box.cfg.slab_alloc_granularity sets the min alignment used for struct
memtx_tuple. By default, it's set to the machine word size (8 bytes
typically), but it can be set to 4 bytes to reduce memory usage.

The problem is that we add memtx tuples to a garbage collection list
using the intrusive list pattern (basically, store a pointer to the next
entry in the first 8 bytes of the `memtx_tuple` struct). If
`box.cfg.slab_alloc_factor` is set to 4, the pointer will be unaligned
while the compiler will generate the machine code assuming it is
naturally aligned. For some architectures (x86), this works fine, but
for others (ARM), this may result in a runtime failure.

To fix this issue, we need to instruct the compiler that the list
pointer stored in `memtx_tuple` may be unaligned.

For now, let's silently ignore granularity < the machine word size,
because we're planning to switch from lifo to stailq for memtx garbage
collection lists, which would result in explicit compiler warnings.

See #7422
Needed for #7185

NO_DOC=invisible to the user
NO_CHANGELOG=invisible to the user
NO_TEST=will be added later when the bug is fixed

It isn't needed - we can check if delay_free_mode > 0 to figure our if
we are in the delayed free mode.

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

Checking #7185 with functional tests is complicated. Let's add a unit
test for this.

Needed for #7185

NO_DOC=test
NO_CHANGELOG=test

Just regular update to bring openssl security fixes into tarantool.

Added a patch for Mac OS to fix a build failure, see
https://github.com/openssl/openssl/issues/18720.

Changelog: https://www.openssl.org/news/cl111.txt
Vulnerabilities: https://www.openssl.org/news/vulnerabilities.html

NO_TEST=security update of a dependency
NO_DOC=security update of a dependency

Now we can make interactive transactions like in case of
local or txt remote consoles if peer supports streams.

Closes #7413

NO_DOC=minor change

The issue is if in binary remote console a error is thrown in expression
like "box.begin() error('something') box.commit()" then it is
overwritten by iproto check for active transactions at the end of eval.
The solution is to rollback active transaction in this case in console
code before iproto check.

Let's also assert in tests a behaviour that after successful transaction
begin it stays active in next evaluations (successful or not) until
explicit rollback.

Closes #7288

NO_DOC=minor change

Starting server is failed after chdir(3) for example if we call
box.cfg{} with work_dir set.

NO_DOC=fix testing harness
NO_TEST=fix testing harness
NO_CHANGELOG=fix testing harness

We keep per server directories there too, not only socket files.

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

For the client_print there are also other checks and actions but they
are unnecesessary.

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

The ; separator was missed after select (check the first hunk) and as
a result the rollback was not executed. So the next test was failed merely
due to the unfinished transaction of the previous test.

We also need to fix the third test to use a different tuple.
Also fix a couple of comments.

NO_DOC=test fix
NO_TEST=test fix
NO_CHANGELOG=test fix

If MP_ERROR contains an error message with invalid UTF-8 sequences,
encode it in Base64 to obtain printable string.

Closes #6781
Closes #6934

NO_DOC=bugfix

Currently the call to internal.format_lua in format_lua_value is covered
by pcall only for `status == true`, however the opposite case also can
raise an exception. This patch adds the missed exception handling.

Part of #6781
Part of #6934

NO_DOC=bugfix
NO_CHANGELOG=minor bug
NO_TEST=not possible to trigger format_lua failure for !status

lua_encode can raise an exception (e.g. from luaL_checkfield) or return
`nil, err` result. Handle it in lbox_console_format_lua, similar to
lbox_console_format_yaml.

When fixed, it will expose another issue - nil value can not be
serialized:

```
tarantool> \set output lua
true;
tarantool> 1, nil, 2
1, {error = "console: exception while formatting the output:
    \"serializer: unexpected data (nd.field.size 0 nd.field.type 5)\""}, 2;
tarantool>
```

Fix this too.

Part of #6781
Part of #6934

NO_DOC=bugfix
NO_TEST=not a visible change
NO_CHANGELOG=not a visible change

Currently the call to internal.format_yaml in output_handlers["yaml"]
is covered by pcall only for `status == true`, however the opposite
case also can raise an exception. This patch adds the missed exception
handling. Two distinct calls are required, because it is not possible to
assing variadic arguments (...) to a variable if some of them is nil.
If the first call fails, internal.format_yaml will be called for the
second time. Hopefully, it will not fail during formatting of the error
message received from libyaml.

Before:
```
tarantool> require('net.box').self:call('\x80')
LuajitError: builtin/box/console.lua:710: expected SCALAR, SEQUENCE-START,
    MAPPING-START, or ALIAS fatal error, exiting the event loop
~/test$ echo $?
0
~/test$
```

After:
```
tarantool> require('net.box').self:call('\x80')
---
- error: 'console: an exception occurred when formatting the output:
    expected SCALAR, SEQUENCE-START, MAPPING-START, or ALIAS'
...

tarantool>
```

Part of #6781
Part of #6934

NO_DOC=bugfix
NO_TEST=see later commits
NO_CHANGELOG=see later commits

This patch adds a missed check for `lua_yaml_encode` return value,
similar to the check in `console_dump_plain`. Now Lua error is raised
if YAML encoding failed for any reason (e.g. OOM or formatting error).

Before:
```
tarantool> box.error.new(box.error.ILLEGAL_PARAMS, '\x80')
~/test$ echo $?
0
~/test$
```

After:
```
tarantool> box.error.new(box.error.ILLEGAL_PARAMS, '\x80')
---
- error: 'console: an exception occurred when formatting the output:
    expected SCALAR, SEQUENCE-START, MAPPING-START, or ALIAS'
...

tarantool>
```

Part of #6781
Part of #6934

NO_DOC=bugfix
NO_TEST=see later commits
NO_CHANGELOG=see later commits

The upgrade script first tries to determine if the node is booted from
old snaps not recoverable on current Tarantool versions. If this is the
case, it sets up special triggers so that snaps are automatically
converted to a suitable format.

This happens before box.cfg{}, so the workdir is not set at this point
in time, and the upgrade script should take configured work_dir into
account explicitly. Fix this.

Closes #7232

NO_DOC=bugfix

Prior to this patch, some opcodes could use a pointer to struct space
that was set during parsing. However, the pointer to struct space is not
really something that defines spaces. A space can be identified by its
ID or name. In most cases, specifying space by pointer works fine, but
truncate() changes the pointer to space, resulting in a sigfault for
prepared statements using the above opcodes. To avoid this problem, a
new opcode has been introduced. This opcode uses the space ID to
determine the pointer to the struct space at runtime and stores it in
the MEM, which is later used in the mentioned opcodes.

Closes #7358

NO_DOC=bugfix

Opcode IteratorReopen is not used and should be dropped.

Part of #7358

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

We used to use it for allocating functional keys, but now we allocate
those as tuples. Let's drop the legacy infrastructure and make alloc and
free MemtxAllocator methods private.

Follow-up #7376

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

Functional index keys are allocated and freed with MemtxAllocator's
alloc and free methods. In contrast to tuples, which are allocated and
freed with alloc_tuple and free_tuple, freeing a functional index key
happens immediately, irrespective of whether there's a snapshot in
progress or not. It's acceptable, because snapshot only uses primary
indexes, which can't be functional. However, to reuse the snapshot
infrastructure for creating general purpose user read views, we will
need to guarantee that functional index keys stay alive until all read
views using them are closed.

To achieve that, this commit turns functional index keys into tuples,
which automatically makes them linger if there's an open read view.
We use the same global tuple format for allocating functional keys,
because the key format is checked in key_list_iterator_next.

Closes #7376

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

Without CA certificates the HTTP client will unable to verify server's
certificate, so the only way to perform an HTTPS request would be use
the `verify_peer = false` option -- disable certificate validation at
all.

The runtime search of system CA bundle/certificates was unintentionally
disabled in 2.10.0 (PR #7119). The patch enabled is back.

The main motivation behind the runtime search is difference in paths on
different systems. Since we ship Tarantool Enterprise Edition as
executable with ability to run on different Linux distributions, we
can't choose one particular path at build time. See details in #5746.

The `CURL_CA_BUNDLE_SET` and `CURL_CA_PATH_SET` options were removed,
because they are not 'real' curl configuration options, but rather
cached values to don't repeat file/directory search at re-configuration.
It looks as internal logic of Curl's CMake script.

NO_DOC=Lack of proper HTTPS support is definitely broken behavior, there
       is no sense to document it or the opposite.
NO_TEST=A simple test would require to send a request to external host.
        It would not work without internet connection or in a sandbox.
        Such test also would be unstable and would fail from time to
        time due to network conditions. I verified the patch manually. I
        have an idea to add more thorough http client testing later.

Fixes #7372

fiber_wakeup has been adapted to spurious wakeups, so this protection
is no longer needed.

Part of #7166

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

Currently the latch doesn't guarantee the order in which it is acquired
by fibers, which requested it. E.g. it is possible to wake up spuriously
a fiber which is yielding in the latch_lock, it will be removed from
l->queue by fiber_make_ready, then it will be inserted to l->queue
again, but for this time, to the head of the list instead of its
original place in the queue.

Fix this by using latch_waiter structure, which is linked into l->queue.

Part of #7166

@TarantoolBot document
Title: Update box_latch_lock description
Since: 2.11

Add "Locks are acquired in the strict order as they were requested." to
the box_latch_lock description in C API reference - Module latch.

This patch fixes the following error while tarantool build on
FreeBSD 13.1:

    [100%] Linking CXX executable watcher.test
    ld: error: undefined symbol: set_sigint_cb
    >>> referenced by console.c:261 (lua/console.c:261)
    >>>               console.c.o:(lbox_console_readline) in
                        archive ../../src/box/libbox.a
    >>> referenced by console.c:342 (lua/console.c:342)
    >>>               console.c.o:(lbox_console_readline) in
                        archive ../../src/box/libbox.a
    >>> referenced by console.c:352 (lua/console.c:352)
    >>>               console.c.o:(lbox_console_readline) in
                        archive ../../src/box/libbox.a
    >>> referenced 1 more times
    c++: error: linker command failed with exit code 1
        (use -v to see invocation)
    gmake[2]: *** [test/unit/CMakeFiles/watcher.test.dir/build.make:152:
        test/unit/watcher.test] Error 1
    gmake[1]: *** [CMakeFiles/Makefile2:10528:
        test/unit/CMakeFiles/watcher.test.dir/all] Error 2
    gmake: *** [Makefile:156: all] Error 2

NO_DOC=minor changes
NO_TEST=minor changes
NO_CHANGELOG=minor changes

The gh_6565 test doesn't stop the hot standby replica it started,
because the replica should fail to initialize and exit eventually
anyway. However, if the replica lingers until the next test due to
https://github.com/tarantool/test-run/issues/345, the next test may
successfully connect to it, which is likely to lead to a failure,
because UNIX socket paths used by luatest servers are not randomized.

For example, here gh_6568 test fails after gh_6565, because it uses the
same alias for the test instance ('replica'):

NO_WRAP
[008] vinyl-luatest/gh_6565_hot_standby_unsupported_>                 [ pass ]
[008] vinyl-luatest/gh_6568_replica_initial_join_rem>                 [ fail ]
[008] Test failed! Output from reject file /tmp/t/rejects/vinyl-luatest/gh_6568_replica_initial_join_removal_of_compacted_run_files.reject:
[008] TAP version 13
[008] 1..1
[008] # Started on Fri Jul  8 15:30:47 2022
[008] # Starting group: gh-6568-replica-initial-join-removal-of-compacted-run-files
[008] not ok 1  gh-6568-replica-initial-join-removal-of-compacted-run-files.test_replication_compaction_cleanup
[008] #   builtin/fio.lua:242: fio.pathjoin(): undefined path part 1
[008] #   stack traceback:
[008] #         builtin/fio.lua:242: in function 'pathjoin'
[008] #         ...ica_initial_join_removal_of_compacted_run_files_test.lua:43: in function 'gh-6568-replica-initial-join-removal-of-compacted-run-files.test_replication_compaction_cleanup'
[008] #         ...
[008] #         [C]: in function 'xpcall'
[008] replica | 2022-07-08 15:30:48.311 [832856] main/103/default.lua F> can't initialize storage: unlink, called on fd 30, aka unix/:(socket), peer of unix/:(socket): Address already in use
[008] # Ran 1 tests in 0.722 seconds, 0 succeeded, 1 errored
NO_WRAP

Let's fix this by explicitly killing the hot standby replica. Since it
could have exited voluntarily, we need to use pcall, because server.stop
fails if the instance is already dead.

This issue is similar to the one fixed by commit 85040161 ("test:
stop server started by vinyl-luatest/update_optimize test").

NO_DOC=test
NO_CHANGELOG=test

Handle status header response like 'HTTP/2 200' with version without
dot.

Closes #7319

NO_DOC=bugfix

Parser does not change its input data on parsing and it's caller as well.

NO_DOC=internal
NO_CHANGELOG=internal
NO_TEST=refactoring

We use LuaJIT 'bit' module for bitwise operations. Due to platform
interoperability it truncates arguments to 32bit and returns signed
result. Thus on granting rights using bit.bor to admin user which
have 0xffffffff rights (from bootstrap snapshot) we get -1 as a result.
This leads to type check error given in issue later in execution.

Closes #7226

NO_DOC=minor bugfix

This patch fixes the following error on OSX systems:

    autoreconf: error: aclocal failed with exit status: 2
    make[3]: *** [build/curl/work/stamp/bundled-libcurl-project-configure]
        Error 2
    make[2]: *** [CMakeFiles/bundled-libcurl-project.dir/all] Error 2
    make[2]: *** Waiting for unfinished jobs....

To resolve this issue we need to install the `autoconf`, `automake`,
and `libtool` packages before tarantool build.

NO_DOC=ci
NO_TEST=ci
NO_CHANGELOG=ci

Installing the `autoconf` and `automake` packages is not enough.
We need to install the `libtool` package as well to avoid errors
like this:

  configure:7323: error: possibly undefined macro: AC_LIBTOOL_WIN32_DLL
    If this token and others are legitimate, please use
    m4_pattern_allow.
    See the Autoconf documentation.
  configure:7324: error: possibly undefined macro: AC_PROG_LIBTOOL
  autoreconf: /usr/bin/autoconf failed with exit status: 1

NO_DOC=ci
NO_TEST=ci
NO_CHANGELOG=ci

The issue was in the `deps_freebsd` target that installed test-run deps
(py38-yaml, py38-six, py38-gevent) via the `pkg` package manager.

At some moment these packages were upgraded to 39 version on the mirrors
and became not available for installation anymore.

Now test-run deps are installed via `pip` package installer to avoid
such issues in the future. Installation via `pip` is more robust because
we have deps versions pinned in test-run and also we have python version
pinned to 3.8.

NO_DOC=ci
NO_TEST=ci
NO_CHANGELOG=ci

Passing format->is_temporary to MemtxAllocator::free_tuple complicates
the allocator API and the code using it. Let's introduce a per tuple
flag for such tuples.

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

Currently, we call memtx_enter_delayed_free_mode() per each index,
in index::create_snapshot_iterator(), but there's actually no need to
bump the snapshot_version more than once per snapshot. Let's move it
to the place where we start checkpoint/join and drop memtx wrappers
around memtx_allocators functions.

This will simplify reworking the memtx read view API, see #7364.

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

Let's hide all the logic regarding delayed freeing of memtx tuples to
MemtxAllocator and provide memtx_engine with methods for allocating and
freeing tuples (not memtx_tuples, just generic tuples). All the tuple
and snapshot version manipulation stuff is now done entirely in
MemtxAllocator.

This is a preparation for implementing a general-purpose tuple read view
API in MemtxAllocator, see #7364.

Note, since memtx_engine now deals with the size of a regular tuple,
which is 4 bytes less than the size of memtx_tuple, this changes the
size reported by OOM messages and the meaning of memtx_max_tuple_size,
which now limits the size of a tuple, not memtx_tuple.

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

It isn't necessary to prefix all static class members with the class
name specifier in the class methods.

NO_DOC=refactoring
NO_TEST=refactoring
NO_CHANGELOG=refactoring

Changelog: https://curl.se/changes.html#7_84_0

New release contains fixes for 4 security problems [1].

Patch adds a new option defined in curl build infrastructure with it's
default value used in 7.84.0. NOTE: PSL is a Public Suffix List (PSL),
it is a list of suffixes used for cookies.

1. https://curl.se/docs/releases.html

NO_DOC=libcurl submodule bump
NO_TEST=libcurl submodule bump

This patch fixes a bug where the ANY field type was replaced by the
SCALAR field type in the ephemeral space used in ORDER BY.

Closes #7345

NO_DOC=bugfix

After this patch, the result type of arithmetic between two unsigned
values will be INTEGER.

Closes #7295

NO_DOC=bugfix

This patch fixes the following error:

    [ 35%] Performing configure step for 'bundled-libcurl-project'
    *** Do not use buildconf. Instead, just use: autoreconf -fi
    ./buildconf: 8: exec: autoreconf: not found

To resolve this issue we need to install the `autoconf` and `automake`
packages before building.

NO_DOC=ci
NO_TEST=ci
NO_CHANGELOG=ci

This makes the test more real-life, and allows not to bother in the
child process with the memory allocated prior to fork.

Closes #7370

NO_DOC=test fix
NO_CHANGELOG=test fix