These were reported when running tests after compiling the debug server
with -O3 -fsanitize=undefined in GCC 7.1.0.

log_crypt(): Do not cast byte* to uint32_t*, because it may break
strict aliasing rules. Instead, cast in the opposite direction.

buf_read_ahead_linear(): Do not display a message if the tablespace
is being deleted.

dtype_print(): Add a missing break statement.

In my merge of the MySQL fix for Oracle Bug#23333990 / WL#9513
I overlooked some subsequent revisions to the test, and I also
failed to notice that the test is actually always failing.

Oracle introduced the parameter innodb_stats_include_delete_marked
but failed to consistently take it into account in FOREIGN KEY
constraints that involve CASCADE or SET NULL.

When innodb_stats_include_delete_marked=ON, obviously the purge of
delete-marked records should update the statistics as well.

One more omission was that statistics were never updated on ROLLBACK.
We are fixing that as well, properly taking into account the
parameter innodb_stats_include_delete_marked.

dict_stats_analyze_index_level(): Simplify an expression.
(Using the ternary operator with a constant operand is unnecessary
obfuscation.)

page_scan_method_t: Revert the change done by Oracle. Instead,
examine srv_stats_include_delete_marked directly where it is needed.

dict_stats_update_if_needed(): Renamed from
row_update_statistics_if_needed().

row_update_for_mysql_using_upd_graph(): Assert that the table statistics
are initialized, as guaranteed by ha_innobase::open(). Update the
statistics in a consistent way, both for FOREIGN KEY triggers and
for the main table. If FOREIGN KEY constraints exist, do not dereference
a freed pointer, but cache the proper value of node->is_delete so that
it matches prebuilt->table.

row_purge_record_func(): Update statistics if
innodb_stats_include_delete_marked=ON.

row_undo_ins(): Update statistics (on ROLLBACK of a fresh INSERT).
This is independent of the parameter; the record is not delete-marked.

row_undo_mod(): Update statistics on the ROLLBACK of updating key columns,
or (if innodb_stats_include_delete_marked=OFF) updating delete-marks.

innodb.innodb_stats_persistent: Renamed and extended from
innodb.innodb_stats_del_mark. Reduced the unnecessarily large dataset
from 262,144 to 32 rows. Test both values of the configuration
parameter innodb_stats_include_delete_marked.
Test that purge is updating the statistics.

innodb_fts.innodb_fts_multiple_index: Adjust the result. The test
is performing a ROLLBACK of an INSERT, which now affects the statistics.

include/wait_all_purged.inc: Moved from innodb.innodb_truncate_debug
to its own file.

In some rare cases queries with UNION ALL
using a derived table specified by
a grouping select with a subquery in WHERE and
impossible HAVING detected after constant row
substitution could hang.
The cause was not a proper return from the
function subselect_single_select_engine::exec()
in the case when the subquery was not optimized
beforehand and the optimization performed
in this function requested for a change of the
subquery engine. This was fixed.
Also a change was applied that avoided execution
of a subquery if impossible having was detected
for the main query at the optimization stage.

Issue FLUSH TABLES to prevent unintended corruption of
crash-unsafe tables when the server is killed.

Throttling only works with when creating backup.  Attempt to use it with
 --copy-back results in crash, since throttle events are not initialized.
Thus, ignore throttling unless --backup is given.

available

lz4.cmake: Check if shared or static lz4 library has LZ4_compress_default
function and if it has define HAVE_LZ4_COMPRESS_DEFAULT.

fil_compress_page: If HAVE_LZ4_COMPRESS_DEFAULT is defined use
LZ4_compress_default function for compression if not use
LZ4_compress_limitedOutput function.

Introduced a innodb-page-compression.inc file for page compression
tests that will also search .ibd file to verify that pages
are compressed (i.e. used search string is not found). Modified
page compression tests to use this file.

Note that snappy method is not included because of MDEV-12615
InnoDB page compression method snappy mostly does not compress pages
that will be fixed on different commit.

This fixes warnings that were emitted when running InnoDB test
suites on a debug server that was compiled with GCC 7.1.0 using
the flags -O3 -fsanitize=undefined.

thd_requested_durability(): XtraDB can call this with trx->mysql_thd=NULL.
Remove the function in InnoDB, because it is not used there.

calc_row_difference(): Do not call memcmp(o_ptr, NULL, 0).

innobase_index_name_is_reserved(): This can be called with
key_info=NULL, num_of_keys=0.

innobase_dropping_foreign(), innobase_check_foreigns_low(),
innobase_check_foreigns(): This can be called with
drop_fk=NULL, n_drop_fk=0.

rec_convert_dtuple_to_rec_comp(): Do not invoke memcpy(end, NULL, 0).

On 64-bit systems, the constant 1 would be 32-bit (int or unsigned)
by default. Cast the constant to ulint before shifting to avoid a
-fsanitize=undefined warning or any potential overflow.

Fix a -fsanitizer=undefined warning that trx_undo_report_row_operation()
was being passed thr=NULL when the BTR_NO_UNDO_LOG_FLAG flag was set.

trx_undo_report_row_operation(): Remove the first two parameters.
The parameter clust_entry!=NULL distinguishes inserts from updates.

This should be a non-functional change (no observable change in
behaviour; slightly smaller code).

Allocate srv_sys statically so that the desired alignment can be
guaranteed. This silences -fsanitize=undefined warnings.
There probably is no performance impact of this, because the
reason for the alignment to ensure the absence of false sharing
between counters. Even with the misalignment, each counter would
have been been aligned at 64 bits, and the counters would reside
in separate cache lines.

The parameter thr of the function btr_cur_optimistic_insert()
is not declared as nonnull, but GCC 7.1.0 with -O3 is wrongly
optimizing away the first part of the condition
UNIV_UNLIKELY(thr && thr_get_trx(thr)->fake_changes)
when the function is being called by row_merge_insert_index_tuples()
with thr==NULL.

The fake_changes is an XtraDB addition. This GCC bug only appears
to have an impact on XtraDB, not InnoDB.

We work around the problem by not attempting to dereference thr
when both BTR_NO_LOCKING_FLAG and BTR_NO_UNDO_LOG_FLAG are set
in the flags. Probably BTR_NO_LOCKING_FLAG alone should suffice.

btr_cur_optimistic_insert(), btr_cur_pessimistic_insert(),
btr_cur_pessimistic_update(): Correct comments that disagree with
usage and with nonnull attributes. No other parameter than thr can
actually be NULL.

row_ins_duplicate_error_in_clust(): Remove an unused parameter.

innobase_is_fake_change(): Unused function; remove.

ibuf_insert_low(), row_log_table_apply(), row_log_apply(),
row_undo_mod_clust_low():
Because we will be passing BTR_NO_LOCKING_FLAG | BTR_NO_UNDO_LOG_FLAG
in the flags, the trx->fake_changes flag will be treated as false,
which is the right thing to do at these low-level operations
(change buffer merge, ALTER TABLE…LOCK=NONE, or ROLLBACK).
This might be fixing actual XtraDB bugs.

Other callers that pass these two flags are also passing thr=NULL,
implying fake_changes=false. (Some callers in ROLLBACK are passing
BTR_NO_LOCKING_FLAG and a nonnull thr. In these callers, fake_changes
better be false, to avoid corruption.)

The parameter path can be passed as NULL.
This error was reported by GCC 7.1.0 when compiling
CMAKE_BUILD_TYPE=Debug with -O3.

The macro UT_LIST_INIT() zero-initializes the UT_LIST_NODE.
There is no need to call this macro on a buffer that has
already been zero-initialized by mem_zalloc() or mem_heap_zalloc()
or similar.

For some reason, the statement UT_LIST_INIT(srv_sys->tasks) in
srv_init() caused a SIGSEGV on server startup when compiling with
GCC 7.1.0 for AMD64 using -O3. The zero-initialization was attempted
by the instruction movaps %xmm0,0x50(%rax), while the proper offset
of srv_sys->tasks would seem to have been 0x48.

This is a non-functional change to make it slightly easier
to read the code. We seem to have some bugs in this
IMPORT TABLESPACE code; see MDEV-12396.

Do not silence uncertain cases, or fix any bugs.

The only functional change should be that ha_federated::extra()
is not calling DBUG_PRINT to report an unhandled case for
HA_EXTRA_PREPARE_FOR_DROP.

Do not silence uncertain cases, or fix any bugs.

The only functional change should be that ha_federated::extra()
is not calling DBUG_PRINT to report an unhandled case for
HA_EXTRA_PREPARE_FOR_DROP.

The issue was that my_errno was not set properly when a repair was killed,
which confused the rpl_killed_ddl script.

I also added an extra test line in varchar.inc to ensure we don't give
duplicate error rows.

Translate "multi:" to SPH_GROUPBY_MULTIPLE.
SPH_GROUPBY_ATTRPAIR is still not supported.

Don't use instal_jar() from Use_Java.cmake -
at least in CentOS 7 (cmake 2.8.12.2) it doesn't yet support
the COMPONENT option.

remove dead code: this "return rc" could've never happened,
because reset_stmt_handle can only possibly fail if RESET_SERVER
flag is used.

But this "return rc" was confusing, it was under DBUG_ENTER (while
not being DBUG_RETURN) and it was skipping required stmt cleanup
at the end of the function.

CSV engine didn't expect CSM files to be read-only

remove useless checks
and a couple of others

bunch of bugs when external_lock() fails on unlock:
* mi_lock_database() used mi_mark_crashed() under share->intern_lock,
  but mi_mark_crashed() itself locks this mutex.
* handler::close() required table to be unlocked, but failed
  external_lock didn't count as unlock
* mysql_unlock_tables() ignored all unlock errors, but they still set
  the error status in stmt_da.

post-review fixes
* move pcre-specific variable out of mysys
* don't use current_thd
* move a commonly used macro to my_sys.h
* remove new sysvar