Fixed archive.archive failure.

Applied remnants of two revisions, which were partially merged.

Rev. 3225.1.1 (5.0 compatibility):

  BUG#11756687 - 48633: ARCHIVE TABLES ARE NOT UPGRADEABLE

  Archive table created by 5.0 were not accessible.

  This patch adds various fixes so that 5.0 archive tables
  are readable and writable. Though it is strongly recommended
  to avoid binary upgrade of archive tables whenever it is
  possible.

Rev. 3710 (due to valgrind warnings):

  Bug#13907676: HA_ARCHIVE::INFO
  In WL#4305 the refactoring of the archive writer,
  it could flush the writer when it was not yet open.

  This was due to if bulk insert was used but no
  rows was actually inserted (write_row was never called),
  the writer was marked dirty even if it was not open.

  Fix was to only mark it as dirty if it was opened.

mysql-test/std_data/bug48633.ARM:
  A test case for BUG#11756687: archive table created by 5.0.95.
mysql-test/std_data/bug48633.ARZ:
  A test case for BUG#11756687: archive table created by 5.0.95.
mysql-test/std_data/bug48633.frm:
  A test case for BUG#11756687: archive table created by 5.0.95.
mysql-test/suite/archive/archive.result:
  Modified a test case for BUG#47012 according to fix for
  BUG#11756687.
  Added a test case for BUG#11756687.
mysql-test/suite/archive/archive.test:
  Modified a test case for BUG#47012 according to fix for
  BUG#11756687.
  Added a test case for BUG#11756687.
  
  No need to remove .ARM files anymore: DROP TABLE will take
  care of them.
storage/archive/azio.c:
  Do not write AZIO (v.3) header to GZIO file (v.1).
  
  Added initialization of various azio_stream members
  to read_header() so it can proceed with v.1 format.
  
  Update data start position only when reading first
  GZIO header. That is only on azopen(), but never on
  azread().
storage/archive/ha_archive.cc:
  Removed guardians that were rejecting to open v.1 archive
  tables.
  
  Reload .frm when repairing v.1 tables - they didn't have
  storage for .frm.
  
  Do not flush write stream when it is not open.
  
  Let DROP TABLE remove 5.0 .ARM files.

Fixed debian/ubuntu build failure.

cmake/ssl.cmake:
  Do not prefer static SSL libraries for WITH_SSL=system|yes as there is
  no guarantee that they will link well with MariaDB shared objects.
  Specifically on debian/ubuntu static SSL libraries are built without
  -fPIC.
  
  Restore 5.6 behavior. 10.0.3 never prefer static SSL libraries.

Fixed failure of one of the test case from innodb_ext_key.test.
The fact is that the innodb code may return statistical data
on record per key values that is far off the real numbers. 
This is exactly what happened in mariadb-5.5 with this test case.

Added an ANALYZE command in this test case to avoid this problem.
The same change will be done in mariadb-5.5

Fixed connect.grant failure.

sql/create_options.cc:
  Keep "first" list intact, allocate new list for merge result.
  
  Normally "first" is options list on TABLE_SHARE. ALTER TABLE may fail
  after the merge and leave share with corrupt list in the table definition
  cache.
sql/create_options.h:
  Construct engine_option_value from another engine_option_value.

Fixed maria.maria and funcs_1.is_tables failure.

sql/sql_table.cc:
  C_ALTER_TABLE_FRM_ONLY lost it's meaning after merge of inplace alter:
  now it is used even if table is to be copied.
  Otherwise: how can row format change go without notifying storage engine?
  Removed check for create_table_mode for now, to be discussed.

Fixed query_response_time build failure.

plugin/query_response_time/query_response_time.cc:
  my_atomic_rwlock_rdlock() may update time_collector_lock structure.
  Remove const modifier from relevant methods.

Fixed debian/ubuntu packages build failure.

debian/patches/33_scripts__mysql_create_system_tables__no_test.dpatch:
  tmp_user has now extra column, updated patch accordingly.

Fixed main.innodb_mysql_sync failure.

TDC_RT_REMOVE_NOT_OWN_KEEP_SHARE is not same as
TDC_RT_REMOVE_NOT_OWN_AND_MARK_NOT_USABLE. Comment in sql_table.cc says:

      Storage engine has requested exclusive lock only for prepare phase
      and we are not under LOCK TABLES.
      Don't mark TABLE_SHARE as old in this case, as this won't allow opening
      of table by other threads during main phase of in-place ALTER TABLE.

At this moment we hold exclusive metadata lock, all we should do
is purge unused TABLE objects.

sql/sql_base.cc:
  Restore 5.6 behavior of TDC_RT_REMOVE_NOT_OWN_KEEP_SHARE.
sql/sql_base.h:
  Restore 5.6 behavior of TDC_RT_REMOVE_NOT_OWN_KEEP_SHARE.

Fixed main.mysql--help failure.

mysql-test/r/mysqld--help.result:
  PFS_MAX_RWLOCK_CLASS was raised to 40 during the merge.

Fixed sql_discovery.simple failure.

storage/sequence/sequence.cc:
  Let sequence share class have unique name to avoid collision with other
  classes.
storage/test_sql_discovery/test_sql_discovery.cc:
  Let test_sql_discovery share class have unique name to avoid collision with
  other classes.

Fixed innodb.help_url failure.

mysql-test/suite/innodb/r/help_url.result:
  Accessing discarded tablespace doesn't cause loud error message anymore.
  Use different way to retrieve refman link.
mysql-test/suite/innodb/t/help_url.test:
  Accessing discarded tablespace doesn't cause loud error message anymore.
  Use different way to retrieve refman link.

Fixed connect compilation failure (with ODBC).

storage/connect/odbconn.cpp:
  min/max are not defined by my_global.h anymore, include osutil.h instead.

- Make THD::raise_condition() call push_warning() after set_error_status() call.
  (they seem to have accidentally exchanged in this merge cset:
   sergii@pisem.net-20130721143919-7cltcw2l9g29f983)
- Rollback the patch from two csets before (the one with comment:
   Update filesort_debug.test (see comment #1 in MDEV-4786 for analysis))

- As of 10.0.2, MariaDB has mysql.gtid_slave_pos table as system table.

Fixed connect compilation failure.

storage/connect/csort.cpp:
  min/max are not defined by my_global.h anymore, include osutil.h instead.
storage/connect/filamfix.cpp:
  min/max are not defined by my_global.h anymore, include osutil.h instead.
storage/connect/ha_connect.cc:
  Renamed MYSQL_ERROR -> Sql_condition.
  Renamed KEY::key_parts -> KEY::user_defined_key_parts.
  Adjusted connect according to:
  WL#4305 - storage-engine private data area per physical table.
storage/connect/ha_connect.h:
  Adjusted connect according to:
  WL#4305 - storage-engine private data area per physical table.
storage/connect/inihandl.c:
  min/max are not defined by my_global.h anymore, include osutil.h instead.
storage/connect/osutil.h:
  min/max are not defined by my_global.h anymore, define them for C as well.
storage/connect/taboccur.cpp:
  Include server headers before osutil.h to avoid min/max redefinition.
storage/connect/tabpivot.cpp:
  Include server headers before osutil.h to avoid min/max redefinition.
storage/connect/tabtbl.cpp:
  Include server headers before osutil.h to avoid min/max redefinition.
storage/connect/tabxcl.cpp:
  Include server headers before osutil.h to avoid min/max redefinition.
storage/connect/value.cpp:
  min/max are not defined by my_global.h anymore, include osutil.h instead.
  Include server headers before osutil.h to avoid min/max redefinition.

Fixed test_sql_discovery compilation failure.

storage/test_sql_discovery/mysql-test/archive/discover.rdiff:
  archive.discover result has changed during merge. Adjusted discover.rdiff
  accordingly.
storage/test_sql_discovery/test_sql_discovery.cc:
  Adjusted test_sql_discovery according to:
  WL#4305 - storage-engine private data area per physical table.

Fixed sequence compilation failure.

storage/sequence/sequence.cc:
  Adjusted sequence according to:
  WL#4305 - storage-engine private data area per physical table.

sql/sql_table.cc:
  Removed false assertion that HA_NO_COPY_ON_ALTER is merge specific.
  This flag is used at least by Spider and Connect.
storage/spider/ha_spider.cc:
  Renamed THD::stmt_da -> THD::get_stmt_da().
  Renamed KEY::key_parts -> KEY::user_defined_key_parts.
  Renamed ALTER_* flags -> Alter_info::ALTER_*.
storage/spider/spd_conn.cc:
  Renamed THD::stmt_da -> THD::get_stmt_da().
storage/spider/spd_db_conn.cc:
  Renamed MYSQL_ERROR -> Sql_condition.
  Renamed KEY::key_parts -> KEY::user_defined_key_parts.
storage/spider/spd_db_include.h:
  current_comment_start argument of generate_parition_syntax()
  was removed during merge.
storage/spider/spd_db_mysql.cc:
  Renamed KEY::key_parts -> KEY::user_defined_key_parts.
storage/spider/spd_ping_table.cc:
  Renamed THD::stmt_da -> THD::get_stmt_da().
storage/spider/spd_table.cc:
  Include my_getopt.h so that it exports my_defaults_file and
  my_defaults_extra_file.
  Renamed KEY::key_parts -> KEY::user_defined_key_parts.

* update results
* don't force HA_CREATE_DELAY_KEY_WRITE on all temp tables,
  (bad for CREATE ... LIKE) instead imply it in myisam/aria
* restore HA_ERR_TABLE_DEF_CHANGED in archive
* increase the default number of rwlock classes in P_S to fit all our rwlocks

includes:
* remove some remnants of "Bug#14521864: MYSQL 5.1 TO 5.5 BUGS PARTITIONING"
* introduce LOCK_share, now LOCK_ha_data is strictly for engines
* rea_create_table() always creates .par file (even in "frm-only" mode)
* fix a 5.6 bug, temp file leak on dummy ALTER TABLE

test suite added to be sure that bug is fixed

MDEV-4647: Valgrind warnings (Conditional jump or move depends on uninitialised value) in Item_equal::fix_fields
Fix to calm down valgrind.

(fix of http://bugs.mysql.com/bug.php?id=69623)

  
In the case of loose scan used as input for order by, end_send()
didn't detect correctly that a loose scan was used, and didn't copy
the non-aggregated fields from the temp table used for ORDER BY.
  
The fix uses the fact that the quick select used for sorting is
attached to JOIN::pre_sort_join_tab instead of JOIN::join_tab.

MDEV-4570 [PATCH] Sys_query_cache_limit initialization depends on initialization in other source files

 REFRESH_FAST do not fit in 32 bit