
("Introduce IPROTO_VOTE command"):
 - Rename 'status' to 'ballot' everywhere in the comments.
 - Rename IPROTO_REQUEST_VOTE to IPROTO_VOTE_DEPRECATED and
   iproto_reply_request_vote to iproto_reply_vote_deprecated
   to emphasize the fact that this iproto command has been
   deprecated and IPROTO_VOTE should be used instead.
 - Only send an IPROTO_VOTE request to a master if it is
   running tarantool 1.10.1 or newer.

The new command is supposed to supersede IPROTO_REQUEST_VOTE, which is
difficult to extend, because it uses the global iproto key namespace.
The new command returns a map (IPROTO_BALLOT), to which we can add
various information without polluting the global namespace. Currently,
the map contains IPROTO_BALLOT_IS_RO and IPROTO_BALLOT_VCLOCK keys,
but soon it will be added info needed for replica rebootstrap feature.

Needed for #461

Schema reference is useless. Remove it.

During tnum removal, convertToWithoutRowidtable became
redundant. Remove it, refactor addPrimaryIndex.
Also fix naming for un-named indexes.
Finally, remove iPKey from struct table and
update testsuite.

Part of #3482

Previously "BEGIN" / "BEGIN TRANSACTION", "COMMIT TRANSACTION" /
"END" / "END TRANSACTION", "ROLLBACK TRANSACTION" could be used
to handle transactions. By changing these commands syntax in
parser we're aiming on getting closer to ANSI SQL. Also found
initialization in ifdef that caused some problems with error
messages in occasions when the wrong syntax was used.

With the patch applied only following commands can be used:
 - "START TRANSACTION" to begin transaction.
 - "COMMIT" to end transaction.
 - "ROLLBACK" to rollback transaction without savepoints.
 - "ROLLBACK TO .." to rollback transaction to savepoint.

Closes #2164

To optimize the select-subquery tarantool uses
subquery flattening optimization which is only used
with respect to some restrictions. When any of
subselects but the last one has an ORDER BY
clause we should raise an error. So subquery
components containing it are not accepted
to be optimized.

Checking part was only checking if the last
subselect contains the clause which led to an
assertion fail. With the patch applied
flattening is not used in case any of
subselects has an ORDER BY.

Closes #3353

_say function was called with invalid arguments.
Thank @sorc1 for patch.

Closes #3433.

Strdup may silently fail without any message from tarantool.
Patch adds this checks.

Mode logging to syslog was broken due to assertion in case
ov vaild configuration:

  box.cfg{log = 'syslog:identity=tarantool'}

Closes #3502

Strings containing "true" and "false" were converted
to a boolean type when serializing. Fixed.
Example:
type(yaml.decode(yaml.encode('false'))) == string
type(yaml.decode(yaml.encode('true'))) == string

Closes #3476.

Function tonumber64 has worked incorrectly with values less
than INT64_MAX.
Now it works in the interval [INT64_MAX, UINT64_MAX] returning
nil otherwise.

Closes #3466.

Net.box didn't pass options containing iterator to
server side.
There were also invalid results for two :count tests in
net.box.result file.

Thanks @ademenev for contributing problem and help with
problem locating.

Closes #3262.

Parser didn't raise error when DELETE FROM can't find
a table specified on statement.
As no error was raised, partially-prepared VDBE code
seemed to be generated fine, that was not true.

Closes #3535, #3511.

Before the patch, big integer constants were silently
converted to floating point values. Fix that by issuing
error message if value doesn't fit in int64_t range.

Also, refactor surrounding code as per Tarantool's code style.

Closes #2347

Currently, this function takes a single boolean argument, but I'm
planning to add another one. Since two bool arguments look rather
confusing, let's turn this arguments into flags.

Needed for #461

A small cleanup to avoid a potentially inefficient use of
gc_consumer_vclock().

Follow up on a patch for gh-461.

In order to check if a replica needs to be rebootstrapped, we need to
know the vclock of the oldest WAL stored on the master, but the garbage
collector works with signatures and hence can't report the vclock it was
last called for. Actually, all gc users have a vclock and can pass it
instead of signature so it's pretty easy to switch garbage collection
infrastructure to vclock.

Needed for #461

We will need it in other places.

 - Remove extra scan of the WAL directory from local_recovery() - we
   scan the directory in recovery_end_vclock() hence we can skip scan in
   recover_remaining_wals() by passing scan_dir = false.

 - Rename recovery_end_vclock() to recovery_scan() to emphasize the fact
   that this function scans the WAL directory. Write a comment to this
   function.

 - Add comments to wal.c explaining why we scan the WAL directory there.

Follow-up 0695fbbb ("box: retrieve end vclock before starting local
recovery").

To bring crash_expected option of "start server" command.

There were no checks for granting and revoking a non-existent
privilege or a privilege to a non-existent entity.
Added the checks, and a test case.

Closes #3417