make CMakeLists.txt to detect if the installed boost can be compiled with the
installed compile and specified set of compiler options.

Background: even sufficiently new Boost cannot be compiled with the sufficiently old gcc
in the presence of -fno-rtti

add missing checks to configure.cmake
remove dead code and unused HAVE_xxx constants from the sources

Don't use ER(xxx) in THD::close_connection(), when current_thd is already reset to NULL.
Prefer ER_THD() or ER_DEFAULT() instead.

      mysql_rm_table_no_locks() function was modified.
      When we construct log record for the DROP TABLE, now we
      look if there's a comment before the first table name and
      add it to the record if so.
  
per-file comments:
  sql/sql_table.cc
  MDEV-340 Save replication comments for DROP TABLE.
        comment_length() function implemented to find comments in the query,
        call it in mysql_rm_table_no_locks() and use the result to form log record.
  mysql-test/suite/binlog/r/binlog_drop_if_exists.result
  MDEV-340 Save replication comments for DROP TABLE.
        test result updated.
  mysql-test/suite/binlog/t/binlog_drop_if_exists.test
  MDEV-340 Save replication comments for DROP TABLE.
        test case added.

Add the check for sys/utsname.h to configure.cmake

use correct macro for Microsoft compiler. It is _MSC_VER , not _MSV_VER

- Make Item_subselect::fix_fields() ignore UNCACHEABLE_EXPLAIN flag when deciding whether 
  the subquery item should be marked as constant.

Fix : use attribute %config(noreplace)  for /etc/my.cnf , instead of (automatically generated) %config

see https://mariadb.atlassian.net/browse/MDEV-11

- index_merge/intersection is unable to work on GIS indexes, because:
  1. index scans have no Rowid-Ordered-Retrieval property
  2. When one does an index-only read over a GIS index, they do not 
     get the index tuple, because index only contains bounding box of the geometry.
     This is why key_copy() call crashed.
This patch fixes #1, which makes the problem go away. Theoretically, it would 
be nice to check #2, too, but SE API semantics is not sufficiently precise to do it.

Now partition engine adds underlying tables to the QC and ask underlying tables engine permittion to cache the query and return result of the query.

Incorrect QC cleanup in case of table registration failure fixe.

Unified interface for myisammrg & partitioned engnes for QC.

MDEV-393. Remove --loose-pbxt=OFF/loose-skip-pbxt from bootstrapper calls to avoid "unknown parameter" warning

fix incorrect merge

Fixed MDEV-385: mysqltest running with continue-on-error crashes on a non-SQL command producing an error 

client/mysqltest.cc:
  Added missing DBUG_VOID_RETURN

Thanks to Ivoz for pointing this out.

- Let QUICK_RANGE_SELECT::init_ror_merged_scan() call  quick->reset() only 
  after we've set the column read bitmaps. 

- Better error messages

This fixes that one again can run the test systems with many threads without having to increase fs.aio-max-nr.


mysql-test/include/mtr_check.sql:
  Ignore the INNODB_USE_NATIVE_AIO variable (may change during execution)
mysql-test/mysql-test-run.pl:
  Ignore warnings for failure to setup AIO
storage/innobase/os/os0file.c:
  Continue without AIO even if we can't allocate resources for AIO
storage/xtradb/os/os0file.c:
  Continue without AIO even if we can't allocate resources for AIO
storage/xtradb/srv/srv0start.c:
  Give an error message (instead of core dump) if AIO can't be initialized

TABLE_LIST::check_single_table made aware about fact that now if table attached to a merged view it can be (unopened) temporary table
(in 5.2 it was always leaf table or non (in case of several tables)).

The bug was the result of the incomplete fix for bug lp bug 1008293.

the new file is fully synced to disk and binlog index. This fixes a window
where a crash would leave next server restart unable to detect that a crash
occured, causing recovery to fail.

MDEV-359: Fix another case where switch-off semisync could cause a race that ended with server crash.

This one was when the code releases and reaquires the lock with pthread_cond_wait() - and semisync is switched off meanwhile.

The semisync code does a fast-but-unsafe check for enabled or not without lock,
followed by a slow-but-safe check under lock. However, if the slow check failed,
the code still referenced not valid data (in an assert() expression), causing a
crash.

Fixed by not running the incorrect assert when semisync is disabled.