It turns out, Field::set_notnull is required for a nullable I_S field.
Rework thread_pool_info.test to test request queueing in threadpool.

If --thread-pool-dedicated-listener is set, worker should not pick up
events. Dedicated listener constantly drains all events, thus polling
also from another thread makes no sense.

splitted permission setting to the 2 separate calls.
according to execute_process() doc, output from first command will be
piped to second which led to race and randomly lost suid bit from
auth_pam_tool.

foreign_push_index_error(): Fix a typo of 'match' that was introduced
in commit 5130f520 (MDEV-20480).

Thanks to Oli Sennhauser for reporting this.

Post-push fix for 10.5+.

The fix influence MDEV-14479. Before the fix
lock_rec_convert_impl_to_expl() did not create explicit lock if caller's
transaction owns found implicit lock(see MDEV-14479 for details). After the fix
lock_rec_convert_impl_to_expl() can create explicit lock under the above
conditions if the requested lock mode is not LOCK_REC_NOT_GAP. And that
is why we need to check if the table is X-locked before
lock_rec_convert_impl_to_expl() call.

The merge commit 4a259572
caused a test failure on Windows. The suppression regexp
needs to accept the backslash.

fil_invalid_page_access_msg(): Simplify the implementation
and invoke sql_print_error() directly.

fil_space_t::io(): Invoke fil_invalid_page_access_msg() only from
one location.

PageConverter::update_index_page(): Always validate the PAGE_INDEX_ID.
Failure to do so could cause a crash when iterating
secondary index pages. This was caught by the 10.4 test
innodb.full_crc32_import.

Delete-marked record is on the secondary index and the clustered index
already purged the corresponding record. We cannot detect if such
record is historical and we should not: the algorithm of
row_ins_check_foreign_constraint() skips such record anyway.

Do swap_blobs() for new partition_read_multi_range mode.

Due to an integer overflow an invalid size of ref_pointer_array could be
allocated.

Using size_t allows this continue. Allocation failures are
handled gracefully if the value is too big.

Thanks to Zuming Jiang for the bug report and fuzzing MariaDB.

Reviewer: Sanja

A test case to reproduce the issue. The actual fix is in galera
library.
Reviewed-by: Jan Lindström <jan.lindstrom@mariadb.com>

MDEV-18734 FIXME: vcol.partition triggers ASAN heap-use-after-free

MDEV-26223 Galera cluster node consider old server_id value even after modification of server_id [wsrep_gtid_mode=ON]

If cluster is bootstrapped in existing database, we should use provided
configuration variables for wsrep_gtid_domain_id and server_id instead
of recovered ones.
If 'new' combination of wsrep_gtid_domain_id & server_id already existed
somewere before in binlog we should continue from last seqno, if
combination is new we start from seqno 0.
Reviewed-by: Jan Lindström <jan.lindstrom@mariadb.com>

Contains following fixes:

* allow TOI commands to timeout while trying to acquire TOI with
override lock_wait_timeout with a LONG_TIMEOUT only after
succesfully entering TOI
* only ignore lock_wait_timeout on TOI
* fix galera_split_brain test as TOI operation now returns ER_LOCK_WAIT_TIMEOUT after lock_wait_timeout
* explicitly test for TOI
Reviewed-by: Jan Lindström <jan.lindstrom@mariadb.com>

Main idea: don't log-and-crash but propogate error to the upper layers of stack
to handle it and show to a user.

If lock type is LOCK_GAP or LOCK_ORDINARY, and the transaction holds
implicit lock for the record, then explicit gap-lock will not be set for
the record, as lock_rec_convert_impl_to_expl() returns true and
lock_rec_convert_impl_to_expl() bypasses lock_rec_lock() call.

The fix converts explicit lock to implicit one if requested lock type is
not LOCK_REC_NOT_GAP.

innodb_information_schema test result is also changed as after the fix
the following statements execution:

SET autocommit=0;
INSERT INTO t1 VALUES (5,10);
SELECT * FROM t1 FOR UPDATE;

leads to additional gap lock requests.

Creation of tables by the three names mysql.user, mysql.host, mysql.db
was being blocked in the function row_mysql_is_system_table().

Since commit 4abb8216 (MDEV-17658),
mysql.user is a view, not a table. Since commit
ead9a34a (MDEV-15851), mysql.host
is not being created at all.

Let us remove the special handling of table names in InnoDB,
and allow mysql.db to be created in InnoDB. The special handling
was originally added in commit e84ef2b7
without any explanation.

Reviewed by: Sergei Golubchik

InnoDB DDL fails when it tries to sync the table
when innodb_force_recovery is set to 2. Problem
is that fts_optimize_wq is not initialized when
there are no background threads. fts_sync_during_ddl()
should check whether fts_optimize_wq is initialized.

Import operation without .cfg file fails when there is mismatch of index
between metadata table and .ibd file. Moreover, MDEV-19022 shows
that InnoDB can end up with index tree where non-leaf page has only
one child page. So it is unsafe to find the secondary index root page.

This patch does the following when importing the table without .cfg file:

1) If the metadata contains more than one index then InnoDB stops
the import operation and report the user to drop all secondary
indexes before doing import operation.

2) When the metadata contain only clustered index then InnoDB finds the
index id by reading page 0 & page 3 instead of traversing the
whole tablespace.

pars_info_bind_id(): Remove the parameter copy_name. It was always
being passed as constant TRUE or true. It turns out that copying
the string is completely unnecessary. In all calls except the one
in fts_get_select_columns_str() and fts_doc_fetch_by_doc_id(),
the parameter is being passed as a compile-time constant, and therefore
the pointer cannot become stale. In that special call, the string
that is being passed is allocated from the same memory heap that
pars_info_bind_id() would have been using.

pars_info_add_id(): Remove (unused declaration).

Fix for previous commit, shell logic for repeat configure
stages corrected to use subshell.

This commit fixes a call to the sockstat utility for FreeBSD,
where this utility requires an extra "-s" parameter to display
the connection status and prints one extra column.

This commit fixes a mistake where the --whole-file option
is used by rsync SST in WAN mode instead of LAN.

This commit adds support for reading new SSL configuration
options (ssl-ca, ssl-cert and ssl-key) if the [sst] section
with old options (tca, tcert and tkey) is missing in the config
file, even if not specified authentication mode via the ssl-mode
option. Before this change, new parameters were read only if the
ssl-mode option was present in the configuration file and it was
not equal to the 'DISABLED' value.

Also added diagnostics (information level) which warns the user
that due to the presence of the tca, tcert and/or tkey parameters
in the [sst] section, new SSL configuration options will be ignored
(if their values do not match the old ones).

Observed in 10.4 however same code in 10.2

mariadb-server-10.4/tests/mysql_client_test.c:18209:5: error: this ‘if’ clause does not guard... [-Werror=misleading-indentation]
18209 |     if (!opt_silent)
      |     ^~
In file included from mariadb-server-10.4/tests/mysql_client_test.c:38:
mariadb-server-10.4/tests/mysql_client_fw.c:133:9: note: ...this statement, but the latter is misleadingly indented as if it were guarded by the ‘if’
  133 |         ((void) ((expr) ? 0 : (die(__FILE__, __LINE__, #expr), 0)))
      |         ^
mariadb-server-10.4/tests/mysql_client_test.c:18212:7: note: in expansion of macro ‘DIE_UNLESS’
18212 |       DIE_UNLESS(tm[i].year == 0);
      |       ^~~~~~~~~~

$ /usr/bin/cc --version
cc (GCC) 11.2.1 20210728 (Red Hat 11.2.1-1)

Problem:
=======
There are two issues that are addressed in this patch:
 1) SHOW BINARY LOGS uses caching to store the binary logs that exist
in the log directory; however, if new events are written to the logs,
the caching strategy is unaware. This is okay for users, as it is
okay for SHOW to return slightly old data. The test, however, can
result in inconsistent data. It runs two connections concurrently,
where one shows the logs, and the other adds a new file. The output
of SHOW BINARY LOGS then depends on when the cache is built, with
respect to the time that the second connection rotates the logs.
 2) There is a race condition between RESET MASTER and SHOW BINARY
LOGS. More specifically, where they both need the binary log lock to
begin, SHOW BINARY LOGS only needs the lock to build its cache. If
RESET MASTER is issued after SHOW BINARY LOGS has built its cache and
before it has returned the results, the presented data may be
incorrect.

Solution:
========
 1) As it is okay for users to see stale data, to make the test
consistent, use DEBUG_SYNC to force the race condition (problem 2) to
make SHOW BINARY LOGS build a cache before RESET MASTER is called.
Then, use additional logic from the next part of the solution to
rebuild the cache.
 2) Use an Atomic_counter to keep track of the number of times RESET
MASTER has been called. If the value of the counter changes after
building the cache, the cache should be rebuilt and the analysis
should be restarted.

Reviewed By:
============
Andrei Elkin: <andrei.elkin@mariadb.com>

Columnstore badly failed on 32bit. The way Debian triggers
somehow doesn't detect the amd64 in the architecture of columnstore
so we explicitly disable it to prevent failures on x86_32.

The architecture from the control file is sufficient to not build
of arm64 and other unsupported achitectures so we don't need to
disable columnstore by default.

The logic around not building columnstore on Travis/Gitlab ci
can be preserved with a autobake-deb.sh restructure.

https://github.com/wolfSSL/wolfssl/issues/2828 lists 4.6.0 a
as the fixed version as we currently have 4.8.0.

Since the time of the above issue, Debian has allowed OpenSSL
linking.

This was a regression introduced in MDEV-18353, where to CONNECT objects
were incorrectly counted for named pipes.

Audit-plugin related fixes.

Add server functions to provide necessary data.