- Skip the undo logging only for the newly added partition.

fixes srpm builds on centos7 and other old platforms (with cmake < 3.1)
fixes pre-uninstall script
fixes parallel builds

1. Store assignment failures on incompatible data types now raise errors if:
- STRICT_ALL_TABLES or STRICT_TRANS_TABLES sql_mode is used, and
- IGNORE is not used

Otherwise, only a warning is raised and the statement continues.

2. Changing the error/warning test as follows:

-ERROR HY000: Illegal parameter data types inet6 and int for operation 'SET'
+ERROR HY000: Cannot cast 'int' as 'inet6' in assignment of `db`.`t`.`col`

so in case of a big table it's easier to see which column has the problem.
The new error text is also applied to SP variables.

A message used to say "failed to read or decrypt"
but the "or decrypt" part was removed in
commit 0b47c126
without adjusting rarely needed error message suppressions in some
encryption tests.

Let us improve the error message so that it mentions the file name,
and adjust all error message suppressions in tests.

Thanks to Oleksandr Byelkin for noticing one test failure.

Make SEL_ARG::make_root() maintain SEL_ARG::weight.

Also, an unrelated change: fix dbug_print_sel_arg() to correctly
print SQL NULL for the right endpoint.

Disable for embedded as mysqldump cannot connect to embedded server.

The test mariabackup.mdev-14447 is inserting relatively much
data while concurrently backing up the data. The test often fails
on CI systems, possibly due to an inherent race condition between
the producer (server) and consumer (backup) that would be solved
if the backup was being produced by the server (MDEV-14992).

The written data volume was increased somewhat by
commit 4179f93d (MDEV-18976).
Let us trim the log volume by not writing PAGE_CHECKSUM records
or row-level undo log records, and make the test cover what was
intended to cover by creating the table in the system tablespace.

Part #2: make sure we allocate space for two JOIN_TABs that
use temporary tables.
The dbug_join_tab_array_size is still set to catch cases where
we try to access more JOIN_TAB object than we thought we would have.

The problem was caused by use of COLLATION(AVG('x')). This is an
item whose value is a constant.
Name Resolution code called convert_const_to_int() which removed AVG('x').
However, the item representing COLLATION(...) still had with_sum_func=1.

This inconsistent state confused the code that handles grouping and
DISTINCT: JOIN::get_best_combination() decided to use one temporary
table and allocated one JOIN_TAB for it, but then
JOIN::make_aggr_tables_info() attempted to use two and made writes
beyond the end of the JOIN::join_tab array.

The fix:
- Do not replace constant expressions which contain aggregate functions.
- Add JOIN::dbug_join_tab_array_size to catch attempts to use more
  JOIN_TAB objects than we've allocated.

- query->intersection fails to get freed if the query exceeds
innodb_ft_result_cache_limit

- errors from init_ftfuncs were not propogated by delete command

This is taken from percona/percona-server@ef2c0bcb9a34aeb06de0058d7c2a2969416b35a7

with its 0049-Selectively-disallow-SHA1-signatures.patch
in the openssl source rpm.

let's allow them for now, this fixes tls_version and tls_version1 tests.

This bug manifested itself for INSERT...SELECT and DELETE statements whose
WHERE condition used an IN/ANY/ALL predicand or a EXISTS predicate with
such grouping subquery that:
 - its GROUP BY clause could be eliminated,
 - the GROUP clause contained a subquery over a mergeable derived table
   referencing the updated table.

The bug ultimately caused a server crash when the prepare phase of the
statement processing was executed. This happened after removal redundant
subqueries used in the eliminated GROUP BY clause from the statement tree.
The function that excluded the subqueries from the did not do it properly.
As a result the specification of any derived table contained in a removed
subquery was not marked as excluded.

Approved by Oleksandr Byelkin <sanja@mariadb.com>

mysqlimport starts many worker threads. when one of the worker
encounters an error, it frees global memory and calls exit().

it suppresses memory leak detector, because, as the comment says
"dirty exit, some threads are still running", indeed, it cannot
free the memory from other threads.

but precisely because some threads are still running, they
might use this global memory, so it cannot be freed.

fix: if we know that some threads are still running and accept
that we cannot free all memory anyway, let's not free global
allocations either

This is particularly important for Azure where there is no
MyISAM support in their MariaDB cloud product.

Like mysqldumper does, a view can satisfy the requirement
like a table, without constraints. The views in frm files are
text form and don't have column limits.

Thanks Thomas Casteleyn for the suggestion.

MDEV-29078 For old binary logs explicit_defaults_for_timestamp presumed to be OFF, server value ignored

don't assume anymore that OPTIONS_WRITTEN_TO_BIN_LOG is fixed once
and forever. Instead, deduct master's OPTIONS_WRITTEN_TO_BIN_LOG
from the master's version in binlog.

save it in binlog, so that CREATE TABLE could be replayed correctly
on the slave

make @@explicit_defaults_for_timestamp session variable

print "Defaults to on" not only for GET_BOOL, but also for GET_BIT

list ${OPENSSL_ROOT_DIR}/lib64 explicitly, because
cmake below version 3.23.0 won't search there.

With a global non-default max-statement-time of a time interval that exceed
the query time mysqldump queries when doing a backup.

To solve both, add a max-statement-time option, defaulting to 0 (unlimited time).

Also like mariabackup, set the session wait_timeout=DEFAULT (28800). The
time/processing between mysqldump times isn't expected to get that
close ever, but let's adopt the standard of mariabackup as no-one has
challenged it has having a detrimental effect.

Reviewer and test case author Daniel Black