This was forgotten in e9de6386

log_group_max_size: is not needed because redo log do not use fil_io() now

if my_realpath() fails, don't return the error code, get_defaults_options()
returns a number of options consumed, not 0=ok/1=error.

instead, ignore the error from my_realpath. If it fails it internally
falls back to my_load_path, which restores 10.4- behavior

Old temporal data types (created with a pre-10.0 version of MariaDB)
are now displayed with a /* mariadb-5.3 */ comment in:

- SHOW CREATE TABLE
- DESCRIBE
- INFORMATION_SCHEMA.COLUMNS.COLUMN_TYPE

For example:

CREATE TABLE `t1` (
  `t0` datetime /* mariadb-5.3 */ DEFAULT NULL,
  `t6` datetime(6) /* mariadb-5.3 */ DEFAULT NULL
) ENGINE=MyISAM DEFAULT CHARSET=latin1

Note, new temporal data types are displayed without a format comment.

followup for 3a3605f4

* It uses stack a lot, limit the recursion depth like
  we used to do for pcre.
* But don't do that for newer pcre2 that uses stack very little
  (and doesn't support limiting recursion depth anyway).
* Fix tests to verify only that deep recursion doesn't crash, but
  ignore results and warnings (which depend on pcre2 version)
* Also different pcre2 versions report different offset in error messages,
  take that into account.

Debian is apparently offended that pcre2-posix implements POSIX API,
thus it renames all posix-compatible symbols in libpcre2-posix to have the
PCRE2 prefix. But Debian doesn't do anything to pcre2posix.h header,
so any unaware application will get POSIX compatible type names
and function prototypes from pcre2, but actual symbols will come
from libc.

To remedy this enormous incongruity we have to redefine POSIX-compatible
function names in pcre2posix to match Debian's hack.

prevent potential stalls for when all slots are used up, and all io threads
are used up, and io callbacks are slow.

Removing redundant space character after the data type comment
in SHOW CREATE TABLE, so the output changes from e.g.:

  a TIME /* mariadb-5.3 */  DEFAULT NULL

to

  a TIME /* mariadb-5.3 */ DEFAULT NULL

This is a prerequisite for MDEV-19906.

- Making classes Field_time, Field_datetime, Field_timestamp abstract
- Adding instantiable Field_time0, Field_datetime0, Field_timestamp0 classes
- Removing redundant cast in field_conv.cc, item_timefunc.cc, sp.cc in calls for set_time() and get_timestamp()
- Replacing store_TIME() to store_timestamp() in log.cc and removing redundant cast

With MDEV-16678, some InnoDB background tasks will acquire MDL,
and we must filter out such requests in the test.
For this test, an alternative might have been to use wait_all_purged.inc.

1. Fix places where data race warnings were relevant.

tls_worker_data::m_state should be modified under mutex protection,
since both maintainence timer and current worker set this flag.

2. Suppress warnings that are legitimate, yet harmless.
Apparently, the dirty reads in waitable_task::get_ref_count() or
write_slots->pending_io_count()

Avoiding race entirely without side-effects here is tricky,
and the effects of race is harmless.

The worst thing that can happen due to race is an extra wait notification,
under rare circumstances.

srv_purge_wakeup() is called under protection of latch in
purge_sys_t::resume()

move sync check code before acquiring that latch.

Now that FIL_TYPE_LOG no longer exists, let us remove some redundant
comparisons of fil_space_t::purpose.

fil_flush_low(): Do not distinguish FIL_TYPE_TEMPORARY.

fil_space_get_space(): Assume that space->purpose is valid.

The merge commit 68fe5f53 missed this.

page_set_autoinc(): Check if any change would take place,
and omit the mtr_t::OPT parameter. This should avoid
unnecessarily redo log writes for ROW_FORMAT=COMPRESSED pages.

The merge a8ed0f77
was accidentally a null-merge.

MDEV-18046: Assortment of crashes, assertion failures and ASAN errors in mysql_show_binlog_events

MDEV-18046: Assortment of crashes, assertion failures and ASAN errors in mysql_show_binlog_events

Problem:
========
SHOW BINLOG EVENTS FROM <pos> reports following assert when ASAN is enabled.

uint32 binlog_get_uncompress_len(const char*):
  Assertion `(buf[0] & 0xe0) == 0x80' failed

Fix:
===
**Part11: Converted debug assert to error handler code**

Problem:
========
SHOW BINLOG EVENTS FROM <pos> reports following ASAN error.

AddressSanitizer: heap-buffer-overflow on address
READ of size 1 at 0x60e00009cf71 thread T28
#0 0x55e37e034ae2 in net_field_length

Fix:
===
**Part10: Avoid reading out of buffer**

Problem:
========
SHOW BINLOG EVENTS FROM <pos> reports following assert when ASAN is enabled.

Query_log_event::Query_log_event(const char*, uint,
    const Format_description_log_event*, Log_event_type):
  Assertion `(pos) + (6) <= (end)' failed

Fix:
===
**Part9: Removed additional DBUG_ASSERT**

Problem:
========
SHOW BINLOG EVENTS FROM <pos> reports following ASAN error

AddressSanitizer: SEGV on unknown address
The signal is caused by a READ memory access.
User_var_log_event::User_var_log_event(char const*, unsigned int,
    Format_description_log_event const*)

Implemented part of upstream patch.
commit: mysql/mysql-server@a3a497ccf7ecacc900551fb1e47ea4078b45c351

Fix:
===
**Part8: added checks to avoid reading out of buffer limits**

Problem:
========
SHOW BINLOG EVENTS FROM <pos> reports following ASAN error
"heap-buffer-overflow on address" and some times it asserts.

Table_map_log_event::Table_map_log_event(const char*, uint,
    const Format_description_log_event*)
Assertion `m_field_metadata_size <= (m_colcnt * 2)' failed.

Fix:
===
**Part7: Avoid reading out of buffer**


Converted debug assert to error handler code.

Problem:
========
SHOW BINLOG EVENTS FROM <pos> reports following ASAN error

AddressSanitizer: heap-buffer-overflow on address 0x60400002acb8
Load_log_event::copy_log_event(char const*, unsigned long, int,
    Format_description_log_event const*)

Fix:
===
**Part6: Moved the event_len validation to the begin of copy_log_event function**

Problem:
========
SHOW BINLOG EVENTS FROM <pos> reports following ASAN error

AddressSanitizer: heap-buffer-overflow on address
String::append(char const*, unsigned int)
Query_log_event::pack_info(Protocol*)

Fix:
===
**Part5: Added check to catch buffer overflow**

Problem:
========
SHOW BINLOG EVENTS FROM <pos> reports following ASAN error

heap-buffer-overflow within "my_strndup" in Rotate_log_event

my_strndup /mysys/my_malloc.c:254
Rotate_log_event::Rotate_log_event(char const*, unsigned int,
    Format_description_log_event const*)

Fix:
===
**Part4: Improved the check for event_len validation**

Problem:
========
SHOW BINLOG EVENTS FROM <pos> reports following crash when ASAN is enabled.

SEGV on unknown address
in inline_mysql_mutex_destroy
in my_bitmap_free
in Update_rows_log_event::~Update_rows_log_event()

Fix:
===
**Part3: Initialize m_cols_ai.bitmap to NULL**