We can't expect that users want to always convert their mysqld_safe
settings on an upgrade. In its current form it will always run, and that
seems unnecessary on every single installation.

Also the script is buggy, leaks mysqld_safe output into the written file
and since it gets syntax errors the whole mariadb.service will fail to
work.

These do now show up automatically due to init and systemd customizations,
so the handiest fix is to add them manually. This has been the praxis
in downstream MariaDB packaging for a couple years now, and works fine.

- Ensure service is loaded and started after installation,
  (fixes service start issues in Debian/Ubuntu upgrades where
  otherwise service mysql status stayed stopped)

- Ensure service stopped before removal/purge
  (fixes unstopped processes detected by piuparts)

- Ensure systemd daemon is reloaded after removal/purge when service
  has been removed

Partially reverts commit a4cc6fb9.

While all current versions of Linux have systemd, support for traditional
init.d is still needed e.g. on Linux subsystem on Windows, kFreeBSD and
special variants of Debian/Ubuntu that for other reasons don't have
systemd.

Thus, re-introduce the init file that was remove, but this time with
then name 'mariadb'.

Supporting traditional sysv init in paraller with systemd is easy, since
Debian has facilities for it.

Also simplify and update salsa-ci.yml install/upgrade testing works
for all previous MariaDB and MySQL releases without any excess quirks.

Note that in fresh installs the salsa-ci.yml needs to run command
'service mariadb status' to control the service, while on upgrades
it is enough to run 'service mysql status', since the init.d/mysql
file is left behind from previous install, along with some other
config files such as /etc/default/mysql and /etc/mysql/* stuff.

Apparently, in stats_reset_table(), the innocuous

  memset(&group->counters, 0, sizeof(group->counters));

is converted by clang to SSE2 instructions.

The problem is that "group" is not correctly aligned,
despite  MY_ALIGNED(CPU_LEVEL1_DCACHE_LINESIZE) in the thread_group_t
declaration.

It is not aligned because it was allocated with my_malloc, since
commit fd9f1638, MDEV-5205. Previously all_groups was a
statically allocated array.

Fix is to remove MY_ALIGNED, and pad the struct instead.

Same array instance in two Item_func_in instances. First Item_func_in
instance is freed on table close. Second one is freed on
cleanup_after_query().

get_copy() depends on copy ctor for copying an item and hence does
shallow copy for default copy ctor. Use build_clone() for deep copy of
Item_func_in.

Most of the volations came from:
sel_arg_range_seq_next(void*, st_key_multi_range*) (opt_range_mrr.cc:342)

Problem was that trx->lock.was_chosen_as_wsrep_victim variable was
not set back to false after it was set true.

wsrep_thd_bf_abort
	Add assertions for correct mutex status and take necessary
	mutexes before calling thd->awake_no_mutex().

innobase_rollback_trx()
	Reset trx->lock.was_chosen_as_wsrep_victim

wsrep_abort_slave_trx()
	Removed unused function.

wsrep_innobase_kill_one_trx()
	Added function comment, removed unnecessary parameters
	and added debug assertions to enforce correct usage. Added
	more debug output to help out on error analysis.

wsrep_abort_transaction()
	Added debug assertions and removed unused variables.

trx0trx.h
	Removed assert_trx_is_free macro and replaced it with
	assert_freed() member function.

trx_create()
	Use above assert_free() and initialize wsrep variables.

trx_free()
	Use assert_free()

trx_t::commit_in_memory()
	Reset lock.was_chosen_as_wsrep_victim

trx_rollback_for_mysql()
	Reset trx->lock.was_chosen_as_wsrep_victim

Add test case galera_bf_kill

When server is compiled with recent VS2019, then executables,
have dependency on  vcruntime140_1.dll

While we include the VC redistributable merge modules into our MSI package,
those merge modules were stale (taken from older VS version, 2017)

Since VS2019 brough new DLL dependency by introducing new exception handling
https://devblogs.microsoft.com/cppblog/making-cpp-exception-handling-smaller-x64
thus the old MSMs were not enough.

The fix is to change logic in win/packaging/CMakeLists.txt to look up for
the correct, new MSMs.

The bug only affects 10.4,as we compile with static CRT before 10.4,
and partly-statically(just vcruntime stub is statically linked, but not UCRT)
after 10.4

For the fix to work, it required also some changes on the build machine
(vs_installer, modify VS2019 installation, add Individual Component
"C++ 2019 Redistributable MSMs")

This reverts commit c675886d.

The bug was introduced in e3f5789a and fixed in 1af74d52

The option was introduced between those two commits and now it's
redundant again.

Type_handler_xxx::Item_const_eq() can handle only non-NULL values.
The code in Item_basic_value::eq() did not take this into account.

Adding a test to detect three different combinations:
- Both values are NULLs, return true.
- Only one value is NULL, return false.
- Both values are not NULL, call Type_handler::Item_const_eq()
  to check equality.

Flush LSN to system tablespace on innodb shutdown if XA is rolled back by
mariabackup.

innobase_get_charset(), innobase_get_stmt_safe(): Remove.
It is more efficient and readable to invoke thd_charset()
and thd_query_safe() directly, without a non-inlined wrapper function.

The function thd_query_safe() is used in the implementation of the
following INFORMATION_SCHEMA views:

    information_schema.innodb_trx
    information_schema.innodb_locks
    information_schema.innodb_lock_waits
    information_schema.rocksdb_trx

The implementation of the InnoDB views is in trx_i_s_common_fill_table().
This function invokes trx_i_s_possibly_fetch_data_into_cache(),
which will acquire lock_sys->mutex and trx_sys->mutex in order to
protect the set of active transactions and explicit locks.
While holding those mutexes, it will traverse the collection of
InnoDB transactions. For each transaction, thd_query_safe() will be
invoked.

When called via trx_i_s_common_fill_table(), thd_query_safe()
is acquiring THD::LOCK_thd_data while holding the InnoDB locks.
This will cause a deadlock with THD::awake() (such as executing
KILL QUERY), because THD::awake() could invoke lock_trx_handle_wait(),
which attempts to acquire lock_sys->mutex while already holding
THD::lock_thd_data.

thd_query_safe(): Invoke mysql_mutex_trylock() instead of
mysql_mutex_lock(). Return the empty string if the mutex
cannot be acquired without waiting.

Not only Ubuntu Focal builds openssl with OPENSSL_TLS_SECURITY_LEVEL=2,
but for some unfathomable reason it patches openssl sources to disable
TLS < 1.2 at security level 2, even though openssl manual says it
should only happen at level 4:
https://www.openssl.org/docs/man1.1.0/man3/SSL_CTX_set_security_level.html

We test TLSv1.1 and TLSv1, so we have to override Focal defaults in mtr.

As part of the SPATIAL INDEX implementation in InnoDB,
dict_index_t was expanded by a rtr_ssn_t field. There are only
3 operations for this field, all protected by rtr_ssn_t::mutex:

* btr_cur_search_to_nth_level() stores the least significant 32 bits
of the 64-bit value that is stored in the index root page.
(This would better be done when the table is opened for the
very first time.)
* rtr_get_new_ssn_id() increments the value by 1.
* rtr_get_current_ssn_id() reads the current value.

All these operations can be implemented equally safely by using
atomic memory access operations.

setup alternative name for rhel8/centos8

use correct C/C version when providing mariadb-connector-c
and other changes for installation over distro packages to work:

  dnf install mariadb-server
  dnf install --allowerasing MariaDB-{server,common,client,shared}*.rpm

The code did not take into account that:
- U+005C (backslash) can occupy more than mbminlen characters (e.g. in sjis)
- Some character sets do not have a code for U+005C (e.g. swe7)

Adding a new function my_wc_to_printable into MY_CHARSET_HANDLER to
cover all special cases easier.