Updated mysqld--help,win.rdiff.

Some compilers are unhappy about mixed struct/class declarations.

Fixed type cast.

Fixed sysvars_server_[not]embedded failure: changed type of
table_open_cache_instances from ulong to uint.

Added casts foratomic operations around tc_active_instances and
tc_contention_warning_reported: needed on some platforms.

Table cache instances autosizing.

Improve scalability by implementing multi-instance table cache.

The condition pushed into WHERE/HAVING of a materialized
view/derived table may differ for different executions of
the same prepared statement. That's why the should be
ANDed with the existing WHERE/HAVING conditions only after all
permanent transformations of these conditions has been
performed.

Do not push conditions into materialized views/derived tables
marked with the flag 'fill_me'.

MDEV-10779 Failing assertion lex->proc_list.elements == 0 or syntax error on PROCEDURE ANALYSE in UNION

Reverting a part of the patch for "MDEV-8909union parser cleanup",
as a parenthesized SELECT with PROCEDURE followed by UNION is not
disallowed by the grammar (only a non-parenthesized SELECT with PROCEDURE
followed by a UNION is disallowed grammatically).

When running with --ps protocol, window functions cleanup method may
be called twice. Enforce proper cleanup by setting pointers to NULL
after deletion. This prevents double deletion on second cleanup.

no need to set net->vio to zero in mysql_real_connect

        Wrong argument for the my_error function fixed.

This bug in the code of Item_ref::build_clone could
cause corruption of items in where conditions.
Also made sure that equality predicates extracted
from multiple equality items to be pushed into
materialized views were cloned.

If there are other threads running (for example binlog background
thread), then the thread count may not drop to zero at the end of
do_handle_bootstrap(). This caused an assertion and missing wakeup of
the main thread. The missing wakeup is because THD::~THD() only
signals the COND_thread_count mutex when the number of threads drops
to zero.
Signed-off-by: Kristian Nielsen <knielsen@knielsen-hq.org>

This can happen with degenerate frame definitions that are always empty.

Cursors now report their current row number as the boundary of the
partition. This is used by Frame_scan_cursor to compute aggregate
functions that do not support removal.

Added an extra virtual method to the Frame_cursor class to allow cursors
to report the row number to which they are pointing.

Currently cursors automatically add values to the sum functions they
manage. There are use cases when we just want to figure out the frame
boundaries, without actually adding/removing values from them.

The 'IS A' relation is more appropriate for Partition_read_cursor. This
also helps with accessing methods available only to Table_read_cursor.

When specifying a RANGE type frame that exceeds the partition size, both
for the top and bottom cursors we end up removing more rows than added
to the aggregate function. This happens because our TOP range cursor,
which removes values from the aggregate function, would be allowed to breach
partition boundaries, while the BOTTOM range cursor would not.

To prevent this from happening, force the TOP range cursor to only move
within the current partition, as does the BOTTOM range cursor.

Perform only one table scan for each window function present. We do this
by keeping keeping cursors for each window function frame bound and
running them for each function for every row.

MDEV-10763: Wrong result - server does not return NULL values from default list partition after ALTER table

Fixed partition pruning for NULL value.

MDEV-10765: Wrong result - query does not retrieve values from default partition on a table partitioned by list columns

Partial matches should be treat as not exact one.

because it is zero always in this place as we have check for this above
which checks that net->vio isn't 0 and exit.

Conflicts:
	sql/mysqld.cc
	sql/slave.cc

When a deadlock kill is detected inside the storage engine, the kill
is not done immediately, to avoid calling back into the storage engine
kill_query method with various lock subsystem mutexes held. Instead the
kill is queued and done later by a slave background thread.

This patch in preparation for fixing TokuDB optimistic parallel
replication, as well as for removing locking hacks in InnoDB/XtraDB in
10.2.

Signed-off-by: Kristian Nielsen <knielsen at knielsen-hq.org>