Moving Item_func_spatial_rel from Item_bool_func to Item_bool_func2.
to make OP(const,field) use indexes.
- MBR functions supported OP(const,field) optimization in 10.0,
but were inintentionally broken in an earlier 10.1 change that introduced
a common parent for Item_func_spatial_mbr_rel and Item_func_spatial_precise_rel.
- Precise functions never supported optimization for OP(const,field).
Now both MBR and precise functions support OP(const,field) optimization.

any of the optimizer related functionality.

count_sargable_conds() instead for Item_func_in, Item_func_null_predicate,
Item_bool_func2. There other Item_int_func descendants that used to set
"sargable" to true (Item_func_between, Item_equal) already have their
own implementation of count_sargable_conds(). There is no sense to
have two parallel coding models for the same thing.

adding a separte class Item_func_spatial_relate for ST_RELATE().
This is a preparatory patch for:
 MDEV-8239 Reverse spatial operations OP(const, field) do not get optimized

Change session only variable enforce_storage_engine to be
session variable and make sure that also global value
is used if session variable is not set.

Item_func_eq's created during conversion of a ROW equality to a conjunction
of scalar equalities did not set cmp_context for its arguments properly,
so some of these created Item_func_eq could be later erroneously eliminated.

MDEV-8373 Zero date can be inserted in strict no-zero mode through CREATE TABLE AS SELECT timestamp_field

don't return NULL-s anymore, and actually they're not BOOLEAN.
Fixed.

The --gtid-ignore-duplicates option was not working correctly with row-based
replication. When a row event was completed, but before committing, there
was a small window where another multi-source SQL thread could wrongly try
to re-execute the same transaction, without properly ignoring the duplicate
GTID. This would lead to duplicate key error or out-of-order GTID error or
similar.

Thanks to Matt Neth for reporting this and giving an easy way to reproduce
the issue.

Introduce a new dummy INFORMATION_SCHEMA.CHANGED_PAGE_BITMAPS table to XtraDB
with reset_table callback to allow FLUSH NO_WRITE_TO_BINLOG CHANGED_PAGE_BITMAPS
to be called from innobackupex.

        The behaviour required by the standard seems too weird to expect.

This is an addition to original patch: cleanups for next_breadth_first_tab().

Overhead change:
next_breadth_first_tab()  0.04% -> out of radar

Split first_breadth_first_tab() into
JOIN::first_breadth_first_optimization_tab() and
JOIN::first_breadth_first_execution_tab().

This allows to eliminate function call and one condition. Adjusted callers
accordingly.

Overhead change:
first_breadth_first_tab()        0.07% -> out of radar
next_breadth_first_tab()         0.04% -> 0.04%
JOIN::cleanup()                  0.15% -> 0.11%
JOIN::save_explain_data_intern() 0.28% -> 0.24%

Moved Apc_target::destroy(), Apc_target::enable() and Apc_targe::disable()
definitions to my_apc.h so that they can be inlined.

Apc_targe::disable() now calls Apc_target::process_apc_requests() only if
there're APC requests. This saves one pthread_mutex_lock() call.

Overhead change:
Apc_target::disable              0.04% -> out of radar
Apc_target::enable               0.03% -> out of radar
Apc_target::process_apc_requests 0.02% -> out of radar
pthread_mutex_lock               0.43% -> 0.42%
pthread_mutex_unlock             0.26% -> 0.25%

        Some cases of the feature's borders were treated incorrectly.

The crash was caused by attempting to use a hashtable without
it getting initialized, due to the --skip-grant-tables switch.

ANALYZE should also record remove_duplicates() operation.

Handle the case where the optimizer decides to use
handler->delete_all_rows(), but then this call returns
HA_ERR_UNSUPPORTED and execution switches to regular
row-by-row deletion.

Do not call handler::rebind_psi() and handler::unbind_psi() when performance
schema is compiled out.

Overhead change:
handler::rebind_psi 0.04% -> out of radar
handler::unbind_psi 0.03% -> out of radar
open_table          0.21% -> 0.18%
close_thread_table  0.05% -> 0.05%

Initialize abs_timeout when it is about to be used. This saves one my_hrtime()
call on hot path (when we acquire MDL lock without waiting).

When filling I_S.PROCESSLIST use THD::start_utime/THD::utime_after_query instead
of THD::start_time. This allows us to save 2 clock_gettime() calls.

Overhead change:
__clock_gettime   0.13% -> 0.11% (122 -> 76 calls per OLTP RO transaction)
my_interval_timer 0.07% -> 0.06%
my_hrtime         0.04% -> 0.01%

Pass THD to Sort_and_group_tracker::report_sorting().
This reduces number of pthread_getspecific() calls from 292 to 290.

Pass THD to Item_field (and all derivatives) constructors.
This reduces number of pthread_getspecific() calls from 322 to 292.

Pass THD to find_all_keys() and Item_equal::Item_equal().
In MRR use table->in_use instead of current_thd.

This reduces number of pthread_getspecific() calls from 354 to 320.

Various fixes to let MariaDB compile with performance schema disabled.

Field_temporal_with_date::validate_for_get_date()
and reusing it in a few places.

Instead of asserting print informative error message to error log
and return failure from innodb_init causing the server to shutdown.

Added new dynamic configuration variable innodb_buf_dump_status_frequency
to configure how often buffer pool dump status is printed in the logs.
A number between [0, 100] that tells how oftern buffer pool dump status
in percentages should be printed. E.g. 10 means that buffer pool dump
status is printed when every 10% of number of buffer pool pages are
dumped. Default is 0 (only start and end status is printed).