In the hexlo function, the element type of the array hex_lo_digit is not
explicitly declared as signed char, causing elements with a value of -1
to be converted to 255 on Arm64. The problem occurs because "char" is
unsigned by default on Arm64 compiler, but signed on x86 compiler. This
problem can be seen in https://godbolt.org/z/rT775xshj

The above issue causes "use-after-poison" exception in my_mb_wc_filename
function. The code snippet where the error occurred is shown below,
copied from below link.
https://github.com/MariaDB/server/blob/5fc19e71375fb39eb85354321bf852d998aecf81/strings/ctype-utf8.c#L2728

2728    if ((byte1= hexlo(byte1)) >= 0 &&
2729     (byte2= hexlo(byte2)) >= 0)
  	{
2731    	int byte3= hexlo(s[3]);
    		…
  	}

At line 2729, when byte2 is 0, which indicates the end of the string s.
(1) On x86, hexlo(0) return -1 and line 2731 is skipped, as expected.
(2) On Arm64, hexlo(0) return 255 and line 2731 is executed, not as
expected, accessing s[3] after the null character of string s, thus
raising the "user-after-poison" error.

The problem was discovered when executing the main.mysqlcheck test.
Signed-off-by: Xiaotong Niu <xiaotong.niu@arm.com>

fil_page_compress_low returns 0 for both innodb_compression_algorithm=0
and where there is compression errors. On the two callers to this
function, don't increment the compression errors if the algorithm was
none.

Reviewed by: Marko Mäkelä

The memory allocated for an instance of the class Item_direct_ref_to_item
was leaked on second execution of a query run as a prepared statement and
involving conversion of strings with different character sets.

The reason of leaking the memory was caused by the fact that a statement
arena could be already set by the moment the method
Type_std_attributes::agg_item_set_converter() is called.

MDEV-24283 Assertion `bitmap_is_set(&m_part_info->read_partitions, m_part_spec.start_part)' failed in ha_partition::handle_ordered_index_scan

ha_partition should not try to search the index
if all partitions were pruned away.

The fix originally by Nayuta Yanagisawa

MyISAM tables no longer take transactional metadata locks
unless there already is an active transaction.

MDEV-27523 main.delayed fails with wrong error code or timeout when executed after main.deadlock_ftwrl

don't forget to reset mdl_context.m_deadlock_overweight when
taking the THD out of the cache - the history of previous connections
should not affect the weight in deadlock victim selection

(small cleanup of the test to help the correct merge)

try harder to punch holes on xfs, don't rely on its heuristics to do
the right thing

Due to the different command line format of the timeout
utility on FreeBSD and Linux, SST scripts for mariabackup
may not work on FreeBSD. This commit fixes this problem
by adding a different command to test options on FreeBSD
and adding proper formatting for the utility options.

When aggregating pairs BIT+NULL and NULL+BIT for result, e.g.
in COALESCE(), preserve the BIT data type (ignore explicit NULLs).

The same fix applied to YEAR.

This bug could affect queries with IN subqueries in WHERE clause and using
derived tables to which split optimization potentially could be applied.

When looking for the best split of a splittable derived table T any key
access from a semi-join materialized table used for lookups S to table T
must be excluded from consideration because in the current implementation
of such tables as S the values from its records cannot be used to access
other tables.

Approved by Oleksandr Byelkin <sanja@mariadb.com>

Author: Sergei Petrunia <sergey@mariadb.com>
Date:   Wed Oct 11 19:02:25 2023 +0300

    MDEV-32301: Server crashes at Arg_comparator::compare_row

    In  Item_bool_rowready_func2::build_clone(): if we're setting
      clone->cmp.comparators=0
    also set
      const_item_cache=0
    as the Item is currently in a state where one cannot compute it.

A subquery in form "(SELECT not_null_value LIMIT 1 OFFSET 1)" will
produce no rows which will translate into scalar SQL NULL value.

The code in Item_singlerow_subselect::fix_length_and_dec() failed to
take the LIMIT/OFFSET clause into account and used to set
item_subselect->maybe_null=0, despite that SQL NULL will be produced.

If such subselect was used in ORDER BY, this would cause a crash in
filesort() code when it would get a NULL value for a not-nullable item.

also made subselect_engine::no_tables() const function.

In  Item_bool_rowready_func2::build_clone(): if we're setting
  clone->cmp.comparators=0
also set
  const_item_cache=0
as the Item is currently in a state where one cannot compute it.

The code inside Item_subselect::fix_fields() could fail to check
that left expression had an Item_row, like this:

  (('x', 1.0) ,1) IN (SELECT 'x', 1.23 FROM ... UNION ...)

In order to hit the failure, the first SELECT of the subquery had
to be a degenerate no-tables select. In this case, execution will
not enter into Item_in_subselect::create_row_in_to_exists_cond()
and will not check if left_expr is composed of scalars.

But the subquery is a UNION so as a whole it is not degenerate.
We try to create an expression cache for the subquery.
We create a temp.table from left_expr columns. No field is created
for the Item_row. Then, we crash when trying to add an index over a
non-existent field.

Fixed by moving the left_expr cardinality check to a point in
check_and_do_in_subquery_rewrites() which gets executed for all
cases.
It's better to make the check early so we don't have to care about
subquery rewrite code hitting Item_row in left_expr.

Part 1

SST for mariabackup may not destroy old files if datadir or
other working directory is declared as a symlink due to the lack
of the "-L" option among the find utility options, similarly SST
for rsync in some cases may not transfer data directories if they
are created as symlinks. This fix adds the missing option and
generally unifies the work with find utility options to avoid
failures in the interpretation of directories and regular
expressions.

While checking for altered column in foreign key constraints,
InnoDB fails to ignore virtual columns. This issue caused
by commit 5f09b53b(MDEV-31086).

- InnoDB should avoid the sync commit operation when
there is nothing in fulltext cache. This is caused by
commit 1248fe72 (MDEV-27582)

the bug with %attr(700,%{mysqld_user},-)
was fixed not right after 2.8.12, but only in 3.10.0

https://gitlab.kitware.com/cmake/cmake/-/commit/a351edd245

MDEV-32249 strings/ctype-ucs2.c:2336: my_vsnprintf_utf32: Assertion `(n % 4) == 0' failed in my_vsnprintf_utf32 on INSERT

The crash inside my_vsnprintf_utf32() happened correctly,
because the caller methods:
  Field_string::sql_rpl_type()
  Field_varstring::sql_rpl_type()
mis-used the charset library and sent pure ASCII data to the
virtual function snprintf() of a utf32 CHARSET_INFO.

It was wrong to use Field::charset() in sql_rpl_type().
We're printing the metadata (the data type) here, not the column data.
The string contraining the data type of a CHAR/VARCHAR column
is a pure ASCII string.

Fixing to use res->charset() to print, like all virtual implementations
of sql_type() do.

Review was done by Andrei Elkin.
Thanks to Andrei for proposing MTR test improvents.

The cmake configuration step is single-threaded and already consuming
too much time. We should not make it worse by adding invocations like
MY_CHECK_CXX_COMPILER_FLAG().

Let us prefer something that works on any supported version
of GCC (4.8.5 or later) or clang, as well as recent versions
of the Intel C compiler.

This replaces commit 1fde7853

followup for 96ae37ab

gcc 13.2.1

When spider_db_delete_all_rows() is called, the supplied spider->conns
may have already been freed. The existing mechanism has spider_trx own
the connections in trx_conn_hash and it may free a conn during the
cleanup after a query. When running a delete query and if the table is
in the table cache, ha_spider::open() would not be called which would
recreate the conn. So we recreate the conn when necessary during
delete by calling spider_check_trx_and_get_conn().

We also reduce code duplication as delete_all_rows() and truncate()
has almost identical code, and there's no need to assign
wide_handler->sql_command in these functions because it has already
been correctly assigned.

Windows C runtime does not implement line buffering mode for stdio.

This sometimes makes output from different tests interleaved in MTR
MTR relies on this buffering (lines won't  output until "\n") to correctly
work in parallel scenarios.

Implement do-it-yourself line buffering on Windows, to workaround.

Windows C runtime does not implement line buffering mode for stdio.

This sometimes makes output from different tests interleaved in MTR
MTR relies on this buffering (lines won't  output until "\n") to correctly
workin parallel scenarios.

Implement do-it-yourself line buffering on Windows, to workaround.

Flush stdout on finalizing of mysqldump/mysqlbinlog output
to avoid truncation.

The same patch has been applied to the mysqltest.cc code with
  commit 34ff714b
  Author: Magnus Svensson <msvensson@mysql.com>
  Date:   Fri Nov 14 11:06:56 2008 +0100
    WL#4189 Make mysqltest flush log file at close if logfile is stdout

but not to mysqldump.c/mysqlbinlog.cc

Fixed tests:
main.secure_file_priv_win - disabled ps2-protocol for queries like
"SELECT ... INTO OUTFILE ..."

copy_back(): Also copy the dummy empty ib_logfile0 so that
MariaDB Server 10.8 or later can be started after
--copy-back or --move-back.

Thanks to Daniel Black for reporting this.

mysqladmin's presumption about the cause of an error by looking
at the error code was presumptious. Server knows best, so pass
the error along.

Avoid returning -1 as a exit code, Linux makes this 255 and Windows
keeps this as -1.

MDEV-30658 lock_row_lock_current_waits counter in information_schema.innodb_metrics may become negative

MONITOR_OVLD_ROW_LOCK_CURRENT_WAIT monitor should has
MONITOR_DISPLAY_CURRENT flag set in its definition, as it shows the
current state and does not accumulate anything.

Reviewed by: Marko Mäkelä

Fix order_by_optimizer_innodb and order_by_innodb tests.

The problem was that the query could be ran before InnoDB was
ready to provide a realistic statistic for #records in the table.

It provided a number that was too low, which caused the optimizer
to decide that range access plan wasn't advantageous and discard it.

MDEV-32275 getting error 'Illegal parameter data types row and bigint for operation '+' ' when using ITERATE in a FOR..DO

An "ITERATE innerLoop" did not work properly inside
a WHILE loop, which itself is inside an outer FOR loop:

outerLoop:
  FOR
   ...
   innerLoop:
    WHILE
      ...
      ITERATE innerLoop;
      ...
    END WHILE;
    ...
  END FOR;

It erroneously generated an integer increment code for the outer FOR loop.
There were two problems:
1. "ITERATE innerLoop" worked like "ITERATE outerLoop"
2. It was always integer increment, even in case of FOR cursor loops.

Background:
- A FOR loop automatically creates a dedicated sp_pcontext stack entry,
  to put the iteration and bound variables on it.

- Other loop types (LOOP, WHILE, REPEAT), do not generate a dedicated
  slack entry.

  The old code erroneously assumed that sp_pcontext::m_for_loop
  either describes the most inner loop (in case the inner loop is FOR),
  or is empty (in case the inner loop is not FOR).

  But in fact, sp_pcontext::m_for_loop is never empty inside a FOR loop:
  it describes the closest FOR loop, even if this FOR loop has nested
  non-FOR loops inside.

  So when we're near the ITERATE statement in the above script,
  sp_pcontext::m_for_loop is not empty - it stores information about
  the FOR loop labeled as "outrLoop:".

Fix:
- Adding a new member sp_pcontext::Lex_for_loop::m_start_label,
  to remember the explicit or the auto-generated label correspoding
  to the start of the FOR body. It's used during generation
  of "ITERATE loop_label" code to check if "loop_label" belongs
  to the current FOR loop pointed by sp_pcontext::m_for_loop,
  or belongs to a non-FOR nested loop.

- Adding LEX methods sp_for_loop_intrange_iterate() and
  sp_for_loop_cursor_iterate() to reuse the code between
  methods handling:
  * ITERATE
  * END FOR

- Adding a test for Lex_for_loop::is_for_loop_cursor()
  and generate a code either a cursor fetch, or for an integer increment.
  Before this change, it always erroneously generated an integer increment
  version.

- Cleanup: Initialize Lex_for_loop_st::m_cursor_offset inside
  Lex_for_loop_st::init(), to avoid not initialized members.

- Cleanup: Removing a redundant method:
    Lex_for_loop_st::init(const Lex_for_loop_st &other)
  Using Lex_for_loop_st::operator(const Lex_for_loop_st &other) instead.