Use post_kill_notification in for one_thread_per_connection scheduler, 
the same as already used in threadpool, to reliably wake a thread  stuck in 
read() or in different poll() variations.

 - Ensure asm parameters are in registers, so we do not de-reference from
   bogus stack pointer.

 - Make return address undefined in DWARF unwind info in my_context_spawn,
   so DWARF-based unwinders will know this is the end of the call stack
   (same as the amd64 fix for the similar issue).

Disable compiling unit tests if WITH_UNIT_TEST is FALSE.

Also, fix CMake code to allow compilation WITHOUT_ARIA_STORAGE_ENGINE

and when safemalloc is enabled (use ucontext, otherwise backtrace function gets confused and crashes)

Changed implementation os_file_rename() on Windows  such as it does not fail if destination file already exists. Now MoveFileEx() with MOVEFILE_REPLACE_EXISTING flag is used, instead of prior MoveFile().
This fixed implementation is better compatible with rename() on POSIX systems.

because it assumes the function prologue that gcc does not
generate.

(and collateral changes)

mysql-test/t/innodb_mysql_lock.test:
  change the variable from the test, not from the .opt file.
  one mysqld restart less.
mysql-test/t/lowercase_table4.test:
  fix dos line endings
sql/handler.cc:
  don't access the uninitialized variable

Ensure semicolons are not lost when concatenating fill_help_tables to  bootstrap.sql in Windows.
5.6 bootstrapper requires semicolons to separate bootstrapper commands.

  always use intptr type when casting a pointer to an integer to avoid sign expansion.
  or, at least, cast identically in socket_summary_by_instance and socket_instances

The following piece of code in trnman.c:

345   trn->min_read_from=  active_list_min.next->trid;
  ...
351   active_list_max.prev= trn->prev->next= trn;
352   trid_min_read_from= active_list_min.next->min_read_from;

on 345 gcc stores active_list_min.next in %ebx
(and trn->min_read_from=[%ebx]->trid)
and on 352 it does trid_min_read_from= [%ebx]->min_read_from;
BUT active_list_min.next was changed on the line 351.
gcc doesn't notice it and continues to use the cached value.

Bug#13510739 63775: SERVER CRASH ON HANDLER READ NEXT AFTER DELETE RECORD.

  shared should provide libmysqlclient.so.18(libmysqlclient_16) too
  don't "use DBD::mysql" explicitly in mytop

This patch undoes the removal of enum store_key_result by the previous patch for mdev-3812.

(because of a width or a short buffer)

MDEV-3812: Remove unneeded extra call to engine->exec() in Item_subselect::exec, remove enum store_key_result

This task fixes an ineffeciency that is a remainder from MySQL 5.0/5.1. There, subqueries
were optimized in a lazy manner, when executed for the first time. During this lazy optimization
it may happen that the server finds a more efficient subquery engine, and substitute the current
engine of the query being executed with the new engine. This required re-execution of the engine.

MariaDB 5.3 pre-optimizes subqueries in almost all cases, and the engine is chosen in most cases,
except when subquery materialization found that it must use partial matching. In this case, the
current code was performing one extra re-execution although it was not needed at all. The patch
performs the re-execution only if the engine was changed while executing.

In addition the patch performs small cleanup by removing "enum store_key_result" because it is
essentially a boolean, and the code that uses it already maps it to a boolean.

Adjusted test case results after the merge 10.0-base, 10.0-monty.
The results are in sync with MySQL 5.6.7.

fix rpl_mdev382.result

do not print return address when callstack is output on Windows, it does not provide any useful info

Do not DBUG_PRINT uninitialized variable. This avoid false positive from runtime checks in debug builds (Windows).

  shared should provide libmysqlclient.so.18(libmysqlclient_16) too
  don't "use DBD::mysql" explicitly in mytop

don't try to define it for plugins, then, as they don't need it.

Fix by Sergey Petrunia.

This patch only prevents the evaluation of expensive subqueries during optimization.
The crash reported in this bug has been fixed by some other patch.
The fix is to call value->is_null() only when  !value->is_expensive(), because is_null()
may trigger evaluation of the Item, which in turn triggers subquery evaluation if the
Item is a subquery.