Amend commit b853b4fd
that was reverted in commit 29150e23.

recv_parse_log_recs(): Do check for corrupted redo log or file
system before checking for len==0, but only read *ptr if
it is not past the end of the buffer (end_ptr).

recv_parse_log_rec(): Report incorrect redo log type
in a consistent way with recv_parse_or_apply_log_rec_body().
This is a follow-up to commit f30c5af4.

The Pool poisoning that was introduced in MDEV-15030 introduced
race conditions in AddressSanitizer builds, because concurrent
poisoning and unpoisoning were not prevented by any synchronization
primitive.

Pool::get(): Protect the unpoisoning by m_lock_strategy.

Pool::mem_free(): Protect the poisoning by m_lock_strategy.

Pool::putl(): Renamed from put(), because now the caller is
responsible for invoking m_lock_strategy.

If trx_free() and trx_create_low() were called while a call to
trx_reference() was pending, we could get a reference to a wrong
transaction object.

trx_reference(): Return NULL if the trx->id no longer matches.

lock_trx_release_locks(): Assign trx->id = 0, so that trx_reference()
will not return a reference to this object.

trx_cleanup_at_db_startup(): Assign trx->id = 0.

assert_trx_is_free(): Assert !trx->n_ref. Assert trx->id == 0,
now that it will be cleared as part of a transaction commit.

Also, replace reverse iteration with forward iteration.

lock_table_has(): Remove a redundant condition.

Allocate trx->lock.rec_pool and trx->lock.table_pool directly from trx_t.
Remove unnecessary use of std::vector.

In order to do this, move some definitions from lock0priv.h to
lock0types.h, so that ib_lock_t will not be an opaque type.

row_log_table_get_pk_col(): Replace a condition that was inadvertently
removed in MDEV-16365. PRIMARY KEY columns are never allowed to be NULL,
and failure to enforce the constraint caused a null pointer to be
dereferenced in mem_heap_dup().

aws_key_management needs current directory to be datadir during
initalization, it scans current directory for encrypted keys.

Fix is to ensure, that plugin initialization in mariabackup happens
after the call to my_setwd(mysql_real_data_home).

The problem occurs because the statement generated by Spider used an
internal function name, ADD_TIME.

This problem has been corrected by the fix for bug MDEV-16878 within the
server, which enables Spider to generate the statement using the actual
SQL function name.  I have made some additional changes within Spider to fix
related problems that I observed while testing.

Author:
  Jacob Mathew.

First Reviewer:
  Alexander Barkov.

Second Reviewer:
  Kentoku Shiba.

Merged:
  Commit 4b6dccc8 on branch bb-10.3-MDEV-16398

Fix a race condition in the test.

The test causes simulated server crashes with DBUG_SUICIDE();.

It also relies on transactions that were committed right before the
crash to be visible after the crash (that is, it requires durability).

Run the test with transaction durability enabled: set
rocksdb-flush-log-at-trx-commit=1.

only run git if .git exists, don't do it for source tarballs

we stopped using -fno-implicit-templates many years ago

applied (at least partially):
* armhf_mroonga_storage_fail.patch (unaligned write)
* mysqld_multi.server_lsb-header.patch (add LSB header)
* fix-spelling-errors.patch (tabxml.cpp)
* hurd_socket.patch (in Platform.pm)
* remove-systemd-obsolete-target.patch

don't use MY_MUTEX_INIT_FAST in constructors of statically
allocated objects.

The package libmariadbclient18 contains the dialog.so plugin, which also
the new libmariadb3 ships. As they both use the exact same path the latter
must be marked as a with Breaks and Replaces relations ship.

Note: This fix is conservative hack for stable releases 10.2 and 10.3.
In 10.4, the development release at the time, we will clean up how the
libmariadb3 packaging and it's -compat packages are done to match that
what is done in downstream Debian official.

recv_parse_log_recs(): Do not check for corruption before
checking for end-of-log-buffer. For some reason, adding the
check to the logical-looking place would cause intermittent
recovery failures in the tests innodb.innodb-index and
innodb_gis.rtree_compress2.

recv_parse_log_recs(): Check for corruption before checking for
end-of-log-buffer.

mlog_parse_initial_log_record(), page_cur_parse_delete_rec():
Flag corruption for out-of-bounds values, and let the caller
dump the corrupted redo log extract.

If recv_sys_justify_left_parsing_buf() has been invoked, it is possible
that recv_previous_parsed_rec_offset is after the current offset.
In this case, we must not dump any bytes before the current record.

If the LOG_BLOCK_HDR_DATA_LEN field is corrupted, scanning the
log records could fail in strange ways. It is better to validate
the field as part of validating each log block.

Display the log record type in hexadecimal, not binary.

The package libmariadbclient18 contains the dialog.so plugin, which also
the new libmariadb3 ships. As they both use the exact same path the latter
must be marked as a with Breaks and Replaces relations ship.

Note: This fix is conservative hack for stable releases 10.2 and 10.3.
In 10.4, the development release at the time, we will clean up how the
libmariadb3 packaging and it's -compat packages are done to match that
what is done in downstream Debian official.