The patch updates curl module to the version 8.3.0 [1] and updates a
CMake module for building curl library.

Changes in CMake module:

- Option `CURL_STATICLIB` is gone and replaced with `BUILD_STATIC_LIBS`.
- Option `CURL_USE_GNUTLS` was added and disabled by default.
- NSS library support was removed and option `CURL_USE_NSS` has been
  removed as well.
- Option `CMAKE_UNITY_BUILD` was added and disabled by default.
- Option `CURL_DISABLE_FORM_API` was added and disabled by default. It is
  in fact depends on `CURL_DISABLE_MIME`, but anyway disabled explicitly.

Changelog: https://curl.se/changes.html#8_3_0

1. https://github.com/curl/curl/releases/tag/curl-8_3_0

NO_DOC=libcurl submodule bump
NO_TEST=libcurl submodule bump

Fixes #9086

This patch fixes the following warning:

NO_WRAP
  CMake Warning (dev) at cmake-3.25/Modules/FindPackageHandleStandardArgs.cmake:438 (message):
    The package name passed to `find_package_handle_standard_args`
    (GetLIBUNWINDVersion.cmake) does not match the name of the calling package
    (LibUnwind).  This can lead to problems in calling code that expects
    `find_package` result variables (e.g., `_FOUND`) to follow a certain
    pattern.
  Call Stack (most recent call first):
    cmake/FindLibUnwind.cmake:82 (find_package_handle_standard_args)
    CMakeLists.txt:552 (find_package)
NO_WRAP

Closes #6998

NO_DOC=build
NO_TEST=build
NO_CHANGELOG=build

All LuaJIT related LSan warnings were suppressed in the scope of the
commit 985548e4 ("asan: suppress all
LSAN warnings related to LuaJIT"), since all compiler flags tweaks were
enclosed in LuaJIT CMake machinery. As a result of the commit in LuaJIT
submodule tarantool/luajit@a86e376 ("build: introduce LUAJIT_USE_ASAN
option") ASan and LSan support has been finally added to LuaJIT runtime,
so it was decided to remove LSan suppressions for LuaJIT functions.
Unfortunately, it was not so easy as it looked like.

At first, Lua global state is not closed properly at Tarantool instance
exit (see <tarantool_free> in src/main.cc and <tarantool_lua_free> in
src/lua/init.c for more info), so LSan false-positive leaks are detected
(for more info, see #3071). Hence, the original LSan suppression for
lj_BC_FUNCC is returned back (temporarily) until the aforementioned
issue is not resolved.

Furthermore, the internal LuaJIT memory allocator is not instrumented
yet, so to find any memory faults it's worth building LuaJIT with system
provided memory allocator (i.e. enable LUAJIT_USE_SYSMALLOC option).
However, again, since Tarantool doesn't finalize Lua universe the right
way, so running Tarantool testing routine with LUAJIT_USE_SYSMALLOC
enabled generates false-positive LSan leaks. Return back here to enable
LUAJIT_USE_SYSMALLOC, when #3071 is resolved.

Last but not least, the default value of fiber stack size is 512Kb, but
several tests in test/PUC-Rio-Lua-5.1-test suite in LuaJIT submodule
(e.g. some cases with deep recursion in errors.lua or pm.lua) have
already been tweaked according to the limitations mentioned in #5782,
but the crashes still occur while running LuaJIT tests with ASan support
enabled. Experiments once again confirm the notorious quote that "640 Kb
ought to be enough for anybody".

Anyway, LuaJIT tests are added to <test-release-asan> target in .test.mk
and LUAJIT_TEST_ENV is extended with required ASan and LSan options.

Follows up #5878

NO_DOC=ci
NO_TEST=ci
NO_CHANGELOG=ci

In scope of the commit 82f4b4a3 ("lib/core/fiber: Increase default
stack size") the default value of fiber stack size is increased up to
512 Kb (you can find the reasons in the aforementioned commit message
and in https://github.com/tarantool/tarantool/issues/3418 description).

Some of the tests in test/PUC-Rio-Lua-5.1-test suite in LuaJIT repo
(e.g. some cases with deep recursion in errors.lua or pm.lua) have
already been tweaked according to the limitations mentioned in
https://github.com/tarantool/tarantool/issues/5782, but the crashes
still occurs while running LuaJIT tests with ASan support enabled.

To make the testing routine more convenient, FIBER_STACK_SIZE option is
introduced to Tarantool CMake machinery. One can provide the size either
by raw digits (i.e. in bytes) or using Kb/Mb suffixes for convenience.

A couple of important nits:
* If the given value is not a multiple of 4Kb, CMake machinery adjusts
  it up to the nearest one greater than this value.
* If the adjusted value is less than 512Kb, configuration fails with the
  corresponding CMake fatal error.

Follows up #3418
Relates to #5782

@TarantoolBot document
Title: introduce FIBER_STACK_SIZE configuration option

To make managing of the default fiber stack size more convenient, the
corresponding CMake option is added.

**NB**: The stack size can't be less than 512Kb and if the given value
is not a multiple of 4Kb, CMake machinery adjusts it up to the nearest
one greater than this value.

There are sporadic segfaults in libunwind during backtrace_collect().
Reproducible with the latest version, and there are open issues with
similar stacks:

https://github.com/libunwind/libunwind/issues/150
https://github.com/libunwind/libunwind/issues/260
https://github.com/libunwind/libunwind/issues/473

Let's disable ENABLE_BACKTRACE for AArch64 Linux until these issues are
resolved in libunwind.

Closes #8572
Part of #8791

NO_DOC=bugfix

It was disabled because ligomp.a on some AArch64 systems (e.g. CentOS)
is compiled without PIC support. Now Tarantool doesn't depend on ligomp,
so it's safe to turn on PIC for AArch64 systems.

Follow-up #7536

NO_DOC=build
NO_TEST=build

It was disabled because dlsym tests failed when PIC is on (#7640).
However, later it turned out that the issue is not related to PIC, and
LuaJIT was turned off for such tests by commit 67e79b15 ("test: turn
LuaJIT off in tests with dlsym"). Now it's safe to turn on PIC for FreeBSD.

Follow-up #7536

NO_DOC=build
NO_TEST=build

OpenMP is no longer used since commit 4f617b70 ("box: introduce
memtx_sort_threads config parameter"). All dependencies on libgomp
should be removed.

Follow-up #7689

NO_DOC=build
NO_TEST=build

`string` with JOIN option is only available since CMake 3.12, but we have
developers using CMake 3.1: implement a utility `string_join` function to
remove this dependency.

Closes #5881

NO_CHANGELOG=<build fix>
NO_DOC=<build fix>
NO_TEST=<build fix>

Closes #3389
Closes #7689
Closes #4646

@TarantoolBot document
Title: new box.cfg parameter memtx_sort_threads

The parameter sets the number of threads used to sort keys of secondary
indexes on loading memtx database. The parameter cannot be changed
dynamically (as it does not make sense).

Maximum value is 256, minimum is 1. Default is to use all available cores.

Usage example:
```
box.cfg{memtx_sort_threads=4}
```

Patch made following changes:
- add a CMake function lua_multi_source that unpacks symbol names and
  Lua source paths in passed list and compiles Lua source into *.lua.c
  sources using lua_source() function. Function is needed for embedding
  Lua modules consisting from a number of files to Tarantool EE.
- add using lua_multi_source instead of lua_source for etcd client module.
- remove function prototypes and declarations related to extra features.
  All this code is moved to Tarantool EE source code.

Part of #8620
Follows up https://github.com/tarantool/tarantool-ee/issues/443

NO_CHANGELOG=refactoring
NO_DOC=refactoring
NO_TEST=refactoring

Patch introduce a macro FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION. It is
a common build macro recommended in libFuzzer's documentation [1].

1. https://llvm.org/docs/LibFuzzer.html#fuzzer-friendly-build-mode

Needed for #4826

NO_CHANGELOG=build
NO_DOC=build
NO_TEST=build

Follows-up #4823

NO_CHANGELOG=internal
NO_DOC=internal
NO_TEST=internal

Added Google's 'libprotobuf-mutator' and 'protobuf' libraries
for developing grammar-based LuaJIT and SQL fuzzers based on
LibFuzzer.

It is needed to build protobuf module from source because
by default, the system-installed version of protobuf is used
by libprotobuf-mutator, and this version can be too old.

Part of #4823

NO_CHANGELOG=<dependencies>
NO_DOC=<dependencies>
NO_TEST=<dependencies>

Updating libcurl from 7.84.0 to 7.87.0 in commit 09f4eca1
("third_party: update libcurl from 7.84.0 to 7.87.0")
made it impossible to built tarantool with bundled curl with nghttp2,
because nghttp2 was not updated to fit libcurl 7.87.0, this led to
missing function
nghttp2_option_set_no_rfc9113_leading_and_trailing_ws_validation
during linking step.

https://github.com/nghttp2/nghttp2/releases/tag/v1.52.0

Closes #8268

NO_DOC=there is no known behavior changes, consider the change as not
       visible for a user
NO_TEST=can only reproduce in specific environment
NO_CHANGELOG=see NO_DOC

Changelog: https://curl.se/changes.html#7_87_0

New release contains fixes for 7 security problems [1]:

https://curl.se/docs/CVE-2022-35252.html
https://curl.se/docs/CVE-2022-32221.html
https://curl.se/docs/CVE-2022-35260.html
https://curl.se/docs/CVE-2022-42915.html
https://curl.se/docs/CVE-2022-42916.html
https://curl.se/docs/CVE-2022-43551.html
https://curl.se/docs/CVE-2022-43552.html

Patch adds a new option ENABLE_WEBSOCKETS defined in curl build
infrastructure with it's default value used in 7.87.0.

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

NO_DOC=libcurl submodule bump
NO_TEST=libcurl submodule bump

Closes #8150

The EE CMakeConfig.txt adds CE as a sub-directory. For this to work,
we should use project dir instead of cmake dir in CE cmake files,
see commit d8097325 ("cmake: align folders dependencies").

Fixes commit 14f93aee ("libunwind: improve incremental
build/rebuild").

Follow-up #8019

NO_DOC=build fix
NO_TEST=build fix
NO_CHANGELOG=build fix

Currently we use ExternalProject_Add to build libunwind subproject. The
usecases it supports are not quite aligned with our. Configuration and
build steps are rerun in case download step sees updated sources. But
we use git submodules to fetch third-party projects and thus subproject
is not rebuild when it's files are changed on disk. Either because of
we doing some sort of experiments locally or because the new sources
are brought by git pull.

Actually libunwind has a CMakeList.txt so that we could try to build it
just as project subdirectory. But it requires quite a fresh CMake
version, currently supports only build on Visual Studio and probably is
not up to date given it's package version variables. Keeping
CMakeLists.txt in sync with main autotools build is additional
maintenance burden. So I'd like to use main autotools build.

So ext_project_autotools function is added to provide nicer build
integration with third-party projects which are build using autotools.
Actually incremental rebuild works only if CMake is at least 3.12 but
I think this should be true for developer installations.

Hopefully we can reuse the function in case we need to bundle more
subprojects with autotools builds.

Follow-up #5665

NO_DOC=build improvement
NO_TEST=build improvement
NO_CHANGELOG=build improvement

Fixed pthread-related CMake checks. The checks code is built with
`-pedantic-errors` and it leads to errors of the following kind on
clang 15:

```
<...>/CMakeFiles/CMakeScratch/TryCompile-78KaOK/src.c:4:17: error: a
    function declaration without a prototype is deprecated in all
    versions of C [-Werror,-Wstrict-prototypes]
        int main() { pthread_setname_np(pthread_self(), ""); }
                ^
                 void
```

Fixed a warning in the SQL code (it's an error in Debug build):

```
<...>/src/box/sql/vdbeaux.c:170:13: error: variable 'n' set but not used
    [-Werror,-Wunused-but-set-variable]
        static int n = 0;
```

Fixed several warnings from lemon.c of the following kind:

```
<...>/extra/lemon.c:173:6: warning: a function declaration without a
    prototype is deprecated in all versions of C and is treated as a
    zero-parameter prototype in C2x, conflicting with a subsequent
    definition [-Wdeprecated-non-prototype]
void FindRulePrecedences();
     ^
<...>/extra/lemon.c:766:6: note: conflicting prototype is here
void FindRulePrecedences(struct lemon *xp)
```

See also https://github.com/tarantool/small/issues/57

Fixes #8110

NO_DOC=build fix
NO_TEST=build fix

Not used anywhere anymore. No need in it, because SHA1 is provided by
the OpenSSL library.

Closes #7987

NO_DOC=code cleanup
NO_TEST=code cleanup
NO_CHANGELOG=code cleanup

This option enables emission of the DWARF CFI (Call Frame Information)
directives to the assembler. It is enabled by default on most compilers,
but on GCC 7 for AArch64 and older it wasn't [1], so turn it on explicitly.
When enabled, the compiler emits .cfi_* directives that are required for
the stack unwinding, and defines __GCC_HAVE_DWARF2_CFI_ASM.

Part of #7960

[1] https://gcc.gnu.org/pipermail/gcc-patches/2018-March/495549.html

NO_DOC=internal
NO_CHANGELOG=internal
NO_TEST=Tested by test/app-luatest/fiber_parent_backtrace_test.lua

In the scope of this commit check for instruction/line hooks for
compiled code is turned on. As a result a special pad is emitted to the
trace validating whether any debug hook is fired while this trace is
being executed.

Closes #7762

NO_DOC=internal

Reviewed-by: Sergey Kaplun <skaplun@tarantool.org>
Reviewed-by: Maxim Kokryashkin <m.kokryashkin@tarantool.org>
Signed-off-by: Igor Munkin <imun@tarantool.org>

Now we base on some unreleased state of libunwind. This is by itself not
a good practice. Yet the main motivation is that in the current version of
libunwind fast path for backtrace detection on x86_64 does not work.
I guess this is because of libunwind/libunwind@400b3f819ad4 (" x86_64: Stop
aliasing RSP and CFA"). See libunwind/libunwind#440. Fortunately this
commit is not present it the latest release.

Using fast or slow path has a great impact on performance of debug build
where collecting fiber gc allocation backtrace is turned on by default.
It surely depends on frequency and pattern of allocation. Test
sql/delete3 depends on backtrace performance most dramatically. On some
installations collecting backtraces slowed down the test up to 10 times.

If fast path is available then collecting backtrace does not affect
performance in a noticeable way.

I propose not to keep autotools products in the libunwind fork repo as
it is done previously. LOG_CONFIGURE is removed because it somehow
incompatible with using && in CONFIGURE_COMMAND command and not critical
to the build.

Also add autotools build dependencies for packpack package specs.
Currently not all packpack images have autotools preinstalled so this is
required for build to success. Anyway we'b better have precise build
requirements in build specs.

Follow-up #5665

NO_DOC=internal
NO_TEST=internal
NO_CHANGELOG=internal

This patch removes an old CMake kludge, which sets CMAKE_SYSTEM_PROCESSOR
to "x86_64" on all 64-bit architectures, even on ARM (on Darwin).

Closes #7495

NO_DOC=bugfix

After revisiting the list of currently disabled compiler warnings it seems
like the following are unjustifiably disabled:
* `-Wunknown-pragmas`;
* `-Wstrict-aliasing`;
* `-Wunused-value`;
* `-Wchar-subscripts`;
* `-Wformat-truncation`.

`-Wunused-value` appeared to be disabled due to a bug in Clang (see also
llvm/llvm-project#14119): according to the LLVM release dates, the last
affected Clang version was 3.1 and it was hopefully fixed in Clang 3.2.
No currently alive (or recently EOLed) distributions offer such old Clang
versions.

Strict aliasing violation warnings coming from libev cannot be disabled
with `system_header` pragma on GNU compilers older than 8, so we have to
leave to leave it disabled for the whole project in this case.

Otherwise, enable the warnings listed above for the sake of code health.

Closes #7862

NO_CHANGELOG=build
NO_DOC=build
NO_TEST=build

The `-Wa,--debug-prefix-map` compiler flag breaks GNU-based LTO, and also
we cannot reliably test this feature.

Follow-up 256da010

NO_CHANGELOG=bugfix
NO_DOC=bugfix
NO_TEST=bugfix

Since our diagnostics use the `__FILE__` macro, they provide absolute
paths, which is kind of redundant and inconsistent: replace them with
relative ones.

As for debugging information, replacing absolute paths with relative ones
also requires an extra command to tell the debugger where to find the
source files, which is not convenient for developers: provide a new
`DEV_BUILD` option (turned off by default), which replaces absolute paths
with relative ones in debugging information if turned off.

Strip the prefix map flags from compiler flags exported to tarantool via
`src/trvia/config.h`.

Closes #7808

NO_DOC=<verbosity>
NO_TEST=<verbosity>

Setting hardening compiler flags is used in three places: default build,
static build and enterprise build — refactor it into a separate module.

Follow-up e6abe1c9

NO_CHANGELOG=refactoring
NO_DOC=refactoring
NO_TEST=refactoring

e6abe1c9 passes compiler flags to dependencies via a `<project>_build` macro
parameter, which is, firstly, inconvenient, and, secondly, as a result, not
all dependencies got the required compiler flags passed: use global
variables instead and pass these flags to skipped dependencies.

Follow-up e6abe1c9

NO_CHANGELOG=refactoring
NO_DOC=refactoring
NO_TEST=refactoring

`add_compile_flags` macro relied on undefined behaviour: instead of
unsetting the `_lang` and `_flag` variables, it tried to unset their
values — fix this.

NO_CHANGELOG=build
NO_DOC=build
NO_TEST=build

With libunwind/libunwind@f67ef28 we can now use `unw_backtrace` with
PAC enabled on AARCH64 and remove the workaround with glibc's `backtrace`
for this case.

Closes #7285

NO_CHANGELOG=refactoring
NO_DOC=refactoring
NO_TEST=refactoring

Introduce cmake option ENABLE_HARDENING, which is TRUE by default for
non-debug regular and static builds, excluding AArch64 and FreeBSD.
It passess compiler flags that harden Tarantool (including the bundled
libraries) against memory corruption attacks. The following flags are
passed:

* -Wformat - Check calls to printf and scanf, etc., to make sure that
  the arguments supplied have types appropriate to the format string
  specified.

* -Wformat-security -Werror=format-security - Warn about uses of format
  functions that represent possible security problems. And make the
  warning into an error.

* -fstack-protector-strong - Emit extra code to check for buffer
  overflows, such as stack smashing attacks.

* -fPIC -pie - Generate position-independent code (PIC). It allows to
  take advantage of the Address Space Layout Randomization (ASLR).

* -z relro -z now - Resolve all dynamically linked functions at the
  beginning of the execution, and then make the GOT read-only.

Also do not disable hardening for Debian and RPM-based Linux distros.

Closes #5372
Closes #7536

NO_DOC=build
NO_TEST=build

The GENERATED and HEADER_FILE_ONLY source file properties should be
followed by a boolean value. CMake 3.20+ warns if it is not so, see [1].

It seems, the module.h properties do not actually change anything
visible in our case.

[1]: https://cmake.org/cmake/help/latest/policy/CMP0118.html

NO_DOC=pure code health patch, no visible changes
NO_TEST=pure code health patch, no visible changes
NO_CHANGELOG=pure code health patch, no visible changes

c-ares relies on libresolv on OS X, and when built statically, brings
the same dependency on to libcurl.

Link libcurl with libresolv explicitly when linking with c-ares

NO_DOC=build fix
NO_TEST=build fix
NO_CHANGELOG=build fix

Updated third_party/zstd submodule from v1.5.0 to 1.5.2 version.
The new version fixes a few bugs found by fuzzing testing.

Note, the new zstd version contains a .S source file so we need to
include ASM to the CMake project languages.

NO_DOC=build
NO_TEST=build

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

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

__gcov_flush was removed in gcc11.

Since gcc11 __gcov_dump calls __gcov_lock at the start and
__gcov_unlock before returning. Same is true for __gcov_reset.
Because of that using __gcov_reset right after __gcov_dump since gcc11
is the same as using __gcov_flush before gcc11.

Closes #7302

NO_CHANGELOG=internal
NO_DOC=internal
NO_TEST=internal

For the sake of maximizing backtrace collection performance, build
the libunwind project submodule with "-O2" compiler flag.

Also, build it with the "-g" compiler flag just in case to simplify
debugging.

Last but not least, pass the same archive-maintaining program used by the
main CMake project to the libunwind project submodule to make the build
homogeneous.

Needed for #7207

NO_CHANGELOG=build
NO_DOC=build
NO_TEST=build

By default CMake generates Makefiles for building a project. However, it
allows to generate Ninja files. Ninja [1] may build project a bit faster
than Make, see [2].

Patch adds fixes for CMake files allowing to use Ninja for building
Tarantool:

1. Fixed dependencies in ExternalProject_Add(), see explanation in [3]
2. Fixed ninja error due to presence of symbol '$' in cmake/rpm.cmake
3. Added propagation of CMAKE_GENERATOR in dependencies that uses CMake
   for building, see [4]

How-to build wit Ninja:

$ cmake -G Ninja -B build -S .
$ ninja -C build/

1. https://ninja-build.org/
2. https://mesonbuild.com/Simple-comparison.html
3. https://stackoverflow.com/a/65803911/3665613
4. https://cmake.org/cmake/help/latest/module/ExternalProject.html

NO_DOC=internal
NO_CHANGELOG=internal
NO_TEST=internal