thd_destructor_proxy() may miss abort signal if innobase_end() is running
concurrently, which causes server hang in pthread_join() on shutdown.

The problem was that aborting wasn't protected by mutex:
proxy thr: while (!myvar->abort)
end thr: running->abort = 1;
end thr: mysql_cond_broadcast(...);
proxy thr: mysql_cond_wait(...); // nobody to awake it
end thr: pthread_join(...); // waits for proxy thr

Also made main.mysqld_option_err reentrant.

dict_table_t::prepare_instant(): Correctly assign instantly
dropped columns to instant->dropped[].

MDEV-11369 (instant ADD COLUMN) introduced a regression in the case the
leftmost leaf page of the clustered index is empty except for the
hidden metadata record.

btr_pcur_store_position(): If the only record in the leftmost leaf page
is the metadata record, store the position before the first user record
in the tree.

InnoDB does not allow creating multiple FULLTEXT INDEX
in ALGORITHM=INPLACE. This constraint was not being properly
enforced after MariaDB started to support ALGORITHM=INSTANT
and instant ADD COLUMN.

As a side effect of this bug, we again allow ALGORITHM=INPLACE
to rebuild a table when one FULLTEXT INDEX survives.
Also, we are returning a more accurate reason for refusing LOCK=NONE.

innobase_fulltext_exist(): Return the number of fulltext indexes.

ha_innobase::check_if_supported_inplace_alter(): If the table
needs to be rebuilt, refuse the operation if multiple fulltext
indexes would remain.

This assertion would fail when a secondary index record for an
instantly added column was accessed.

It is unclear to me why this code path is executed so rarely.
I was unable to cover it even when using FORCE INDEX.

row_sel_sec_rec_is_for_clust_rec(): Remove the assertion, and use
the proper function rec_get_nth_cfield().

row_sel_store_mysql_field_func(): Simply use rec_get_nth_cfield()
instead of duplicating its logic.

row_drop_table_for_mysql(): Fix a regression introduced in MDEV-16515.
Similar to the follow-up fixes MDEV-16647 and MDEV-17470, we must make
the internal tables of FULLTEXT INDEX immune to kills, to avoid noise
and resource leakage on DROP TABLE or ALTER TABLE. (Orphan internal tables
would be dropped at the next InnoDB startup only.)

call in sql_acl.cc fixed.

json_locate_key() implemented.
get rid of 'key_len' argument in functions.

server shutdown.

Prior to adding current thd to the global thd list, initialize
thd->mysys_var->current_cond / thd->mysys_var->current_mutex.

Otherwise thd_destructor_proxy can miss the abort condition, which is set
by the shutdown thread.

Type conversion bug fixed.

commit_cache_norebuild(): Restore the MDEV-17901 fix that was reverted
as part of the MDEV-18076/MDEV-18077 fix. It was not redundant after all.

Problem:
========
MLOG_FILE_WRITE_CRYPT_DATA redo log fails to apply type for
the crypt_data present in the space. While processing the double-write
buffer pages, page fails to decrypt. It leads to warning message.

Fix:
====
Set the type while parsing MLOG_FILE_WRITE_CRYPT_DATA redo log.
If type and length is of invalid type then mark it as corrupted.

json_t unittest added.

JSON api implementations and tests pushed.
sql_acl.cc fixed with the new function names.

Corrects 94d722b6

- There is no need to wait for crypt thread to stop accessing space while
dropping the garbage encrypted tablespace during recover.

dict_table_t::init_instant(): Correctly initialize the length of
variable-length instantly dropped columns.

row_ins_index_entry_set_vals(): For variable-length instantly dropped
columns, write 0 bytes of data. For dropped fixed-length NOT NULL
columns, write the fixed length of NUL bytes as data.

Allow table definition cache be bigger than open table cache (due to problem with VIEWs and prepared statements).

Implement according to standard SQL specification 2008.
The check_constraints table is used for fetching metadata about
the constraints defined for tables in all databases.

There were some result files which failed after running mtr.
These files are updated with newly create record with mtr --record.

This might apply to 10.1 and 10.2 as well, but I cannot run that test
there without installing third-party software. The test would expect
"innobackupex" to be in path. If I add a wrapper script that invokes
mariabackup --innobackupex, then on 10.2 the cluster would fail to start
up due to a version number mismatch.