Before this patch, the existence of an index for a DROP INDEX statement
was checked at runtime. This is quite inconvenient for #4467, so this
will now be checked when creating the VDBE.

Needed for #4467

NO_DOC=no user-visible changes
NO_TEST=no user-visible changes
NO_CHANGELOG=no user-visible changes

This patch refactors functions in pragma.c and fixes an issue with
incorrectly ordered arguments in sql_pragma_index_info().

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

This patch removes the EP_InfixFunc, EP_DblQuoted and EP_Alias flags as
they are not used.

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

This patch removes the sql_id_eq_str_expr() and
vdbe_emit_stat_space_clear() functions as they are no longer used.

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

Since tuple format cannot be created in IPROTO thread, we only save the
tuple format data and later create a tuple format map from it when
processing `call` and `eval` requests.

Closes #8633

@TarantoolBot document
Title: Tuple formats in IPROTO

New MsgPack extension was added, `MP_TUPLE`, which is used exclusively
in IPROTO. It has the following structure:
+----------+-----------------+----------+
|MP_EXT = 7|      MP_UINT    | MP_ARRAY |
+----------+-----------------+----------+
            ^                 ^
            format identifier tuple data

New IPROTO features were added:
* `IPROTO_FEATURE_DML_TUPLE_EXTENSION = 7` — tuples in `IPROTO_DATA` are
   encoded as `MP_TUPLE` and tuple format is sent in response to DML
   requests.
* `IPROTO_FEATURE_CALL_RET_TUPLE_EXTENSION = 8` —tuples in `IPROTO_DATA`
  are encoded as `MP_TUPLE` and tuple format is sent in response to `call`
  and `eval` requests.
* `IPROTO_FEATURE_CALL_ARG_TUPLE_EXTENSION = 9` — tuples in `IPROTO_TUPLE`
  of `call` and `eval` requests are encoded as `MP_TUPLE` with their
  formats encoded as described below.

The `IPROTO_FEATURE_DML_TUPLE_EXTENSION` feature can be enabled in
`net.box` by setting the `fetch_schema` option to `false` when creating a
new connection.

A new IPROTO key for bodies of responses with data and bodies of `call` and
`eval` requests: `IPROTO_TUPLE_FORMATS = 0x60` — a `MP_MAP` consisting of
`MP_UINT` keys (format identifiers) and `MP_STRING` values (tuple format
clauses).

A new backward compatibility option was added (and a new page,
`https://tarantool.io/compat/box_tuple_extension`, needs to be created for
it): `BOX_TUPLE_EXTENSION_BRIEF` — it controls the
`IPROTO_FEATURE_CALL_RET_TUPLE_EXTENSION` and
`IPROTO_FEATURE_CALL_ARG_TUPLE_EXTENSION` feature bits. It was added for
clients that assume to receive `MP_ARRAY`s in `call` and `eval` requests
arguments and response data.

Please refer to the design document for [details](https://www.notion.so/tarantool/Schemafull-IPROTO-cc315ad6bdd641dea66ad854992d8cbf?pvs=4#7d86000c69cb4f8087a289f2caad73a0).

Extend the MsgPack port to store a pointer to a MsgPack context (the
lifetime of a MsgPack port is static, so we simply allocate it on the
stack), avoiding copying. This context can later be used for dumping
MsgPack to Lua, and possibly acquired by a MsgPack object when the
'MsgPack object' mode of `port_dump_lua` is used.

Needed for #8147

NO_CHANGELOG=refactoring
NO_DOC=refactoring
NO_TEST=refactoring

Tuple formats are sent in a separate IPROTO_TUPLE_FORMATS response body
field, which is namely a mapping from format identifiers to serialized
tuple formats.

The algorithm for parsing IPROTO response body is the following:
1. Parse tuple formats and build a tuple format index.
2. Parse data passing the tuple format index in the decoding context.

Add new backward compatibility option `box_tuple_extension` which controls
the server's `IPROTO_FEATURE_CALL_RET_TUPLE_EXTENSION` bit: needed for
users that don't expect the new tuple extension to be returned in call and
eval responses instead of basic MsgPack types. This option will
also be used for `IPROTO_FEATURE_CALL_ARG_TUPLE_EXTENSION` bit with the
same purpose.

Closes #8147

NO_DOC=<added in a later commit together with documentation of tuple
        formats in IPROTO `call` and `eval` arguments>

In scope of tarantool/tarantool#8147, a new context-dependent extension for
box tuples, `MP_TUPLE`, is introduced. The buffer source uses a buffer with
raw MsgPack, which does not allow for passing the context required for
decoding `MP_TUPLE`, so, in order to decode it, we need to manually skip
the extension header and the tuple format identifier to get to the tuple
data and create a tuple. We can ignore the tuple format identifier (and the
tuple format that was originally sent for this tuple), since the format is
provided by the merger itself.

Needed for #8147

NO_CHANGELOG=<internal change>
NO_TEST=<tested by integration tests>
NO_DOC=<internal change>

Extend the MsgPack object creation interface to push a MsgPack context
required for decoding tuples coming from IPROTO, the context ownership is
acquired by the new object to avoid copying (thus, a virtual `move`
function is added to the MsgPack context).

When a MsgPack context is owned by a MsgPack object, the lifetime of the
former is controlled by the latter, hence a virtual `destroy` function is
added to the MsgPack context.

Needed for #8147

NO_CHANGELOG=refactoring
NO_DOC=refactoring
NO_TEST=refactoring

Add new MsgPack extension for tuples MP_TUPLE with the following structure:
+------+-----------------+----------+
|MP_EXT|      MP_UINT    | MP_ARRAY |
+------+-----------------+----------+
        ^                 ^
        format identifier tuple data

Needed for #8147

NO_CHANGELOG=<internal change>
NO_DOC=<internal change>

When data is coming from IPROTO and is passed to Lua, it is convenient to
pass raw MsgPack data (see `box.schema.func::takes_raw_args` option).
Currently, the port's MsgPack is retrieved directly via `port_get_msgpack`
and then a MsgPack object is constructed via `luamp_push`. In the future,
ports can have a MsgPack decoding context that will also be needed to be
passed to the MsgPack object constructor, so we encapsulate the MsgPack
object construction completely into a port method by introducing a new
operation mode to `port_dump_lua`.

This new operation mode is essentially a no-op for a MsgPack port (which is
the case for data coming from IPROTO), and is completely inefficient for
other types of ports (since an intermediate conversion to MsgPack occurs),
but we do not expect users to be calling Lua stored procedures with the
`takes_raw_args` locally. To handle the latter case we introduce a
`port_dump_lua_mp_object_mode_slow` helper function which facilitates
retrieving MsgPack data from ports via region allocation, creating a
Msgpack object and freeing the MsgPack data.

Needed for #8147

NO_CHANGELOG=refactoring
NO_DOC=refactoring
NO_TEST=refactoring

Currently, `port_dump_lua` has 2 modes of operation: 'flat' and 'tabular',
which are controlled via a boolean flag. We plan to extend it with another
mode of operation, 'Msgpack object', so we need to replace the flag with a
enumeration. Plus, this refactoring should improve readability.

Needed for #8147

NO_CHANGELOG=refactoring
NO_DOC=refactoring
NO_TEST=refactoring

In scope of #8633 call and eval encoding will require an additional
argument (whether to encode tuples as extensions and send formats), and
method encoders already have 3 arguments (MsgPack stream, IPROTO sync and
IPROTO stream identifier): to simplify further extension of method encoders
move arguments to a netbox_method_encode_ctx struct.

Needed for #8633

NO_CHANGELOG=refactoring
NO_DOC=refactoring
NO_TEST=refactoring

The MsgPack extension encoder has a strange interface: it returns the type
it has encoded the value to which at the same time indicates whether it has
succeeded in encoding the value — instead, let's return the type in an
output parameter and return the encoding status as a boolean value.

Needed for #8147

NO_CHANGELOG=refactoring
NO_DOC=refactoring
NO_TEST=refactoring

The MsgPack context structure is the essential data structure for
collecting and retrieving various meta information during MsgPack
encoding/decoding. This structure will be a base for a box-specific MsgPack
context (and, possibly, others), so we make it extendable by adding a
padding (to avoid dynamic allocations).

Refactor existing MsgPack encoding interfaces to use the MsgPack context
structure:
1. extend the port dumping interface, so that we can collect meta
information about dumped tuples' formats;
2. extend the port MsgPack retrieval interface, so that we can also
retrieve meta information about tuple formats for decoding tuples coming
from IPROTO (needed when decoding is postponed to a MsgPack object), the
context ownership is acquired by the caller to avoid copying;
3. extend the MsgPack encoding/decoding interfaces to accept a MsgPack
context to collect meta information during encoding or to use this meta
information for decoding — the context is passed down up to the extension
encoder/decoder, which is the point where future box tuple extension will
be processed.

Needed for #8147

NO_CHANGELOG=refactoring
NO_DOC=refactoring
NO_TEST=refactoring

The tuple format map is the essential data structure for collecting tuple
formats during box tuple encoding and for looking up tuple formats during
box tuple decoding.

Needed for #8147

NO_CHANGELOG=<internal data structure>
NO_DOC=<internal data structure>

Co-authored-by: Aleksandr Lyapunov <alyapunov@tarantool.org>

At the last itaration of review it was decided to throw an alerts,
when no UUID was passed to config and name is not set in _cluster.

This leads to alerts, thrown during replicaset bootstrap. However,
_cluster:on_replace trigger wasn't updated for that change, it
asssumed that when insert of a new replica is done, no alerts was
thrown.

Let's fix the behavior of on_replace trigger so that it deletes
alerts as soon as replica joined with name.

Follow-up #8978

NO_DOC=bugfix
NO_TEST=<already tested>
NO_CHANGELOG=bugfix

New commits:

- test: fix test build for Debian 9 and alike
- test: fix unused variable warning in matras test

NO_TEST=submodule bump
NO_CHANGELOG=submodule bump
NO_DOC=submodule bump

It's impossible to set names on Tarantool below 3.0.0, as all DDL
is forbidden before schema upgrade.

Let's make names NoOp on schema below Tarantool 3.0.0 and set names
automatically only when schema upgrade is done.

Follow-up #8978

NO_DOC=tarantool/doc#3661

This commit moves all code, related to working with versions and which
was used in box/lua/upgrade.lua, into a separate module and exports it
to Lua API as 'internal.version'

This is needed, as in the following commit we set names automatically
only when schema version is more than 3.0.0. This module is used their
in order to avoid code duplication.

Follow-up #8978

NO_DOC=internal
NO_TEST=<already tested>
NO_CHANGELOG=internal

For now it's impossible to use config module in order to recover
from snaps, which don't have names set in them. Calling box.cfg
with names on recovery fails with MISMATCH error, which is caused
by difficult implementation of setting names on first box.cfg,
as names can be set only on rw instance.

This commit doesn't call box.cfg with names, if these names are
missing from the snapshot file. Instead it creates a fiber, which
will set names, when it's possible to do so. Only master sets the
names for the whole cluster.

Closes #8978

NO_DOC=tarantool/doc#3661
NO_CHANGELOG=following commits

When the name is manually set on master by replace in _cluster space,
calling box.cfg on replica with the same name causes its hang. The
problem is the fact, that resubscribe is initiated and waiting for
APPLIER_REGISTERED status is started. As applier knows, that no
registration should be done, this never happens.

Let's don't initiate registration, when instance name is already set.

Needed for #8978

NO_DOC=bugfix
NO_CHANGELOG=not released yet

For now it's impossible to drop created alert in any way except
manual searching for _alerts table. However, we need to drop alerts
on missing names, when the names are set.

Let's introduce simple key-value alerts in order to easily drop them
by key.

Needed for #8978

NO_DOC=internal
NO_TEST=following commit
NO_CHANGELOG=internal

Currently only instance_uuid is validated before recovery process.
All names and replicaset_uuid are checked only when recovery is done,
which can take a long time. It can be frustrating to users, which
have been waiting for several hours only to get name mismatch error.

Let's read the small part of snapshot file before calling box.cfg
in order to figure out, whether the names and uuids, passed to
configuration match the ones, saved inside the snapshot.

During config reload there's no sense in reading snapshot file, as
data is already saved inside spaces, let's read them. We still check
that names in config and names in spaces don't contradict during
config reload.

This commit also introduces methods, for getting names, which are
not set in snap (or memory), this'll be used in consequent commits
to set names automatically.

Needed for #8978

NO_DOC=tarantool/doc#3661

box.schema has a number of constants, e.g. IDs of system spaces,
which may be useful for the user. Let's allow to access box.schema
before box.cfg is called.

It is used in checking names, as we need to know, which snapshot spaces
to scan.

Needed for #8978

NO_DOC=minor change

This commit introduces the new method for xlog module: xlog.meta().
It opens an xlog file, reads and returns the meta block of the file,
which includes its filetype, instance_uuid and vclocks.

It's needed in order to introduce checking of names inside the config
module in the following commit.

Needed for #8978

@TarantoolBot document
Title: xlog.meta(file-name) method

Description: Open an xlog file, and return its meta block.
Possible errors: Failed to open a file or it does not contain properly
formatted meta block.

Example:

```lua
tarantool> xlog = require('xlog')
---
...

tarantool> xlog.meta('00000000000000000000.snap')
---
- filetype: SNAP
  prev_vclock: {}
  instance_uuid: 87b2e60f-275c-4efa-9b0e-e9562e309692
  vclock: {}
...
```

Config's box.cfg applier scans snapshot_dir in order to find out,
whether recovery is going to be done. It's needed in order to determine,
whether the instance should be started into ro mode firstly.

Let's move info about snapshot into separate file in utils. The commit
also introduces snapshot_path, which will be used in the following
commits in order to validate names.

Needed for #8978

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

Closes #9237

Add exports for fiber_set_name_n, fiber_name, fiber_id, fiber_csw &
fiber_find.

Also make fiber_set_joinable, fiber_set_ctx & fiber_get_ctx interpret
NULL as the current fiber.

@TarantoolBot document
Title: add basic fiber api to ffi exports.

5 basic functions can now be used via ffi api, which were previously
only accessible via lua api: fiber_set_name_n, fiber_name, fiber_id,
fiber_csw & fiber_find.

fiber_set_joinable now interprets NULL as current fiber.

Old: "Failed to allocate 2147483648 bytes in hash_table for key"
New: "Failed to allocate 16384 bytes in hash_table for key"

ERRINJ_INDEX_ALLOC cannot be used to test this error, because it fails
earlier, so ERRINJ_HASH_INDEX_REPLACE is introduced.

Follow-up #3594

NO_DOC=minor
NO_CHANGELOG=minor

The issue is fixed in the `small' submodule by the following commit:

* matras: fix matras_view::block_count overflow

Closes #3594

NO_TEST=The test requires more than 64 GB of RAM.

@TarantoolBot document
Title: Document maximum number of tuples in hash index
Product: Tarantool
Root document: https://www.tarantool.io/en/doc/latest/book/box/limitations/

Number of tuples in hash index: 2147483648

The bug was fixed in the small library:
 - slab: fix NULL ptr deref in assertion in slab_get
   https://github.com/tarantool/small/commit/ef77efacd452cb90caea2caf22d266f791c95ec3
 - slab: fix uint32_t overflow in slab_capacity
   https://github.com/tarantool/small/commit/77203600a7c645d97bce56f901eec25de0b29d6e

The small library submodule was updated in commit ebafd684
("small: bump version").

Closes #9218

NO_DOC=changelog
NO_TEST=changelog

Recently added persistent instance/replicaset/cluster names have certain
validation rules (see #5029 and #9148). An instance name and a
replicaset name that are provided in a declarative configuration are
stored in the database, so they should follow the same rules.

This patch implements the validation for instance/replicaset/group
names, for `--name` CLI option and `TT_INSTANCE_NAME` environment
variable.

Part of #8862
Related to #9148

NO_DOC=The user visible change is about a better error message if a
       wrong name is passed. The naming rules are documented in
       https://github.com/tarantool/doc/issues/3466,
       https://github.com/tarantool/doc/issues/3467,
       https://github.com/tarantool/doc/issues/3468.

Sometimes shell quoting is needed in tests to trigger a validation
error. For example, if the argument is empty or contains whitespaces.

Left the default unchanged to don't affect existing tests.

Part of #8862

NO_DOC=testing helper change
NO_CHANGELOG=see NO_DOC
NO_TEST=see NO_DOC

We hit #3807 in release/2.11 for release ASAN build with ASAN-friendly
small allocators.

Follow-up #7327

NO_CHANGELOG=internal
NO_DOC=internal

Closes #9277

@TarantoolBot document
Title: Document `box_space_execute_priv` compatibility option

Historically, it was possible to grant the `execte` privilege on
a space although this action had no effect. Since Tarantool 3.0
it isn't allowed anymore. The new `compat` module option
`box_space_execute_priv` was added to revert to the old behavior.

Please create a documentation page for the new compatibility option at
https://tarantool.io/compat/box_space_execute_priv

Example:

```
tarantool> box.cfg{log_level = 'error'}
---
...

tarantool> box.schema.user.create('alice')
---
...

tarantool> box.schema.user.grant('alice', 'execute', 'space')
---
- error: Unsupported space privilege 'execute'
...

tarantool> require('compat').box_space_execute_priv = 'old'
---
...

tarantool> box.schema.user.grant('alice', 'execute', 'space')
---
...
```

Configuring log modules work differently with log.cfg and box.cfg:
box.cfg{log_modules=...} overwrites the current config completely while
log.cfg{modules=...} overwrites the currently config only for the
specified modules. Let's fix this inconsistency by making log.cfg behave
exactly as box.cfg.

Closes #7962

NO_DOC=bug fix

This reverts commit 0ea1ba87.

Bump is reverted due to regression in Curl 8.4.0.

Needed for #9283

NO_DOC=libcurl submodule bump
NO_TEST=libcurl submodule bump
NO_CHANGELOG=libcurl submodule bump

We decided to allow underscore in names to provide easier support
for many existent installations.

Closes #9148

NO_CHANGELOG=no updates to the feature description
NO_DOC=update existent doc tickets 3466, 3467, 3468

Message body resides in one of rotating input buffers for the
connection. When we don't need message body anymore we advance the
reading end of the input buffer by the size of the message. But message
processing order can differ from the order of messages in the wire.
Thus this advancing a bit hacky. Let's instead mark the data in the
input buffer as read when we process all the messages in the input
buffer. We cannot reuse the buffer any earlier anyway.

Follow-up #7327

NO_TEST=refactoring
NO_CHANGELOG=refactoring
NO_DOC=refactoring

The API functions additionally poison related data in ASAN build.

Follow-up #7327

NO_TEST=refactoring
NO_CHANGELOG=refactoring
NO_DOC=refactoring