Let lock_print_info_all_transactions() iterate rw_trx_hash instead of
rw_trx_list.

When printing info of locks for transactions, InnoDB monitor doesn't
attempt to read relevant page from disk anymore. The code was prone
to race conditions.

Note that TrxListIterator didn't work as advertised: it iterated
rw_trx_list only.

Let trx_rollback_recovered() iterate rw_trx_hash instead of rw_trx_list.

Let lock_validate_table_locks(), lock_rec_other_trx_holds_expl(),
lock_table_locks_lookup(), trx_recover_for_mysql(), trx_get_trx_by_xid(),
trx_roll_must_shutdown(), fetch_data_into_cache() iterate rw_trx_hash
instead of rw_trx_list.

Removed trx_sys_validate_trx_list(): with rw_trx_hash elements are not
required to be ordered by transaction id. Transaction state is now guarded
by asserts in rw_trx_hash_t.

Removed trx_sys_t::n_prepared_recovered_trx: never used.

Removed trx_sys_t::n_prepared_trx: used only at shutdown, we can perfectly
get this value from rw_trx_hash.

Determine minimum transaction id by iterating rw_trx_hash, not rw_trx_list.

It is more expensive than previous implementation since it does linear
search, especially if there're many concurrent transactions running. But in
such case mutex is much bigger evil. And since it doesn't require
trx_sys->mutex protection it scales better.

For low concurrency performance difference is neglible.

Replaced UT_LIST_GET_LEN(trx_sys->rw_trx_list) with
trx_sys->rw_trx_hash.size().
Moved freeing of trx objects at shutdown to rw_trx_hash destructor.
Small clean-up in trx_rollback_recovered().

Reduce divergence between trx_sys_t::rw_trx_hash and trx_sys_t::rw_trx_list
by not adding recovered COMMITTED transactions to trx_sys_t::rw_trx_list.

Such transactions are discarded immediately without creating trx object.

This also required to split rollback and cleanup phases of recovery. To
reflect these updates the following renames happened:
trx_rollback_or_clean_all_recovered() -> trx_rollback_all_recovered()
trx_rollback_or_clean_is_active -> trx_rollback_is_active
trx_rollback_or_clean_recovered() -> trx_rollback_recovered()
trx_cleanup_at_db_startup() -> trx_cleanup_recovered()

Also removed a hack from lock_trx_release_locks(). Instead let recovery
rollback thread to skip committed XA transactions.

when number of NULLs in IN list reaches in_predicate_conversion_threshold

The bug was fixed by removing an assertion that had been set in order
just to test whether the code could be ever executed.

If InnoDB is killed while ALTER TABLE...ALGORITHM=COPY is in progress,
after recovery there could be undo log records some records that were
inserted into an intermediate copy of the table. Due to these undo log
records, InnoDB would resurrect locks at recovery, and the intermediate
table would be locked while we are trying to drop it. This would cause
a call to row_rename_table_for_mysql(), either from
row_mysql_drop_garbage_tables() or from the rollback of a RENAME
operation that was part of the ALTER TABLE.

row_rename_table_for_mysql(): Do not attempt to parse FOREIGN KEY
constraints when renaming from #sql-something to #sql-something-else,
because it does not make any sense.

row_drop_table_for_mysql(): When deferring DROP TABLE due to locks,
do not rename the table if its name already starts with the #sql-
prefix, which is what row_mysql_drop_garbage_tables() uses.
Previously, the too strict prefix #sql-ib was used, and some
tables were renamed unnecessarily.

Follow-up to commit 9ec19b9b

instead of skipping invalid items in setup_conds(),
don't pass them into a JOIN at all

(test case in versioning.select2)

Revert "MDEV-14786 Server crashes in Item_cond::transform on 2nd execution of SP querying from a view [fixes #436]"

This reverts commit 7069071d

And add a test to show that optimization steps that
a) are repeated for every execution
b) create new items
cannot be done on the statement arena

srv_prepare_to_delete_redo_log_files(): Initialize srv_start_lsn.

1. Moving the following methods from THD to Item_change_list:
   nocheck_register_item_tree_change()
   check_and_register_item_tree_change()
   rollback_item_tree_changes()
   as they work only with the "change_list" member and don't
   require anything else from THD.
2. Deriving THD from Item_change_list

This change will help to fix "MDEV-14603 signal 11 with short stacktrace" easier.

Replace some !rw_lock_own() assertions with the stronger
!btr_search_own_any(). Remove some redundant btr_get_search_latch()
calls.

btr_search_update_hash_ref(): Remove a duplicated assertion.

btr_search_build_page_hash_index(): Remove a duplicated assertion.
rw_lock_s_lock() asserts that the latch is not being held.

btr_search_disable_ref_count(): Remove an assertion. The only caller
is acquiring all adaptive hash index latches.

innodb_init_param(): Initialize btr_ahi_parts=1 for Mariabackup.

btr_search_enabled: Let the adaptive hash index be disabled
in Mariabackup. This would potentially only matter during --export,
and --export performs a table scan, not many index lookups.

innobase_start_or_create_for_mysql(): Only start the data dictionary
and transaction subsystems in normal server startup and during
mariabackup --export.

btr_cur_update_in_place(): Read block->index only once,
so that it cannot change to NULL after the first read.
When block->index != NULL, it must be equal to index.

This way all the libraries are listed in a logical order and it will later
be easier to compare the control file contents to downstream versions.

There was a lot of files generated by the sources that were not installed
in any package. This fixes most of those issues, but not all. Files still
outside any package are in the not-installed, which is used by dh_install.

Also make sure all configuration files are installed to the correct
location that matches layout used by Debian official packages.

All man pages that stem from the actual sources should be installed along
everything else in .install files.

Matching libmariadb-dev package contents in official Debian repositories,
the MariaDB Connector/C files should go into a folder using the 'mariadb'
name. For compatibility with sources that expect to find 'mysql' stuff,
create a symlink from include/mysql to include/mariadb.

Using this package any source that expects MySQL Client libraries should
build without changes. Sources that expect MariaDB Connector C should
build using only libmariadb-dev.

Current MariaDB Connector/C does not build quite identically in
this source tree compared to how it build in Debian official.
The remaining differences are tracked in MDEV-14921.

No path is required. These files always reside in the same directory
anyway.

We already ship libmysqlclient18 that replaces any existing package
with the same name and install symlinks that point to out libmariadb.so.

Instead of creating more packages, we add just one more that does the
same and provides both libmysqlclient19 and -20 replacements symlinks.

This is a clean solution, because systems that only use (and are built
against) libmariadb will not need them and if somebody has done a mistake
in the linker parameters, they will spot it and can easily update the
filename to libmariadb.

- tokuft_logprint should be shipped in the TokuDB plugin
- mysql_install_db is needed for installing/upgrading the datadir
  so it must be in the server core package so users can roll their
  own upgrades

There is no point in having version suffixes in plugin packages. They
will via their control file anyway depend on the correct server or client
version. Having version suffices has no benefit and it just complicates
upgrades when a new major version name package needs to replace all
possible previous version.

This issue was overlooked when merging commit
844a5759 from 10.2 into 10.3, so it
needs to be manually fixed now.

The correct format for new Debian packages is to define install lists
as *.install files. This has already been fixed for 10.3, but it when
merging commit c5a525bd from 10.2, this
change was overlooked and a file of the old format slipped in.