Adding a new virtual function MY_CHARSET_HANDLER::copy_abort().
Moving character set specific code into the correspoding implementations
(for simple, multi-byte and mbmaxlen>1 character sets).

error messages.

* remove useless suppressions from the test file, when a plugin doesn't
  load, the file isn't executed anyway
* add the suppression to mysql-test-run.pl instead

Take into account that PIE binaries are loaded at some offset, so
addresses cannot be directly resolved with addr2line. Find this
offset and subtract it before resolving an address.

MDEV-7000 Assertion `0' failed in Protocol::end_statement() on executing DDL under innodb_fake_changes=1

correct the if() condition to match the behavior of the old code
that this if() was supposed to replace

There were two issues:

* set_var_default_role::user was overwritten with a new value,
  allocated in the thd->mem_root, which is reset between executions.
  That was causing the crash. Solved by introducing set_var_default_role::real_user

* when privilege tables were opened on EXECUTE, the reprepare_observer
  would abort the statement (as privilege tables are opened using
  the local TABLE_LIST that doesn't preserve metadata from PREPARE, so
  reprepare_observer thought they're changed). This issue also applied
  to SET PASSWORD. Solved by disabling reprepare_observer.

Added logic to sort gtid list based on domain_id before
populating them in string. Added a test case.

Removed 'galera-sst-mode' option from mysqldump and added logic
in wsrep_sst_mysqldump script to retrieve gtid_binlog_state from
donor node and send it to the joiner node.

Added an include file to check galera library version.

* Changed loopback detection to be done via ifa->ifa_flags
* Removed unused function wsrep_guess_address()

by adding a new class String_copier.

This is a pre-requisite patch for MDEV-6566 and MDEV-6572,
to avoid adding more similar code.

Create_tmp_table was called incorrectly called in
select_materialized_with_stats::create_result_table, having keep_row_order
passed for the do_not_open parameter and keep_row_order always set to false.

file_key_management_plugin is used

Fixed error handling and added disabling InnoDB redo log encryption
if encryption key management plugin is not there.

- Remove the compiler warning, add assert statements.
- make select_describe() not call mysql_explain_union() for
  views that were "merged for INSERT".

file_key_management_plugin is used

Fixed error handling and added disabling InnoDB redo log encryption
if encryption key management plugin is not there.

of encryption/compression pages and that will happen if and only
if dirty pages are written to the disk.

    MDEV-7399: Add support for INFORMATION_SCHEMA.INNODB_MUTEXES
    MDEV-7618: Improve semaphore instrumentation

    Introduced two new information schema tables to monitor mutex waits
    and semaphore waits. Added a new configuration variable
    innodb_intrument_semaphores to add thread_id, file name and
    line of current holder of mutex/rw_lock.

Added support for compression method snappy for page compression.

Changed wsrep plugin from mysql_plugin to maria_plugin to
set up plugin_maturity to Stable.

encrypted tables

Introduced a new innodb_default_page_encryption_key configuration
variable to allow user to set the default key identifier.

put cracklib_password_check into a separate package.
build it conditionally on Debian/Ubuntu

namely, remove the code to support:
etch, lenny, hardy, intrepid, jaunty, karmic, maverick, natty

Remove ONLY_IF clause in MYSQL_ADD_PLUGIN and the requirement
that every plugin's CMakeLists.txt *must* do MYSQL_ADD_PLUGIN
for PLUGIN_XXX=YES to work. This was very fragile and cannot be
relied on.

Use a different implementation of =YES check - iterate all
PLUGIN_* variables and see which one doesn't have a matching target.

Revert all ONLY_IF changes in CMakeLists.txt files.

we don't want mysys static data structures to be overwritten
when a dynamic plugin is loaded.

This reverts commit aa61e4c7
that broke the compilation on few platforms (e.g. centos5).

These warnings are already fixed properly in 10.0-connect.