The logic and the implementation scheme are similar with the
MDEV-9197 Pushdown conditions into non-mergeable views/derived tables

How the push down is made on the example:

select * from t1
where a>3 and b>10 and
 (a,b) in (select x,max(y) from t2 group by x);

-->

select * from t1
where a>3 and b>10 and
  (a,b) in (select x,max(y)
            from t2
            where x>3
            group by x
            having max(y)>10);

The implementation scheme:

1. Search for the condition cond that depends only on the fields
   from the left part of the IN subquery (left_part)
2. Find fields F_group in the select of the right part of the
   IN subquery (right_part) that are used in the GROUP BY
3. Extract from the cond condition cond_where that depends only on the
   fields from the left_part that stay at the same places in the left_part
   (have the same indexes) as the F_group fields in the projection of the
   right_part
4. Transform cond_where so it can be pushed into the WHERE clause of the
   right_part and delete cond_where from the cond
5. Transform cond so it can be pushed into the HAVING clause of the right_part

The optimization is made in the
Item_in_subselect::pushdown_cond_for_in_subquery() and is controlled by the
variable condition_pushdown_for_subquery.

New test file in_subq_cond_pushdown.test is created.

There are also some changes made for setup_jtbm_semi_joins().
Now it is decomposed into the 2 procedures: setup_degenerate_jtbm_semi_joins()
that is called before optimize_cond() for cond and setup_jtbm_semi_joins()
that is called after optimize_cond().
New setup_jtbm_semi_joins() is made in the way so that the result of its work is
the same as if it was called before optimize_cond().

The code that is common for pushdown into materialized derived and into materialized
IN subqueries is factored out into pushdown_cond_for_derived(),
Item_in_subselect::pushdown_cond_for_in_subquery() and
st_select_lex::pushdown_cond_into_where_clause().

Less ifdefs, disable main.connect-abstract until libmariadb is updated.

The functionality of the socket system variable is extended
here such that a preciding '@' indicates that the socket
will be an abstract socket. Thie socket name wil be
the remainder of the name after the '@'. This is consistent
with the approached used by systemd in socket activation.

Thanks to Sergey Vojtovich:

On OS X sockaddr_un is defined as:

struct sockaddr_un
{
  u_char sun_len;
  u_char sun_family;
  char  sun_path[104];
};

There is a comment in man 7 unix (on linux):

"
On Linux, the above offsetof() expression equates to the same value as sizeof(sa_family_t),
but some other implementations include other fields before sun_path, so the offsetof()
expression more portably describes the size of the address structure.
"

As such, use the offsetof for Linux and use the previous sizeof(UNIXaddr)
for non-unix platforms as that's what worked before and they don't
support abstract sockets so there's no compatibility problem..

strace -fe trace=networking mysqld --skip-networking --socket @abc ...
...
[pid 10578] socket(AF_UNIX, SOCK_STREAM|SOCK_CLOEXEC, 0) = 22
[pid 10578] setsockopt(22, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
[pid 10578] bind(22, {sa_family=AF_UNIX, sun_path=@"abc"}, 6) = 0
[pid 10578] listen(22, 80)              = 0
...
Version: '10.3.6-MariaDB-log'  socket: '@abc'  port: 0  Source distribution

$ lsof -p 10578
mysqld  10578  dan   22u  unix 0x00000000087e688c       0t0 4787815 @abc type=STREAM

Recent changes in semisync initialization via MDEV-13073 introduced
instantiation of THD too early from the server components
pov which led to segfault.

Fixed with relocating the semisync component initialization
to later time when thread specific memory can be used.

Fixed typo on select that is executed only when something unexpected
happens.

MDEV-12632: Source and destination overlap in memcpy, encryption.innodb-discard-import-change fails in buildbot with valgrind

Use block->page.offset for checking page number.

Problem was that we did not delete explain information
when Galera must replay a query.

Could not find easily repeatable test case that would not
cause other problems.

followup for 339b9055

don't use $opt_gdb for --client-gdb

mark CONNECTION_ID() as returning *unsigned* number

fix versioning.partition failures:
* Make the test stable by setting timestamp to fixed values.

The error occurs because of how the character set and collation are chosen for
stored procedure parameters that have a character data type.  If the character
set and collation are not explicitly stated in the declaration, the server
chooses the database character set and collation in effect at routine creation
time.

To fix the problem, I added explicit character set and collation attributes
for the stored procedure parameters in the install_spider.sql script.

Author:
  Jacob Mathew.

Reviewer:
  Kentoku Shiba.

Merged From:
  bb-10.3-MDEV-15692

It was on for alpha stages of 10.3, and now we're past that.

MDEV-12632: Source and destination overlap in memcpy, encryption.innodb-discard-import-change fails in buildbot with valgrind

Problem was that if tablespace was encrypted we try to copy
also page 0 from read buffer to write buffer that are in
that case the same memory area.

fil_iterate
	When tablespace is encrypted or compressed its
        first page (i.e. page 0) is not encrypted or
	compressed and there is no need to copy buffer.

MDEV-12632: Source and destination overlap in memcpy, encryption.innodb-discard-import-change fails in buildbot with valgrind

Problem was that if tablespace was encrypted we try to copy
also page 0 from read buffer to write buffer that are in
that case the same memory area.

fil_iterate
	When tablespace is encrypted or compressed its
        first page (i.e. page 0) is not encrypted or
	compressed and there is no need to copy buffer.

Wait until rotation has ended and shutdown before grep to make sure
that dirty pages are on datafiles.

If we use hostnames, it will fail during name resolution on UNIX systems

Also fix perfschema.hostcache_ipv6_max_con.
The test case makes use of a debug switch to execute some special code.
The code does hostname replacement. Every hostname sent during connect
phase becomes santa.claus.ipv4.example.com (or ipv6). This causes a
connection from root@localhost to fail, as root is not registered as
santa claus user. The failure is only apparent on Windows as Unix
systems make use of sockets, which bypass the name resolution check
entirely.

Fix 10.2-specific bug - copy-back is not prepared to handle system
tablespaces with absolute path.

Wait until rotation has ended and shutdown before grep to make sure
that dirty pages are on datafiles.

row_undo_step(): If innodb_fast_shutdown=3 has been requested,
abort the rollback of any non-DDL transactions. Starting with
MDEV-12323, we aborted the rollback of recovered transactions. The
transactions would be rolled back on subsequent server startup.

trx_roll_report_progress(): Renamed from trx_roll_must_shutdown(),
now that the shutdown check has been moved to the only caller.

trx_commit_low(): Allow mtr=NULL for transactions that are aborted
on rollback.

trx_rollback_finish(): Clean up aborted transactions to avoid
assertion failures and memory leaks on shutdown. This code was
previously in trx_rollback_active().

trx_rollback_to_savepoint_low(), trx_rollback_for_mysql_low():
Remove some redundant assertions.

The implementations of the convert_to_basic_const_item() virtual
function for the Item_cache classes should call cache_value() when
value_cached == NULL.

Also, shorten the test so it usually runs 2 seconds, not 3.

There was a missing argument to service_manager_extend_timeout call
and the signness of arguments did not match.

numerous fixes for CREATE ... SELECT with system versioning:
In CREATE ... SELECT the table is created based on the result set,
field properties do not count. That is
* field invisibility is *not* copied over
* AS ROW START/END is *not* copied over
* the history is *not* copied over
* system row_start/row_end fields can *not* be created from the SELECT part

[closes tempesta-tech#472]

MDEV-15391 Server crashes in JOIN::fix_all_splittings_in_plan or Assertion `join->best_read < double(1.79...e+308L)' failed

vers_setup_conds() used to AND all conditions on row_start/row_end
columns and store it either in the WHERE clause or in the ON
clause for some table. In some cases this caused ON clause
to have conditions for tables that aren't part of that ON's join.

Fixed to put a table's condition always in the ON clause of the
corresponding table.

Removed unnecessary ... `OR row_end IS NULL` clause, it's not needed
in the ON clause.

Simplified handling on PS and SP.

fix parsing of the AS OF clause

get_schema_column_record(): print 'WITHOUT SYSTEM VERSIONING` in 'EXTRA'
for such fields

--gdb now accepts an argument, it will be passed to gdb as a command.
multiple commands can be separated by a (non-standard and not escapable)
delimiter - semicolon (;).

Old usage with a bare --gdb continues to work too, of course.

disallow system versioning for tables in mysql database

use named methods instead.

the function xxx_eq(a,b) returns true if two elements
are equal and false if they are not.

strmake_buf() macro should only be used with char[] arrays,
and never with char* pointers. To distinguish between the two
we create a new variable of the same type and initialize it
using array initializer, this causes compilation failure
with pointers. The variable is unused and will be removed by the
compiler. It's enough to do this check only with gcc, so
it doesn't have to be portable.