The problem was when a GTID event was part of a group commit, and so contained
a commit id. The code that replaces GTID with a BEGIN event for old slaves did
not correctly handle this case.

Fix the code so that the GTID with commit id can also be properly replaced
with a BEGIN query event. The extra two bytes are in the BEGIN event replaced
with a dummy, empty time zone string.

Older master has no GTID events, so such events are not available for
deciding on scheduling of event groups and so on.

With this patch, we run such events from old masters single-threaded, in the
sql driver thread.

This seems better than trying to make the parallel code handle the data from
older masters; while possible, this would require a lot of testing (as well as
possibly some extra overhead in the scheduling of events), which hardly seems
worthwhile.

With parallel replication, there can be any number of events queued on
in-memory lists in the worker threads.

For normal STOP SLAVE, we want to skip executing any remaining events on those
lists and stop as quickly as possible.

However, for START SLAVE UNTIL, when the UNTIL position is reached in the SQL
driver thread, we must _not_ stop until all already queued events for the
workers have been executed - otherwise we would stop too early, before the
actual UNTIL position had been completely reached.

The code did not handle UNTIL correctly, stopping too early due to not
executing the queued events to completion. Fix this, and also implement that
an explicit STOP SLAVE in the middle (when the SQL driver thread has reached
the UNTIL position but the workers have not) _will_ cause an immediate stop.

Clean up and improve the parallel implementation code, mainly related to
scheduling of work to threads and handling of stop and errors.

Fix a lot of bugs in various corner cases that could lead to crashes or
corruption.

Fix that a single replication domain could easily grab all worker threads and
stall all other domains; now a configuration variable
--slave-domain-parallel-threads allows to limit the number of
workers.

Allow next event group to start as soon as previous group begins the commit
phase (as opposed to when it ends it); this allows multiple event groups on
the slave to participate in group commit, even when no other opportunities for
parallelism are available.

Various fixes:

 - Fix some races in the rpl.rpl_parallel test case.

 - Fix an old incorrect assertion in Log_event iocache read.

 - Fix repeated malloc/free of wait_for_commit and rpl_group_info objects.

 - Simplify wait_for_commit wakeup logic.

 - Fix one case in queue_for_group_commit() where killing one thread would
   fail to correctly signal the error to the next, causing loss of the
   transaction after slave restart.

 - Fix leaking of pthreads (and their allocated stack) due to missing
   PTHREAD_CREATE_DETACHED attribute.

 - Fix how one batch of group-committed transactions wait for the previous
   batch before starting to execute themselves. The old code had a very
   complex scheduling where the first transaction was handled differently,
   with subtle bugs in corner cases. Now each event group is always scheduled
   for a new worker (in a round-robin fashion amongst available workers).
   Keep a count of how many transactions have started to commit, and wait for
   that counter to reach the appropriate value.

 - Fix slave stop to wait for all workers to actually complete processing;
   before, the wait was for update of last_committed_sub_id, which happens a
   bit earlier, and could leave worker threads potentially accessing bits of
   the replication state that is no longer valid after slave stop.

 - Fix a couple of places where the test suite would kill a thread waiting
   inside enter_cond() in connection with debug_sync; debug_sync + kill can
   crash in rare cases due to a race with mysys_var_current_mutex in this
   case.

 - Fix some corner cases where we had enter_cond() but no exit_cond().

 - Fix that we could get failure in wait_for_prior_commit() but forget to flag
   the error with my_error().

 - Fix slave stop (both for normal stop and stop due to error). Now, at stop
   we pick a specific safe point (in terms of event groups executed) and make
   sure that all event groups before that point are executed to completion,
   and that no event group after start executing; this ensures a safe place to
   restart replication, even for non-transactional stuff/DDL. In error stop,
   make sure that all prior event groups are allowed to execute to completion,
   and that any later event groups that have started are rolled back, if
   possible. The old code could leave eg. T1 and T3 committed but T2 not, or
   it could even leave half a transaction not rolled back in some random
   worker, which would cause big problems when that worker was later reused
   after slave restart.

 - Fix the accounting of amount of events queued for one worker. Before, the
   amount was reduced immediately as soon as the events were dequeued (which
   happens all at once); this allowed twice the amount of events to be queued
   in memory for each single worker, which is not what users would expect.

 - Fix that an error set during execution of one event was sometimes not
   cleared before executing the next, causing problems with the error
   reporting.

 - Fix incorrect handling of thd->killed in worker threads.

As a side-effect of purge_relay_logs(), sql_slave_skip_counter
was silently ignored in GTID mode.

But sql_slave_skip_counter in fact is not a good match with GTID.
And it is not really needed either, as users can explicitly set
@@gtid_slave_pos to skip specific GTIDs, in a way that matches
well how GTID replication works.

So with this patch, we give an error on attempts to set
sql_slave_skip_counter when using GTID, with a suggestion to use
gtid_slave_pos instead, if needed.

    MDEV-4984: Implement MASTER_GTID_WAIT() and @@LAST_GTID.
    MDEV-4726: Race in mysql-test/suite/rpl/t/rpl_gtid_stop_start.test
    MDEV-5636: Deadlock in RESET MASTER

Fix a race in test case rpl_gtid_basic.
include/sync_with_master_gtid.inc got confused due to two different
GTIDs both with the same seq_no=9.

The problem is a deadlock between MYSQL_BIN_LOG::reset_logs() and
MYSQL_BIN_LOG::mark_xid_done(). The former takes LOCK_log and waits for the
latter to complete. But the latter also tries to take LOCK_log; this can lead
to a deadlock.

There was already code that tries to deal with this, with the flag
reset_master_pending. However, there was still a small opportunity for
deadlock, when an previous mark_xid_done() is still running when reset_logs()
is called and is at the precise point where it first releases LOCK_xid_list
and then re-aquires both LOCK_log and LOCK_xid_list.

Solve by setting reset_master_pending in reset_logs() before taking
LOCK_log. And also count how many invocations of LOCK_xid_list are in the
progress of releasing and re-aquiring locks, and in reset_logs() wait for that
number to drop to zero after setting reset_master_pending and before taking
LOCK_log.

Rewrite the gtid_waiting::wait_for_gtid() function.
The code was rubbish (and buggy). Now the logic is
much clearer.

Also fix a missing slave sync that could cause test failure.

Couple of small fixes following buildbot testing.

Some GTID test cases were using include/wait_condition.inc with a
condition like SELECT COUNT(*)=4 FROM t1 to wait for the slave to
catch up with the master. This causes races and test failures, as the
changes to the tables become visible at the COMMIT of the SQL thread
(or even before in case of MyISAM), but the changes to
@@gtid_slave_pos only become visible a little bit after the COMMIT.

Now that we have MASTER_GTID_WAIT(), just use that to sync up in a
GTID-friendly way, wrapped in nice include/save_master_gtid.inc and
include/sync_with_master_gtid.inc scripts.

MASTER_GTID_WAIT() is similar to MASTER_POS_WAIT(), but works with a
GTID position rather than an old-style filename/offset.

@@LAST_GTID gives the GTID assigned to the last transaction written
into the binlog.

Together, the two can be used by applications to obtain the GTID of
an update on the master, and then do a MASTER_GTID_WAIT() for that
position on any read slave where it is important to get results that
are caught up with the master at least to the point of the update.

The implementation of MASTER_GTID_WAIT() is implemented in a way
that tries to minimise the performance impact on the SQL threads,
even in the presense of many waiters on single GTID positions (as
from @@LAST_GTID).

 SET(CMAKE_INCLUDE_DIRECTORIES_PROJECT_BEFORE ON)
to find header files from the bundled libraries
(jemalloc, yassl, readline, pcre, etc)
before the ones installed in the system.

The problem was a race between the SQL driver thread and the worker threads.
The SQL driver thread would set rli->last_master_timestamp to zero to
mark that it has caught up with the master, while the worker threads would
set it to the timestamp of the executed event. This can happen out-of-order
in parallel replication, causing the "caught up" status to be overwritten
and Seconds_Behind_Master to wrongly grow when the slave is idle.

To fix, introduce a separate flag rli->sql_thread_caught_up to mark that the
SQL driver thread is caught up. This avoids issues with worker threads
overwriting the SQL driver thread status. In parallel replication, we then
make SHOW SLAVE STATUS check in addition that all worker threads are idle
before showing Seconds_Behind_Master as 0 due to slave idle.

appears to be unset, or when it contains 'root' even though the user
does not have real root permissions

Set of JOIN_TYPE_OUTER made only once to avoid interference with optimization joins which made only once per query.

on server shutdown after SELECT with CONVERT_TZ

It's wrong to return my_empty_string from val_str().
Removing my_empty_string. Using make_empty_result() instead.

            INSTALL PLUGIN

There was mixed lock order between LOCK_plugin, LOCK_global_system_variables
and LOCK_system_variables_hash. This patch ensures that write-lock on
LOCK_system_variables_hash doesn't intersect with LOCK_plugin.

Fixed by moving initialization/deinitialization of plugin options from
plugin_add()/plugin_del() to plugin_initialize()/plugin_deinitalize().
So that plugin options are handled without protection of LOCK_plugin.

add patch as a "suggested" dependency to mariadb-test-5.5

if two threads were calling mi_assign_to_key_cache() for the same table,
one could change share->key_cache while the other was having
share->key_cache->op_lock locked. The other thread would crash then,
trying to unlock share->key_cache->op_lock (because it would be a different mutex).

fixed by caching the value of share->key_cache in a local variable. The thread can still
call flush_key_blocks() for an unassigned keycache, but it's harmless.

Assertion `mon > 0 && mon < 13' failed.'

in some VMs readline thinks that the window size is zero. ignore it.

MDEV-5461 Assertion `length <= column->length' fails in write_block_record with functions in select list, GROUP BY, ORDER BY

Old code in create_tmp_table(), that created an extra one-byte field (recinfo)
before every NULL-able grouping field (Field) in the tmp table, did not actually work.
Because the matching code in end_update(), that was supposed to update this byte,
was using a wrong offset, updating the first byte of the Field, not a byte before it.
Normally this wasn't an issue, because the Field value (written later in end_update)
was overwriting this byte anyway. But in this bug the Field was Field_null, with zero
length, so end_update() was overwriting the first byte of the following field.
And the following field was not-nullable constant, which was stored only once in
create_tmp_table and never updated later.

Fixed by removing the code that didn't do any useful work anyway.