Make mysqltest to use --ps-protocol more

use prepared statements for everything that server supports
with the exception of CALL (for now).

Fix discovered test failures and bugs.

tests:
* PROCESSLIST shows Execute state, not Query
* SHOW STATUS increments status variables more than in text protocol
* multi-statements should be avoided (see tests with a wrong delimiter)
* performance_schema events have different names in --ps-protocol
* --enable_prepare_warnings

mysqltest.cc:
* make sure run_query_stmt() doesn't crash if there's
  no active connection (in wait_until_connected_again.inc)
* prepare all statements that server supports

protocol.h
* Protocol_discard::send_result_set_metadata() should not send
  anything to the client.

sql_acl.cc:
* extract the functionality of getting the user for SHOW GRANTS
  from check_show_access(), so that mysql_test_show_grants() could
  generate the correct column names in the prepare step

sql_class.cc:
* result->prepare() can fail, don't ignore its return value
* use correct number of decimals for EXPLAIN columns

sql_parse.cc:
* discard profiling for SHOW PROFILE. In text protocol it's done in
  prepare_schema_table(), but in --ps it is called on prepare only,
  so nothing was discarding profiling during execute.
* move the permission checking code for SHOW CREATE VIEW to
  mysqld_show_create_get_fields(), so that it would be called during
  prepare step too.
* only set sel_result when it was created here and needs to be
  destroyed in the same block. Avoid destroying lex->result.
* use the correct number of tables in check_show_access(). Saying
  "as many as possible" doesn't work when first_not_own_table isn't
  set yet.

sql_prepare.cc:
* use correct user name for SHOW GRANTS columns
* don't ignore verbose flag for SHOW SLAVE STATUS
* support preparing REVOKE ALL and ROLLBACK TO SAVEPOINT
* don't ignore errors from thd->prepare_explain_fields()
* use select_send result for sending ANALYZE and EXPLAIN, but don't
  overwrite lex->result, because it might be needed to issue execute-time
  errors (select_dumpvar - too many rows)

sql_show.cc:
* check grants for SHOW CREATE VIEW here, not in mysql_execute_command

sql_view.cc:
* use the correct function to check privileges. Old code was doing
  check_access() for thd->security_ctx, which is invoker's sctx,
  not definer's sctx. Hide various view related errors from the invoker.

sql_yacc.yy:
* initialize lex->select_lex for LOAD, otherwise it'll contain garbage
  data that happen to fail tests with views in --ps (but not otherwise).

This solves the following issues:

* unlike lex->m_sql_cmd and lex->sql_command, thd->query_plan_flags
  is not reset in Prepared_statement::execute, it survives
  till the log_slow_statement(), so slow logging behaves correctly in --ps

* using thd->query_plan_flags for both slow_log_filter and
  log_slow_admin_statements means the definition of "admin" statements
  for the slow log is the same no matter how it is filtered out.

Added metadate info after prepare EXPLAIN/ANALYZE.

Prepare os ANALYZE now respond as EXPLAIN.

row_merge_create_fts_sort_index(): Initialize dict_col_t in
an unambiguous way. GCC 6 and later appear to be able to optimize
away the memset() that is part of mem_heap_zalloc() in the
placement new call. Let us avoid using placement new in order
to ensure that the objects will actually be initialized.

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71388

https://gcc.gnu.org/ml/gcc/2016-02/msg00207.html

While the latter reference hints that the optimization is only
applicable to non-POD types (and dict_col_t does not define
any member functions before 10.2), it is most consistent to
use the same initialization across all versions.

ALTER TABLE ... ADD FOREIGN KEY may trigger assertion failure when
it has LOCK=EXCLUSIVE clause or concurrent FLUSH TABLES is being
executed.

In both cases being altered table is marked as flushed, which forces
subsequent attempt to open parent table to re-open. Which in turn is
not allowed while transaction is running.

Rather than opening parent table, just take appropriate MDL lock.

Also removed table_already_fk_prelocked() check: MDL itself has much
better methods to handle duplicate locks. E.g. the former won't acquire
MDL_SHARED_NO_WRITE if it already has MDL_SHARED_READ.

This can be useful:
UBSAN_OPTIONS=log_path=/some/path

clang users may want to increase stack size in include/my_pthread.h or enable
some optimizations

Clean up after commit 0957d257
which introduced some disorder (unsorted or duplicated test names).

modified:   suite/galera/t/galera_autoinc_sst_mariabackup.test

remove WSREP_SST_OPT_SUFFIX_VALUE, checking [mysqld] is covered in the parse_cnf --mysqld case also in wsrep_sst_xtrabackup-v2

wsrep_sst: remove WSREP_SST_OPT_SUFFIX_VALUE, checking [mysqld] is covered in the parse_cnf --mysqld case

modified:   mysql-test/suite/galera/disabled.def
modified:   mysql-test/suite/galera/r/galera_many_rows.result
modified:   mysql-test/suite/galera/t/galera_kill_nochanges.test
new file:   mysql-test/suite/galera/t/galera_many_rows.cnf
modified:   mysql-test/suite/galera/t/galera_many_rows.test
modified:   mysql-test/suite/galera/t/galera_var_dirty_reads.test
modified:   scripts/wsrep_sst_mariabackup.sh

galera.partition and galera.galera_binlog_stmt_autoinc regularly display
mismatching values for AUTO_INCREMENT columns.

galera.MW-336 often times out while waiting for something in PROCESSLIST.

Also, sort the test names, remove the redundant "galera." prefix and
fix typos in 2 test names.

- Fetch innodb_compression_level from the running server.Add the value
of innodb_compression_level in backup-my.cnf file during backup phase.
So that prepare can use the innodb_compression_level variable from
backup-my.cnf

row_merge_create_index_graph(): Relay the internal state
from dict_create_index_step(). Our caller should free the index
only if it was not copied, added to the cache, and freed.

row_merge_create_index(): Free the index template if it was
not added to the cache. This is a safer variant of the logic
that was introduced in 65070bef in 10.2.

prepare_inplace_alter_table_dict(): Add additional fault injection
to exercise a code path where we have already added an index
to the cache.

row_mysql_handle_errors(): Correct the wrong error handling for
the code DB_FOREIGN_EXCEED_MAX_CASCADE that was introduced in

https://github.com/mysql/mysql-server/commit/c0923d396aef46799883390e9dcf7bbf173e4a03

    commit 35f5429e
    Author: Jimmy Yang <jimmy.yang@oracle.com>
    Date:   Wed Oct 6 06:55:34 2010 -0700

        Manual port Bug #Bug #54582 "stack overflow when opening many tables
        linked with foreign keys at once" from mysql-5.1-security to
        mysql-5.5-security again.

        rb://391 approved by Heikki

No known test case exists for repeating the bug before MariaDB 10.2.
The scenario should be that DB_FOREIGN_EXCEED_MAX_CASCADE is returned,
then InnoDB wrongly skips the rollback to the start of the current
row operation, and finally the SQL layer commits the transaction.
Normally the SQL layer would roll back either the entire transaction or
to the start of the statement. In the faulty scenario, InnoDB would
leave the transaction in an inconsistent state, and the SQL layer could
commit the transaction.

fixes these test failures in ASAN builds (in 10.1 and 10.4):
* main.signal_demo3
* main.sp
* sys_vars.max_sp_recursion_depth_func
* mroonga/storage.foreign_key_delete_existent
* mroonga/storage.foreign_key_delete_nonexistent
* mroonga/storage.foreign_key_insert_existent
* mroonga/storage.foreign_key_update_existent
* mroonga/storage.foreign_key_update_nonexistent
* mroonga/storage.function_command_auto-escape
* mroonga/storage.function_command_select
* mroonga/storage.variable_enable_operations_recording_insert

thd->lex->m_sql_cmd was not cleared between queries.
log_slow_query() could crash (when running mtr --ps) because of this.

I know no test case for this bug in 10.1. So a test case will be
committed separately in 10.2

fts_reset_get_doc(): properly initialize fts_get_doc_t::cache

fts_fetch_index_words(): Restore the initialization len=0.
The test innodb_fts.create in 10.2 would end up in an infinite loop
if this assignment is removed, because a following iteration of the
while() loop would assign zip->zp->avail_in=len with the original value
instead of the 0 that was reset in the previous iteration.

Fix the warnings issued by GCC 8 -Wstringop-truncation
and -Wstringop-overflow in InnoDB and XtraDB.

This work is motivated by Jan Lindström. The patch mainly differs
from his original one as follows:

(1) We remove explicit initialization of stack-allocated string buffers.
The minimum amount of initialization that is needed is a terminating
NUL character.
(2) GCC issues a warning for invoking strncpy(dest, src, sizeof dest)
because if strlen(src) >= sizeof dest, there would be no terminating
NUL byte in dest. We avoid this problem by invoking strncpy() with
a limit that is 1 less than the buffer size, and by always writing
NUL to the last byte of the buffer.
(3) We replace strncpy() with memcpy() or strcpy() in those cases
when the result is functionally equivalent.

Note: fts_fetch_index_words() never deals with len==UNIV_SQL_NULL.
This was enforced by an assertion that limits the maximum length
to FTS_MAX_WORD_LEN. Also, the encoding that InnoDB uses for
the compressed fulltext index is not byte-order agnostic, that is,
InnoDB data files that use FULLTEXT INDEX are not portable between
big-endian and little-endian systems.

row_merge_create_fts_sort_index(): Initialize dict_col_t.

This fixes an access to uninitialized dict_col_t::ind when a debug
assertion in MariaDB 10.4 invokes is_dropped() in
rec_get_converted_size_comp_prefix_low(). Older MariaDB versions
seem to be unaffected by the uninitialized values, but it should
not hurt to initialize everything.

Only starting with MariaDB 10.3.8 (MDEV-16365), InnoDB can actually
handle ALTER IGNORE TABLE correctly when introducing a NOT NULL
attribute to a column that contains a NULL value. Between
MariaDB Server 10.0 and 10.2, we would incorrectly return an error
for ALTER IGNORE TABLE when the column contains a NULL value.

The problem with the InnoDB table attribute encryption_key_id is that it is
not being persisted anywhere in InnoDB except if the table attribute
encryption is specified and is something else than encryption=default.
MDEV-17320 made it a hard error if encryption_key_id is specified to be
anything else than 1 in that case.

Ideally, we would always persist encryption_key_id in InnoDB. But, then we
would have to be prepared for the case that when encryption is being enabled
for a table whose encryption_key_id attribute refers to a non-existing key.

In MariaDB Server 10.1, our best option remains to not store anything
inside InnoDB. But, instead of returning the error that MDEV-17320
introduced, we should merely issue a warning that the specified
encryption_key_id is going to be ignored if encryption=default.

To improve the situation a little more, we will issue a warning if
SET [GLOBAL|SESSION] innodb_default_encryption_key_id is being set
to something that does not refer to an available encryption key.

Starting with MariaDB Server 10.2, thanks to MDEV-5800, we could open the
table definition from InnoDB side when the encryption is being enabled,
and actually fix the root cause of what was reported in MDEV-17320.

1. Always drop merged_for_insert flag on cleanup (there could be errors which prevent TABLE to be assigned)
2. Make more precise cleanup of select parts which was touched

st_select_lex::handle_derived() and mysql_handle_list_of_derived() had
exactly the same implementations.

- Adding a new method LEX::handle_list_of_derived() instead
- Removing public function mysql_handle_list_of_derived()
- Reusing LEX::handle_list_of_derived() in st_select_lex::handle_derived()