Test was affected by incompletely closed preceding connections.

Make test agnostic to concurrent connections by querying
performance_schema.threads only for connections that it uses.

Embedded mode produces different thread identifiers, which makes
replace_result replace unwanted occurences. Use temporary table
for filtering out unwanted threads such that replace_result is
not needed anymore.

This is regression after 0eebe3a.

…UND mismatch)

Test was affected by incompletely closed preceding connections.

Make test agnostic to concurrent connections by querying
performance_schema.threads only for connections that it uses.

Revert part of bead24b, which broke this test. The wait is still
needed to make variables of preceding connection disappear.

Test was affected by incompletely closed preceding connections.

Make test agnostic to concurrent connections by querying
performance_schema.user_variables_by_thread only for connections
that it uses.

Test was affected by incompletely closed preceding connections.

Make test agnostic to concurrent connections by querying
performance_schema.events_statements_current only for connections
that it uses.

Similar to MDEV-37483, use file name encoding for dbnames to create
directories. Adapt mariadb-import to convert the names back.

…ecution

Routine name resolution performed on a temporary memory root
during execution. So on the second execution LEX::spname members
pointed to a cleared memory.

Fixing to peform the resolution to parse time, like the CALL statement does.
During the execution time LEX::spname members now stay untouched.

Send ok packet earlier for SELECT queries that does not have any updates.
This is done in select_send::send_eof()

Axel's select benchmarks shows that this has a notable speed improvement:
1  Thread:  28% TPS speedup
8  Threads: 23%
64 Threads  6%

This was meassured with t_oltp_point_select With InnoDB running the client
over sockets.

Other things:
- Moved error reporting of LIMIT ROWS EXAMINED from handle_select() to
  check_limit_rows_examined(). This is to ensure that the error is
  reported before send_eof() is called.
- Removed duplicate "Query execution was interrupted" messages for the
  same query. killed_for_exceeding_limit_rows_warning_given was introduced
  for this purpose. We cannot use 'killed' flag in
  killed_for_exceeding_limit_rows() to detect if we already have produced
  a warning as the 'killed' in this case is reset for each union.

Reviewer: Sergei Golubchik <[email protected]>

correct the length check.
remove assertions that a file read from disk contains a specific substring

…e-level CREATE but not with global CREATE

CHECK TABLE was inconsistently requiring SELECT privilege on global/db level
or any privilege on the table/column level.

Change to require any table-applicable privilege on any level.

don't trust the content of a file read from disk

… queries to fail

don't copy field default values and check constraints in
CREATE ... SELECT.

CREATE ... SELECT means a table is created from a *result set*
not from some other table.

For backward compatibility, though, let's keep copying constant
default values and the "compressed" attribute.

…DEFAULT

VALUE() should only use table->insert_values when
table->insert_values contains row values.

table->insert_values gets row values for the ODKU clause
so if VALUE() is used before that it shouldn't use table->insert_values

…sions 10.11.X and beyond

my_getcputime() returns "cpu time in 1/10th on a microsecond (1e-7 s)"

otherwise it causes random failures in some later test that lists
files in $datadir/test

don't let the parser create ridiculously deep joins that
will be rejected later anyway

the "Test that bad value for plugin enum option is rejected correctly"
needed multiple fixes:

1. don't set plugin-dir based on $MYSQLTEST_VARDIR, all plugins are
  in var/plugins, but $MYSQLTEST_VARDIR is var/1/, var/2/, etc if
  --parallel is used (that is, practically always), thus the ha_example.so
  cannot be loaded, because cannot be found. Test fails with
  "unknown option --plugin-example-enum" as the plugin is not loaded

2. force --plugin-maturity=experimental, otherwise even if not parallel
   the plugin will fail to load because of low maturity, test still
   fails with "unknown option --plugin-example-enum"

3. don't specify .so extension explicitly otherwise the plugin still
   doesn't load on windows, even if paths and maturity are fixed

4. set --plugin-example=FORCE otherwise plugin fails to load after reading
   --plugin-example-enum-var=noexist because of unknown enum value,
   the server ignores the failure and starts normally. the test hangs.

5. This needs the fix in sql_plugin.cc to detect that the plugin is
   forced even when some options failed to parse. It used to consider
   plugin forced only if all options parsed correctly, which was wrong.

Now the test passes, testing what it was supposed to test - failure
to parse an enum value of a plugin option.

Without these fixes the test hanged as in 4 when run on the main branch
in non-prarallel (e.g. one test only) mode.

…keys on vault errors

* let use_cache_on_timeout apply to other errors
* enable use_cache_on_timeout by default and deprecate it
* increase cache_timeout to max and deprecate it
* change it from long to portable longlong
* delete both in 13.3

* put autocommit/commit outside of LOCK/UNLOCK.
* use uppercase like all other commands
* restore the old value of autocommit

with `ORDER BY number` if the number doesn't refer to a valid
result column, use this number in the error message not '???'.

restore OPTIMIZE/ANALYZE replication under @read_only
that was disabled in b62101f

apparently a file can be present in *more than one* rpm,
e.g. /usr/bin/dtrace on rhel10 is present *both* in
systemtap-sdt-dtrace and in systemtap-sdt-devel.

Make sure there's a separator between entries.

don't use such a greedy regex_replace pattern

event scheduler was printing a lot of info in [Note] in error log.

change to print its startup/shutdown notes only when log_warnings>0.
and runtime notes only when log_warnings>2.
one note was an abnormal error, change to [Error].

if ((res=item->val_str(str)) != str)

is incorrect way to detect whether res can be safely used,
because Item_char_typecast::val_str() can return res
which is different from str, but shares the same buffer.

Master_Server_Id was not cleared after CHANGE MASTER or RESET SLAVE,
showing a stale value until the slave reconnected. Reset master_id
and prev_master_id to 0 in both code paths.

The reset value (0) will be present in SHOW SLAVE STATUS until it is
re-evaluated to the id of a new connected master server.

Signed-off-by: Varun Deep Saini <[email protected]>

There is a lot of more work needed to make this code compatitble with
the rest of the MariaDB server.

- All std:: usage should be removed
- Most templatest should be removed
- Usage of <unordered_map>, <string_view>, <optional>, <unordered_set>
  should be removed and replaced with my_sys, sql_string functions

It would be nice of one could also have default values for all the
CHANGE MASTER variables in my.cnf.  This would allow one to have
much shorter CHANGE MASTER commands

…ns (fix)

Correct compilation on clang-20 by using a bitwise & rather than the
incorrect &&.