and other fixes for gcc-4.9.1 on sid

don't install them at all

don't send an OK packet if the SHOW PROCESSLIST was killed

* remove bundled jemalloc, use the system one
* force jemalloc in release builds on linux

Avoided exponential recursive calls of JOIN_CACHE::join_records() in the case
of non-nested outer joins.
A different solution is required to resolve this performance problem for
nested outer joins.

fix readline/libedit detection:
* search in readline/, editline/ and edit/readline/
* fix typos CMAKE_REQUIRES_LIBRARIES -> CMAKE_REQUIRED_LIBRARIES
* use correct libedit API
* use different cmake variables for libedit and readline

MDEV-6605 Multiple Clients Inserting Causing Error: Failed to read auto-increment value from storage engine

* handler::get_auto_increment() was not expecting any errors from the storage engine.
  That was wrong, errors could happen.
* ha_partition::get_auto_increment() was doing index lookups in partition under a mutex.
  This was redundant (engine transaction isolation was covering that anyway)
  and harmful (causing deadlocks).

if it has to be a mess of ifdefs, then mess of ifdefs it is

MDEV-6595 [PATCH] HPPA: storage/xtradb/os/os0stacktrace.c:88:54: error: invalid operands to binary & (have 'void *' and 'long unsigned int')

fix precedence as in the patch from Brian Evans

  if it's run on a real disk
  
  Made test smaller.

only look at lex->option_type if it's a SHOW command, not a SELECT

MDEV-6483 - Deadlock around rw_lock_debug_mutex on PPC64

This problem affects only debug builds on PPC64.

There are at least two race conditions around
rw_lock_debug_mutex_enter and rw_lock_debug_mutex_exit:

- rw_lock_debug_waiters was loaded/stored without setting
  appropriate locks/memory barriers.
- there is a gap between calls to os_event_reset() and
  os_event_wait() and in such case we're supposed to pass
  return value of the former to the latter.

Fixed by replacing self-cooked spinlocks with system mutexes.
These days system mutexes offer much better performance. OTOH
performance is not that critical for debug builds.

MDEV-6450 - MariaDB crash on Power8 when built with advance tool
            chain

InnoDB mutex_exit() function calls __sync_test_and_set() to release
the lock. According to manual this function is supposed to create
"acquire" memory barrier whereas in fact we need "release" memory
barrier at mutex_exit().

The problem isn't repeatable with gcc because it creates
"acquire-release" memory barrier for __sync_test_and_set().
ATC creates just "acquire" barrier.

Fixed by creating proper barrier at mutex_exit() by using
__sync_lock_release() instead of __sync_test_and_set().

DB-713 separate some long running tokudb tests so that valgrind runs without --big-test can exclude them

MDEV-6601 Assertion `!thd->in_active_multi_stmt_transa ction() || thd->in_multi_stmt_transaction_mode()' failed on executing a stored procedure with commit

Don't restore the whole of thd->server_status after a routine invocation,
only restore SERVER_STATUS_CURSOR_EXISTS and SERVER_STATUS_LAST_ROW_SENT,
as --ps --embedded needs.
In particular, don't restore SERVER_STATUS_IN_TRANS.

Part of this work is based on Stewart Smitch's memory barrier and lower priori
patches for power8.

- Added memory syncronization for innodb & xtradb for power8.
- Added HAVE_WINDOWS_MM_FENCE to CMakeList.txt
- Added os_isync to fix a syncronization problem on power
- Added log_get_lsn_nowait which is now used srv_error_monitor_thread to ensur
  if log mutex is locked.

All changes done both for InnoDB and Xtradb

Change a couple of permissions that cause lintian warnings in .deb packaging and don't really hurt to fix.

sporadically

Fix: Modify test to be smaller so that testcase timeout does not
trigger. We already have a test for --big-test setup
(innodb.innodb_simulate_comp_failures).

Followup. Don't compile threadpool_unix.cc when thread pool is disabled