there was a SELECT/DELETE race

Re-enable some Galera tests that should have been enabled.

Add client_ed25519.so to debian/libmariadb3.install;
merge e47a143f correctly.

Remove a duplicated #include from wsrep_mysqld.cc.

* do not allow versioned table to be without versioned (non-system) fields
* prohibit changing field versioning, when removing table versioning
* handle CREATE...SELECT as well

The setting innodb_change_buffering_debug=2 was supposed to inject
a crash during change buffer merge. There is no public test for
that functionality, and even if there were, it would be better
to use DEBUG_SYNC to halt the thread that does change buffer merge,
force a redo log flush from another thread, and finally kill the
server externally.

The most likely cause of the crash is that a timer fired, after it was closed.

MSDN documents such a possibility, in the documentation for
CloseThreadpoolTimer() function, and recommends disabling the timer before
calling WaitForThreadpoolTimerCallbacks()/CloseThreadpoolTimer().

The fix follows this recommendation.

Note, that 5.5-10.1 disabled the timer before close, but this code
was lost in threadpool refactoring in 10.2

MYSQL_MAINTAINER_MODE is ERR

Remove debug output,
remove overriding of the Windows C runtime flags(linker warning)
do not add code that depends on restsdk if library is not going
to be linked.

freaking Connect

For Visual Studio generator, use a per-config .def/.lib files with symbols
exported from mysqld.exe

Functions exported from mysqld.exe may differ between debug/optimized
compilation, e.g dbug functions are missing in release config.

After SST from master node (the one where event is ENABLED) - you will end up with the event enabled on two nodes, hence it's now being executed twice. It can be solved by comparing event's originator with server_id. if not equal, then change its status to 'SLAVESIDE_DISABLED'

Changes to be committed:
	new file:   mysql-test/suite/galera/r/galera_events2.result
	new file:   mysql-test/suite/galera/t/galera_events2.test
	modified:   sql/events.cc

Test changes only.

Enable after SST script changes.

Try to use more deterministic floating-point operations.
Apparently, 2.2 > 2.2 wrongly holds on many platforms, but
not ppc64le on the compiler used on Red Had Enterprise Linux 8.
The reason could be an infinite binary presentation:
2.2 = 0b10.001100110011…

With t1_f = 2.5 = 0b10.1, t1_f > 2.5 would no longer hold on AMD64.
Let us replace the 2.2 with 2.5 and compare t1_f >= 2.5 in order to
get more consistent results across all platforms.

have the same path as in source rpms, don't build them differently

Part3: MDEV-18156 Assertion `0' failed or `btr_validate_index(index, 0, false)' in row_upd_sec_index_entry or error code 126: Index is corrupted upon DELETE with PAD_CHAR_TO_FULL_LENGTH

Don't break compatibility in GA releases.
Warn the user, but don't refuse to create a table until 10.5

With cmake -DWITH_INNODB_AHI=OFF -DCMAKE_BUILD_TYPE=RelWithDebInfo
the variable 'i' in fseg_free_extent() was declared but not used.

The test fails because it reuses mysqltest perl code to copy directory
tree, and this code contains Windows-specific piece which outputs some
diagnostic information.

The patch introduces new parameter for that Windows-specific perl code to
have the ability to suppress diagnostic output on the corresponding
mysqltest perl module initialization.

Make the test stable: after DROP TABLE, make sure the compaction is
run and finishes.

If we don't do this, the post-drop compaction may run during the next
testcase. It will cause a record from the next testcase to be compacted
away when the test logic doesn't expect it and the test will fail

trx_t::is_recovered: Revert most of the changes that were made by the
merge of MDEV-15326 from 10.2. The trx_sys.rw_trx_hash and the recovery
of transactions at startup is quite different in 10.3.

trx_free_at_shutdown(): Avoid excessive mutex protection. Reading fields
that can only be modified by the current thread (owning the transaction)
can be done outside mutex.

trx_t::commit_state(): Restore a tighter assertion.

trx_rollback_recovered(): Clarify why there is no potential race condition
with other transactions.

lock_trx_release_locks(): Merge with trx_t::release_locks(),
and avoid holding lock_sys.mutex unnecessarily long.

rw_trx_hash_t::find(): Remove redundant code, and avoid starving the
committer by checking trx_t::state before trx_t::reference().

this is not ideal and needs to be fixed eventually,
but it's consistent over all forms of INSERT.

MDEV-20403 Assertion `0' or Assertion `btr_validate_index(index, 0)' failed in row_upd_sec_index_entry or error code 126: Index is corrupted upon UPDATE with TIMESTAMP..ON UPDATE

remove a special treatment of a bare DEFAULT keyword that made it
behave inconsistently and differently from DEFAULT(column).
Now all forms of the explicit assignment of a default column value
behave identically, and all count as an explicitly assigned value
(for the purpose of ON UPDATE NOW).

followup for c7c481f4

Fix the test case.

MDEV-20079 When setting back the system time while mysqld is running, NOW() and UNIX_TIMESTAMP() results get stuck

typo. system_time.start wasn't updated when system_time.sec
and system_time.sec_part were.

MDEV-5817 query cache bug (returning inconsistent/old result
set) with aria table parallel inserts, row format = page

The problem is that for transactional aria tables
(row_type=PAGE and transactional=1), maria_lock_database()
didn't flush the state or the query cache.
Not flushing the state is correct for transactional tables as
this is done by checkpoint, but not flushing the query cache
was wrong and could cause concurrent SELECT queries to not
be deleted from the cache.

Fixed by introducing a flush of the query cache as part of commit, if the table has changed.
t for transactional aria tables (row_type=PAGE and transactional=1), maria_lock_table() didn't flush their state or the query cache.