INSTALL SONAME ignores attempts to load the same plugin twice, it's not an error
(because one can load one plugin by name and then install soname for the rest).
But Federated and FederatedX are different plugins, despite having the same name.

Now plugin_add() only considers two plugins identical if their names are the same
string (compared as pointers). Otherwise it reports an error.,

first update system tables, then the rest

remove ipv4-mapped-ipv6 test, this feature is not always available

            rpl.rpl_gtid_crash fail on PPC64

This is an addition to the original patch.

Restored show binlog events output and adjusted filters to
replace [\d-\d-\d,\d-\d-\d,\d-\d-\d] with [#-#-#].

fix_privilege_tables: use ALTER TABLE to fix last_update column in innodb stat tables 

Remove the "don't update the row for b'' and store uninitialized bytes on disk" change.
Update test cases to allow DEFAULT b'', because b'' is a valid expression elsewhere.

Don't ignore errors from Count_distinct_field::add(),
pass them to the caller, so that it could abort the data
collection loop.

when binding to wildcard addresses (no --bind is specified or --bind=*),
bind to both ipv6 and ipv4 as mysql-5.6 is doing.

when looking for my.cnf files: if DEFAULT_SYSCONFDIR (or INSTALL_SYSCONFDIR)
is specified (for rpms it always is), use that instead of hardcoded /etc path.

            tool chain

Reverted addition to the original patch:
InterlockedExchangeAcquire and InterlockedAndRelease are
supported only on Itanium-based systems.

disable tokudb paranoid asserts

Fix the bug properly (plugin cannot be unloaded as long as it's locked).
Enable and fix the test case.
Significantly reduce number of LOCK_plugin locks for semisync
(practically all locks were removed)

~40% bugfixed(*) applied
~40$ bugfixed reverted (incorrect or we're not buggy)
~20% bugfixed applied, despite us being not buggy
(*) only changes in the server code, e.g. not cmakefiles

            tool chain

This is an addition to the original patch. On Windows
InterlockedExchange implies full memory barrier, whereas
only acquire/release barriers required.