After patch that introduced "none" collation (a953051e),
box.internal.bootstrap() started to fail due to inability to drop
mentioned collation. Lets turn off system triggers for _collation space
in order to process its complete purification during bootstrap.

This patch introduces two new collation sequences: "none" and "binary".
Despite the fact that they use the same comparing algorithm (simple
byte-by-byte comparison), they don't mean the same. "binary" collation
get to the format if user explicitly points it: either specifies this
collation in space format manually or adds <COLLATE BINARY> clause to
column definition within CREATE TABLE statement. "none" collation is
used when user doesn't specify any collation at all. "none" collation always
comes with id == 0 and it can't be changed (since its id vastly used
under the hood as an indicator of absence of collation).
Difference between these collations is vital for ANSI SQL: mixing
"binary" with other collations is prohibited, meanwhile "none" collation
can be used alongside with others.
In this respect current patch extends list of available collations: now
not only ICU collations are allowed, but also BINARY.

Note, that in SQL some queries have changed their query plan. That
occurred due to the fact that our parser allows using <COLLATE> clause
with numeric fields:

CREATE TABLE (id INT PRIMARY KEY);
SELECT id COLLATE "binary" ...

In the example collation of LHS (id column) is NULL, but collation of
RHS is "binary". Before this patch both collations were NULL. Hence,
usage of certain indexes may not be allowed by query planner.
On the other hand, this feature is obviously broken, so that doesn't
seem to be big deal.

Needed for #3185

The space is a blackhole. It will be used for writing deferred DELETE
statements generated by vinyl compaction tasks to WAL so that we can
recover deferred DELETEs that hadn't been dumped to disk before the
server was restarted.

Since the new space doesn't depend on other system spaces, let's assign
the minimal possible id to it, i.e. 257.

Needed for #2129

When granting or revoking a privilege on an entire entity, id 0 was used
to indicate the fact that we don't grant a privilege on a single object,
but on a whole entity. This caused confusion, because for entity USER,
for example, id 0 is a valid object id (user 'guest' uses it).
Any non-zero id dedicated to this cause obviously may be confused as well.
Fix this by creating separate schema_object_types for entities:
SC_ENTITY_SPACE, SC_ENTITY_SEQUENCE, etc.

Closes #3574
Needed for #3524

This patch introduces new system space to persist foreign keys
constraints. Format of the space:

_fk_constraint (space id = 358)

[<constraint name> STR, <parent id> UINT, <child id> UINT,
 <is deferred> BOOL, <match> STR, <on delete action> STR,
 <on update action> STR, <child cols> ARRAY<UINT>,
 <parent cols> ARRAY<UINT>]

FK constraint is local to space, so every pair <FK name, child id>
is unique (and it is PK in _fk_constraint space).

After insertion into this space, new instance describing FK constraint
is created. FK constraints are held in data-dictionary as two lists
(for child and parent constraints) in struct space.

There is a list of FK restrictions:
 - At the time of FK creation parent and child spaces must exist;
 - VIEWs can't be involved into FK processing;
 - Child space must be empty;
 - Types of referencing and referenced fields must be comparable;
 - Collations of referencing and referenced fields must match;
 - Referenced fields must compose unique index;
 - Referenced fields can not contain duplicates.

Until space (child) features FK constraints it isn't allowed to be
dropped. Implicitly referenced index also can't be dropped
(and that is why parent space can't be dropped). But :drop() method
of child space firstly deletes all FK constraint (the same as SQL
triggers, indexes etc) and then removes entry from _space.

Part of #3271

Review fixes

As we would like to lookup triggers by space_id in future
on space deletion to delete associated _trigger tuples we
need to introduce new field space_id as secondary key.

Part of #3273.

Merge upgrade to 1.8.2/.4 into 2.1.0. Versions 1.8.2/.4 are
broken because of invalid _trigger creation and deprecated
versioning policy.

Vinyl/Xlog upgrade tests from 1.7.7 are broken too since they
actually contains 1.8 snapshots, so remove them. Add the test on
upgrade from 1.10 to ensure the new upgrade_to_2_1_0 works.

Introduce _vsequence system space.

Prerequisite of #3250

_schema represented as key-value storage for various
values common for Tarantool, like next id for space creation.
SQL requires format to be fully specified for columns being
access. Unfortunatelly, _schema is inserted into _space
before _space's format is set and since DD triggers are
disabled during upgrade, format for _schema tuple in _space
stays the same. So, set value nullable field in upgrade,
regenerate initial snap, update tests.
Also, as far as _schema's tuple in _space is not updated:
relax fieldno check in sql.c

If Tarantool is being starteed using data from older
version, where stat tables were not system, it will definitely
crash, since SQL statistics analyzer will try to access non-
existing statistics tables (which will be added during upgrade).

To avoid such crash: create dummy statistics spaces during
schema init.

Also, fix types for statistics space format: prohibit null values
for statistic values and make sample format stronger.

  Making `stat[14]` a system spaces enables us to allow users to create
spaces which starts with `_`, because it cannot affect internal state
anymore.
  Since now the only constraints on table names is:
    - consists of printable symbols only
    - length < 65k

Closes #2126

Add a role which contains all ACLs.

Fixes gh-3022.

Useful for any quick start:

box.schema.user.grant('guest', 'super')

Add system privileges 'session' and 'usage'

* 'session' privilege lets user connect to database server
* 'usage' privilege lets user use his/her rights on database objects
* Both privileges are assigned to all users by default.

Implementation details:
* system privileges are special grant rights to 'universe'.
Therefore, they can be granted only by admin. Because of this fact,
during creation or deletion of user, we have to switch to 'admin' to
grant or revoke these rights.

Important changes:
* changed bootstrap.snap due to need to start admin with new privileges
* added auto upgrade script for 1.7.7

Fixes gh-2898.

With contributions by @kostja.

+ Don't use id=0 for collations

Follow up #2649

Add a new space that contains collation definitions for future
index collation and more.

Add a collation cache for fast retrieving a collation object by
its name.

Needed for #2649

The value returned by space.auto_increment() is not guaranteed to be
monotonically growing, because this method uses max+1 for the next
generated value, which can be less than a value generated before in case
the max key is deleted. Besides, calling max() may be expensive, e.g. it
requires a disk access in case of Vinyl. So this patch implements auto
increment based on persistent sequences.

To be used for auto increment, a sequence needs to be attached to a
space. To attach an existing sequence to a space, pass 'sequence=ID' in
index options when creating or modifying the primary key. Here ID is the
name or the numeric identifier of the sequence to attach. Note,
sequences can be attached only to spaces with integer primary key.
After a sequence is attached, it will be used for auto increment when
the user passes 'nil' for the primary key value. If the user passes a
value different from 'nil', the value will be used to update the
attached sequence.

To detach a sequence from the space it was attached to, either drop the
primary key of the space or alter the primary key with 'sequence=false'.

It is not necessary to pre-create a sequence to use it for auto
increment - a sequence can be automatically generated. To generate a
sequence automatically, pass 'sequence=true' in index options.
Automatically generated sequences are named '_auto_ID', where ID is the
space id, and removed when the space is dropped or the sequence is
detached.

A sequence cannot be dropped as long as it is attached to any spaces.

Closes #389

This patch implements a new object type, persistent sequences. Sequences
are created with function box.schema.sequence.create(name, options).
Options include min/max, start value, increment, cache size, just like
in Postgresql, although 'cache' is ignored for now. All sequences can be
accessed via box.sequence.<name>, similarly to spaces. To generate a
sequence value, use seq:next() method. To retrieve the last generated
value, use seq:get(). A sequence value can also be reset to the start
value or to any other value using seq:reset() and seq:set() methods.

Needed for #389

Needed for #2754

Now all upgrade_to_1_... can be removed, except 1_8_2(). Inital
function initializes schema for 1.7.5, not for 1.6.0.

We can not support upgrade directly from <1.7.5, because these
versions contains space format conflicts, which are checked in 1.8.

Needed for #2652

* upgrade handler for 1.8.2 must be set to auto instead of
  1.7.5;
* _trigger's first field has type string, but index on this field
  has type unsigned - it is conflict.

Needed for #2652

Now system space formats contain conflicts between space:format,
their index parts and actual tuple fields.

Such formats can not be used to check tuple fields in SQL.

Fix them in order of #2652

_truncate space write access is required for space truncation.
Add this privilege to bootstrap.snap.

See #2674

Currently, space.create must insert an entry into 'truncate' system
space - if it does not, space.truncate won't work. This is incontinent,
as it makes it impossible to create spaces by simply inserting a tuple
into the 'space' system space via iproto API. So let's insert entries
into truncate space lazily, on the first space truncation.

Closes #2524

We added _truncate space to 1.7.5 and we are going to add new system
spaces for storing sequences and triggers. Without upgrade, the
corresponding operations won't work. Since 1.7.5 is a minor upgrade,
users may not call box.schema.upgrade(), so we need to call it for them
automatically. This patch introduces infrastructure for automatic
upgrades and sets upgrade to 1.7.5 to be called automatically.

While we are at it, rename schema version 1.7.4 to 1.7.5 (1.7.4 has
already been released).

Closes #2517

Persistency for SQL triggers supported by introducing
new system space called `_trigger` which should store all
SQL triggers registered. Add space itself, make
visible to Lua, update bootstrap code and regenerate
bootstrap.snap.
Insertion into the space is done by introducing new VDBE
opcode `OP_ParseSchema3` which provides a pair of trigger
name and SQL which creates the trigger.
During system restore `_trigger` is scanned and its contents
parsed back into VDBE.
Update tests and extend test for persistency w/ trigger case.

Closes #2320

Space truncation that we have now is not atomic: we recreate all indexes
of the truncated space one by one. This can result in nasty failures if
a tuple insertion races with the space truncation and sees some indexes
truncated and others not.

This patch redesigns space truncation as follows:

 - Truncate is now triggered by bumping a counter in a new system space
   called _truncate. As before, space truncation is implemented by
   recreating all of its indexes, but now this is done internally in one
   go, inside the space alter trigger. This makes the operation atomic.

 - New indexes are created with Handler::createIndex method, old indexes
   are deleted with Index::~Index. Neither Index::commitCreate nor
   Index::commitDrop are called in case of truncation, in contrast to
   space alter. Since memtx needs to release tuples referenced by old
   indexes, and vinyl needs to log space truncation in the metadata log,
   new Handler methods are introduced, prepareTruncateSpace and
   commitTruncateSpace, which are passed the old and new spaces. They
   are called before and after truncate record is written to WAL,
   respectively.

 - Since Handler::commitTruncateSpace must not fail while vylog write
   obviously may, we reuse the technique used by commitCreate and
   commitDrop methods of VinylIndex, namely leave the record we failed
   to write in vylog buffer to be either flushed along with the next
   write or replayed on WAL recovery. To be able to detect if truncation
   was logged while recovering WAL, we introduce a new vylog record
   type, VY_LOG_TRUNCATE_INDEX which takes truncate_count as a key: if
   on WAL recovery index truncate_count happens to be <= space
   truncate_count, then it it means that truncation was not logged and
   we need to log it again.

Closes #618
Closes #2060

Check magic in xlog_cursor_read_tx.

Issue #1720

* Rename fields types in data dictionary and test suite:
 - 'num' => 'unsigned'
 - 'str' => 'string'
 - 'int' => 'integer'
* Force coversion to lowercase
* Add aliases to space:create_index():
 - 'str' => 'string'
 - 'num' => 'unsigned' -- logs "field type 'num' is deprecated" warning
 - 'uint' => 'unsigned'
 - 'int' => 'integer'
* Add prefix to `enum field_type` members
* Add upgrade_to_1_7_2() script

Fixes #942
Fixes #1534

This is required to do the following:

tarantoolctl user@host
tarantoolctl guest@host
tarantoolctl guest:@host

Currently all of those will lead to error from net.box. Either because
password is not specified, or because guest user can't login with empty
password.

This change effectively sets guest password to an empty string, to make
it less of a special case to authentication.

And since now guest can be authenticated with an empty password, net.box
will replace nil passwords with empty strings, when user name is
provided.

We still support the case when username=nil and password=nil. In this
case net.box doesn't send authentication packet at all. Guest login is
just assumed.

The bootstrap snapshot will now have guest user with an empty password,
instead of no password.

Fixes #1545

* Set version to {1, 6, 8} in box.space._schema
* Add 'LUA' language to built-in stored procedures

Make more abstract version of key_opts parser to re-use code
for space options.

* 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

Add _vspace, _vindex, _vuser, _vfunc and _vpriv virtual spaces (views)
which contains only objects allowed to the current user.

This patch is needed for client connectors to work under non-admin user.

Extract xlog.test.py into an own test suite, in preparation
for a split into many tests.
xlog.test.py restarts the server all the time anyway, there is
no point to keep all tests in the same file.

Deal with collateral damange of this seemingly innocuous move:

 * fix an assertion failure in user.drop(), user_has_data().
   The check was using a wrong index.
 * fix test-run.py to leave cores around in case of a server crash
   (in a friendly gesture to develoeprs, and joining the ranks of
   such extremely succefful applications as apport and systemd, it used to
   delete them)
 * update bootstrap.snap to provide format for _index space
 * make a bunch of tests safe to the order in which they are executed
 * use space name rather than space id for error messages more often,
   since space id is very volatile after introduction of max_id sequence
   for space ids.

Use a monotonically growing sequence for space identifiers,
to ensure they are never reused.

Update tests.

Update tests to use this role.

Introduce 'public' role.
Auto-grant 'public' to every user.
Update the snapshot.