my_rwlock_destroy=>rwlock_destroy

- Fix  active lock in freed memory in ha_archive (share mutex was not released prior to free())
- Do not attempt vio_fastsend or vio_keepalive on named pipes and shared memory.

LPBUG#782223 : Memory released by Query_cache::resize() or Query_cache::free() contains active rwlocks.
The bug was found by application verifier. 
Fixed by destroying locks prior to free(),

LPBUG#782269 : critical sections are initialized twice in   xt_xn_init_db() 
LPBUG#782431: active lock in memory released by   xt_ind_exit()
LPBUG#782433 : xt_heap_release() does not release spinlock  hp->h_lock initialized in xt_heap_new().
LPBUG#782435: xt_exit_row_locks() tries to release unallocated locks

Simplifies handling 5.1 in buildbot.

On Windows, do not attempt access() for special device names like
CON, PRN etc. access() would return 0, this does not mean that fiile
with this name exists.

mysql-test/r/create.result:
  test of renaming
mysql-test/r/upgrade.result:
  Now such behaviour prohibited to avoid problems.
mysql-test/t/create.test:
  test of renaming
mysql-test/t/upgrade.test:
  Now such behaviour prohibited to avoid problems.
sql/mysql_priv.h:
  Function to test table name presence added.
sql/sql_rename.cc:
  Rename fixed.
sql/sql_table.cc:
  Function to test table name presence added.
  Create fixed.

no unnecessary casting from void*, more type safety.
typos fixed.

Problem was setting DEBUG_SYNC twice in a row too fast in the test case; this
could cause the second setting to override the first before the code had time
to react to the first, causing the signal to get lost.

Fixed by waiting for the code to receive the first signal before overwriting
it in the test case.

(Test failed if we added my_sleep(200000) in event_queue::cond_wait() just before pthread_cond_wait();  Not likely scenario but better to get that fixed too)

Fixed race condition in event that could cause hang when stopping event scheduler with SET GLOBAL event_scheduler=OFF 

client/mysqltest.cc:
  Write command to be executed to the log BEFORE executing the command.
  This makes it easier to debug crashes as the log will contain the fatal command.
mysql-test/r/mysqltest.result:
  Updated results (we now get more things logged)
sql/event_queue.cc:
  Fixed race condition in event that could cause hang when stopping event scheduler with SET GLOBAL event_scheduler=OFF.
  The reason was that a kill signal could be sent between last check of thd->killed and before thd->enter_cond() in which case the signal
  would be missed and we would be stuck in Event_scheduler::stop() forever.

Reverted unittest/unit.pl back to Test::Harness as some of our build machines didn't support the new recommended TAP::Harness module

unittest/mytap/tap.c:
  Fixed output for some tests that didn't call plan()

storage/maria/unittest/ma_test_loghandler-t.c:
  Don't sync during test
storage/maria/unittest/ma_test_loghandler_multigroup-t.c:
  Don't sync during test
storage/maria/unittest/ma_test_loghandler_multithread-t.c:
  Don't sync during test
unittest/mysys/bitmap-t.c:
  Don't test all bit combinations (not needed)
unittest/mysys/thr_template.c:
  Remove sleep as old bug should be fixed nowadays
unittest/mysys/waiting_threads-t.c:
  Only run test with --big
unittest/mytap/tap.c:
  Print total time at end of test.
unittest/unit.pl:
  Use TAP::Harness instead of Test::Harness (recommended according to manual)
  Add times to tests.

Fixed compiler warnings

client/readline.cc:
  Fixed compiler warning
mysql-test/suite/innodb/t/innodb_bug60049.test:
  This test failed when running with --mysqld=--loose-innodb-fast-shutdown=2 which we do on some machines
mysql-test/t/mysqldump.test:
  Only run test if utf8 is used
sql/log.cc:
  Fixed compiler warning
sql/mysql_priv.h:
  Fixed compiler warnings
tests/mysql_client_test.c:
  Don't abort test if ucs2 is not in use.

Fixed compiler warnings

client/readline.cc:
  Fixed compiler warning
mysql-test/t/mysqldump.test:
  Only run test if utf8 is used
sql/log.cc:
  Fixed compiler warning
sql/mysql_priv.h:
  Fixed compiler warnings
tests/mysql_client_test.c:
  Don't abort test if ucs2 is not in use.

into 5.1

The reason for mtr warning is that collect_mysqld_features() starts mysqld with --datadir=/tmp and this 
directory does not exist on Windows.

Fix : instead of passing --datadir=$opt_vardir/tmp in collect_mysqld_features() just use --datadir=.
mysqld does not need a correct directory, just an existing one, as it is started with --help ---verbose -skip-grant-tables.