MDEV-4820: Empty master does not give error for slave GTID position that does not exist in the binlog

The main bug here was the following situation:

Suppose we set up a completely new master2 as an extra multi-master to an
existing slave that already has a different master1 for domain_id=0. When the
slave tries to connect to master2, master2 will not have anything that slave
requests in domain_id=0, but that is fine as master2 is supposedly meant to
serve eg. domain_id=1. (This is MDEV-4485).

But suppose that master2 then actually starts sending events from
domain_id=0. In this case, the fix for MDEV-4485 was incomplete, and the code
would fail to give the error that the position requested by the slave in
domain_id=0 was missing from the binlogs of master2. This could lead to lost
events or completely wrong replication.

The patch for this bug fixes this issue.

In addition, it cleans up the code a bit, getting rid of the fake_gtid_hash in
the code. And the error message when slave and master have diverged due to
alternate future is clarified, as requested in the bug description.

Includes all post-review fixes as well. 

(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-4782: Valgrind warnings (Conditional jump or move depends on uninitialised value) with InnoDB, semijoin
- in sub_select(): don't call table->file->position() when reading the first record
  produced an error.

Backport the fix olav.sandstaa@sun.com-20101102184747-qfuntqwj021imy9r:
"Fix for Bug#52660 Perf. regr. using ICP for MyISAM on range queries on an index containing TEXT" 
(together with further fixes in that code) into MyISAM and Aria.

mdev-4173: Wrong result (extra row) with semijoin=on, joins in outer query, LEFT JOIN in the subquery
Apply the patch from Patryk Pomykalski:
- create_internal_tmp_table_from_heap() will now return information whether
  the last row that we tried to write was a duplicate row.
(mysql-5.6 also has this change)

- Make JOIN::cleanup(true) also work correctly when the query is KILLed
  after join optimization was started but before a query plan was produced

MDEV-4536: ...sql/sql_base.cc:1598: bool close_thread_table(THD*, TABLE**): Assertion
- Make JOIN::cleanup(full=true) always free join optimization tabs.

- merge_same_index_scans() may put the same SEL_ARG tree in multiple result plans.
  make it call incr_refs() on the SEL_ARG trees that it does key_or() on, because 
  key_or(sel_arg_tree_1, sel_arg_tree_2) call may invalidate SEL_ARG trees pointed 
  by sel_arg_tree_1 and sel_arg_tree_2.
  

When a new master is provisioned that does not have any old binlogs,
the @@gtid_slave_pos is used to know where in the GTID history the
provisioning happened. A slave is allowed to connect at the point of
this value of @@gtid_slave_pos, even if that GTID is not in the
binlogs on the new master.

The code to handle this case when the binlog on the newly provisioned
master is completely empty was just wrong (couple of typos). Clearly it
had never been tested ... :-/

When a new master is provisioned that does not have any old binlogs,
the @@gtid_slave_pos is used to know where in the GTID history the
provisioning happened. A slave is allowed to connect at the point of
this value of @@gtid_slave_pos, even if that GTID is not in the
binlogs on the new master.

But --gtid-strict-mode did not correctly handle this case. When strict
mode was enabled, an attempt to connect at the position would cause an
error about holes in the binlog, which is not correct.

This patch adds a hash of GTIDs that need to be treated specially by
GTID strict mode to deal correctly with this case.

MDEV-4409 - Fix deadlock in MySQL key cache code, that can happen if there is a key cache resize running in parallel with an update.

If there is a key cache resize,a  thread writing to key cache, will pause waiting  until resize finishes. However this thread is won't be woken, because resize does not  signaling waiters anymore. This is a regression introduced in WL#86(segmented MyISAM key cache)
The fix is to unconditionally release  threads waiting on resize_queue when resize  finishes, as in pre-WL#86 code.

modified:
  mysql-test/suite/plugins/r/fulltext_plugin.result
  mysql-test/suite/plugins/t/fulltext_plugin.test
  plugin/fulltext/plugin_example.c
  sql/sql_show.cc
pending merges:
  Alexander Barkov 2013-07-09 [merge] Merge from 5.1
    Sergei Golubchik 2013-07-06 Bug #69682 - mysqld crashes after uninst...
    Sergei Golubchik 2013-05-24 MDEV-4575 MySQL client doesn't strip off...

modified:
  mysql-test/suite/plugins/r/fulltext_plugin.result
  mysql-test/suite/plugins/t/fulltext_plugin.test
  plugin/fulltext/plugin_example.c
  sql/sql_show.cc
pending merges:
  Sergei Golubchik 2013-07-06 Bug #69682 - mysqld crashes after uninstall ...
    Sergei Golubchik 2013-05-24 MDEV-4575 MySQL client doesn't strip off...

modified:
  include/my_time.h
  libmysql/libmysql.c
  mysql-test/r/adddate_454.result
  mysql-test/r/cast.result
  mysql-test/r/date_formats.result
  mysql-test/r/func_sapdb.result
  mysql-test/r/func_time.result
  mysql-test/r/mdev316.result
  mysql-test/r/parser.result
  mysql-test/r/partition_datatype.result
  mysql-test/r/partition_pruning.result
  mysql-test/r/type_date.result
  mysql-test/r/type_datetime.result
  mysql-test/suite/vcol/r/vcol_misc.result
  mysql-test/t/cast.test
  sql-common/my_time.c
  sql/field.cc
  sql/field_conv.cc
  sql/filesort.cc
  sql/item.cc
  sql/item.h
  sql/item_cmpfunc.cc
  sql/item_func.cc
  sql/item_strfunc.cc
  sql/item_timefunc.cc
  sql/sql_time.cc
pending merges:
  Sergei Golubchik 2013-07-03 MDEV-4667 DATE('string') incompability betwe...

MDEV-4643 MariaDB crashes consistently when trying a SELECT on VIEW with a UNION and an additional JOIN in SELECT

open derived temp tables *before* trying QUICK_SELECT for them,
handler::multi_range_read_info() needs an open table.

don't ignore the return value fix_fields()

MDEV-4257 Assertion `!table || (!table->read_set || bitmap_is_set(table->read_set, field_index))' fails on FROM subquery with fulltext search, derived_merge=on

remove emtpty Item_func_match::update_used_tables() method

Fix of nested join parsing of illegal query.

Cleanup: remove TIME_FUZZY_DATE.

Introduce TIME_FUZZY_DATES which means "very fuzzy, the resulting
value is only used for comparison. It can be invalid date, fine, as long as it can be
compared".

Updated many tests results (they're better now).

for field terminators, uchar was compared with char

pending merges:
  Alexander Barkov 2013-06-28 A clean-up for MDEV-4634

modified:
  sql/time.cc

modified:
  mysql-test/r/type_date.result
  mysql-test/r/type_datetime.result
  mysql-test/r/type_time.result
  mysql-test/t/type_date.test
  mysql-test/t/type_datetime.test
  mysql-test/t/type_time.test
  sql/item_func.cc
  sql/item_timefunc.cc
  sql/sql_time.cc
  sql/sql_time.h
pending merges:
  Alexander Barkov 2013-06-28 MDEV-4634 Crash in CONVERT_TZ

Item_func_min_max::get_date() did not check the
returned value against the fuzzy_date flags, so
it could return a bad value to the caller that
expects a good date (e.h. CONVERT_TZ).


modified:
  mysql-test/r/type_date.result
  mysql-test/r/type_datetime.result
  mysql-test/r/type_time.result
  mysql-test/t/type_date.test
  mysql-test/t/type_datetime.test
  mysql-test/t/type_time.test
  sql/item_func.cc
  sql/item_timefunc.cc
  sql/mysql_priv.h
  sql/time.cc

Do not use x64  assembler implementation in  x32.