Author: Marko Mäkelä
Author Date: 2018-12-13 19:25:12 UTC

MDEV-17989 InnoDB: Failing assertion: dict_tf2_is_valid(flags, flags2)

With innodb_default_row_format=redundant, InnoDB would crash when
using table options that are incompatible with ROW_FORMAT=REDUNDANT.

create_table_info_t::m_default_row_format: Cache the value of
innodb_default_row_format.

create_table_info_t::check_table_options(): Validate ROW_TYPE_DEFAULT
with m_default_row_format.

create_table_info_t::innobase_table_flags(): Use the
cached m_default_row_format.

create_table_info_t: Never read m_form->s->row_type.
Use m_create_info->row_type instead.

dict_tf_set(): Never set invalid flags for ROW_FORMAT=REDUNDANT.

ha_innobase::truncate(): Set info.row_type based on the ROW_FORMAT
of the current table.

Author: Sergei Golubchik
Author Date: 2018-12-13 18:51:40 UTC

remove unsed variable

Author: Sergey Vojtovich
Author Date: 2018-12-12 15:58:20 UTC

MDEV-16987 - ALTER DATABASE possible in read-only mode

Forbid ALTER DATABASE under read_only.

Author: Oleksandr Byelkin
Author Date: 2018-12-13 16:06:26 UTC

MDEV-16278: Missing DELETE operation in COM_STMT_BULK_STMT

Allow array binding for DELETE, test it.

Author: Oleksandr Byelkin
Author Date: 2018-12-13 14:28:18 UTC

MDEV-16240: Assertion `0' failed in row_sel_convert_mysql_key_to_innobase

Set table in row ID position mode before using this function.

Author: Marko Mäkelä
Author Date: 2018-12-13 11:37:21 UTC

Merge 10.0 into 10.1

Also, apply the MDEV-17957 changes to encrypted page checksums,
and remove error message output from the checksum function,
because these messages would be useless noise when mariabackup
is retrying reads of corrupted-looking pages, and not that
useful during normal server operation either.

The error messages in fil_space_verify_crypt_checksum()
should be refactored separately.

Author: Alexander Barkov
Author Date: 2018-12-13 11:09:03 UTC

MDEV-17995 INET6_NTOA(ucs2_input) erroneously returns NULL

Author: Daniele Sciascia
Author Date: 2018-12-11 13:26:42 UTC

Record test galera_ist_mariabackup_innodb_flush_logs

Author: Thirunarayanan Balathandayuthapani
Author Date: 2018-12-13 06:50:14 UTC

commit 37ad3e55c6281095b9e867a2b5d5076e0a963e89
Author: Thirunarayanan Balathandayuthapani <thiru@mariadb.com>
Date: Tue Dec 11 18:51:47 2018 +0530

    MDEV-17957 Make Innodb_checksum_algorithm stricter for strict_* values

- Pushed the fix for innodb_zip.innochecksum test case failure

Author: Varun Gupta
Author Date: 2018-12-12 18:27:38 UTC

fixing compile error

Author: Kentoku SHIBA
Author Date: 2018-12-12 18:13:31 UTC

MDEV-16787 optimistic parallel replication fails on spider
Add a system variable spider_slave_trx_isolation.
- spider_slave_trx_isolation
  The transaction isolation level when Spider table is used by slave SQL thread.
  -1 : OFF
   0 : READ UNCOMMITTED
   1 : READ COMMITTED
   2 : REPEATABLE READ
   3 : SERIALIZABLE
  The default value is -1

Miscellaneous Spider typos

Author: Kentoku SHIBA
Author Date: 2018-12-12 18:11:57 UTC

MDEV-16787 optimistic parallel replication fails on spider
Add a system variable spider_slave_trx_isolation.
- spider_slave_trx_isolation
  The transaction isolation level when Spider table is used by slave SQL thread.
  -1 : OFF
   0 : READ UNCOMMITTED
   1 : READ COMMITTED
   2 : REPEATABLE READ
   3 : SERIALIZABLE
  The default value is -1

Miscellaneous Spider typos

Author: midenok
Author Date: 2018-11-27 19:25:11 UTC

Replace dict_instant_t::non_pk_col_map with field_map

Author: Robert Bindar
Author Date: 2018-12-03 15:05:25 UTC

MDEV-13095 Implement User Account locking, fix embedded tests, rebase

Author: Vladislav Vaintroub
Author Date: 2018-12-11 17:23:54 UTC

MDEV-16266 - New command FLUSH SSL to reload server's SSL certificate(private key,CRL,etc)

Author: Marko Mäkelä
Author Date: 2018-12-12 11:46:06 UTC

Merge 10.2 into 10.3

Author: Sergei Golubchik
Author Date: 2018-11-24 13:13:41 UTC

MDEV-17658 change the structure of mysql.user table

Implement User_table_json.
Fix scripts to use mysql.global_priv.
Fix tests.

Author: Thirunarayanan Balathandayuthapani
Author Date: 2018-12-11 15:12:06 UTC

MDEV-17958 Keep big endian variant of innodb_checksum_algorithm=crc32 only in big-endian machine

Problem:
========
In MySQL 5.7, it was noticed that files are not portable between
big-endian and little-endian systems (such as SPARC and x86),
because the original implementation of innodb_checksum_algorithm=crc32
was not byte order agnostic. Basically, it weakened the
innodb_checksum_algorithm for crc32 algorithm for mariadb-10.2

Fix:
====
Keep the bug-compatible variant for crc32 algorithm only present
on big-endian systems (#ifdef WORDS_BIGENDIAN).

Author: Thirunarayanan Balathandayuthapani
Author Date: 2018-12-11 10:52:12 UTC

MDEV-17957 Make Innodb_checksum_algorithm stricter for strict_* values

Problem:

 Innodb_checksum_algorithm checks for all checksum algorithm to
validate the page checksum even though the algorithm is specified as
strict_crc32, strict_innodb, strict_none.

Fix:

  Remove the checks for all checksum algorithm to validate the page
cehcksum if the algo is specified as strict_* values.

Author: Thirunarayanan Balathandayuthapani
Author Date: 2018-12-11 10:46:11 UTC

MDEV-17957 Make Innodb_checksum_algorithm stricter for strict_* values

Problem:

  Innodb_checksum_algorithm checks for all checksum algorithm to
validate the page checksum even though the algorithm is specified as
strict_crc32, strict_innodb, strict_none.

Fix:

   Remove the checks for all checksum algorithm to validate the page
checksum if the algo is specified as strict_* values.

Author: Sergei Golubchik
Author Date: 2018-12-11 10:38:30 UTC

debian install/upgrade fixes

1. don't run full mysql_upgrade on every server restart,
   use --version-check to do it only once

2. fix syslog tag name in the postinst script, don't pretend
   mysqld_safe generated all these messages. Auto-detect the version
   to simplify maintenance

Author: Vladislav Vaintroub
Author Date: 2018-12-11 00:22:54 UTC

generate certificate to test FLUSH SSL

Author: Sergei Golubchik
Author Date: 2018-12-10 18:04:07 UTC

another attempt to fix Debian upgrades

Author: Sergei Golubchik
Author Date: 2018-12-10 13:53:50 UTC

debian: don't "start" mysqld after it was just restarted

mariadb-server-10.2.postinst has "invoke start" at the end,
unconditionally added by a debhelper. We don't want to get there
after a triggered restart.

Author: Varun Gupta
Author Date: 2018-12-09 19:04:41 UTC

MDEV-17778: Alter table leads to a truncation warning with ANALYZE command

Alter statement changed the THD structure by setting the value to FIELD_CHECK_WARN
and then not resetting it back. This led ANALYZE to throw a warning which previously
it didn't.

Author: Sergey Vojtovich
Author Date: 2018-12-06 15:23:24 UTC

Moved early check for table existance to mysql_execute_command()

MDEV-17772 - 3 way lock : ALTER, MDL, BACKUP STAGE BLOCK_DDL

While waiting for a (potentially long) RO transaction or SELECT, DDL and
LOCK TABLES ... WRITE hold protection against FTWRL and BACKUP STAGE.

This effectively makes FTWRL/BACKUP STAGE indirectly wait for this RO
transaction or SELECT to finish. Which is not great, as otherwise we
could do something useful meanwhile.

With this patch BACKUP lock is attempted to be acquired after TABLE/SCHEMA
locks. If this attempt fails, TABLE/SCHEMA locks gets released and we
start waiting for BACKUP lock. When wait finishes, BACKUP lock is released
(to avoid deadlocks) and we attempt to acquire all locks once again.

Other changes:
- Take MDL lock before testing if table exists as part of
  CREATE TABLE ... IF EXISTS. This change was an effect of changes in
  lock_table_name and removes an inconsistency where one could get
  different error messages from CREATE TABLE .. IF EXISTS depending on
  active mdl locks.
  One effect of this change is that we don't binary log CREATE TABLE IF
  EXISTS if the table exists. This was done because old code was sometimes
  behaving inconsistenly (it was logged some time and not other times)
  and sending the query to the slave could make the slave even more
  inconsistent as there is not guarantee that the new table will have
  the same definition as the old table on the master.

Author: Sergey Vojtovich
Author Date: 2018-12-09 10:29:32 UTC

Fixed false positive from close_cached_connection_tables

Author: Varun Gupta
Author Date: 2018-12-09 07:55:27 UTC

MDEV-17255: New optimizer defaults and ANALYZE TABLE

Added to new values to the server variable use_stat_tables.
The values are COMPLEMENTARY_FOR_QUERIES and PREFERABLY_FOR_QUERIES.
Both these values don't allow to collect EITS for queries like
    analyze table t1;
To collect EITS we would need to use the syntax with persistent like
   analyze table t1 persistent for columns (col1,col2...) index (idx1, idx2...) / ALL

Changing the default value from NEVER to PREFERABLY_FOR_QUERIES.

Author: varun
Author Date: 2018-05-31 10:21:59 UTC

MDEV-15253: Default optimizer setting changes for MariaDB 10.4

Author: Alexander Barkov
Author Date: 2018-12-08 12:10:44 UTC

MDEV-17928 Conversion from TIMESTAMP to VARCHAR SP variables does not work well on fractional digits

Author: Igor Babaev
Author Date: 2018-12-07 20:13:26 UTC

MDEV-16188: Fixed problems of the implementation for rowid filter pushdown
            in InnoDB engine.

Fixed some other problems at the SQL layer related to rowid filters.
After this patch all tests in the main test suite passed including
execution with --ps-protocol using both Debug and Release builds.

Author: Marko Mäkelä
Author Date: 2018-12-07 14:32:32 UTC

Merge 10.4 into HEAD

Author: Sergei Golubchik
Author Date: 2018-12-06 23:33:51 UTC

yet another test: Eight Queen problem in one recursive CTE

Author: Varun Gupta
Author Date: 2018-12-06 21:34:34 UTC

index merge test adjusted

Author: Varun Gupta
Author Date: 2018-12-06 20:42:22 UTC

MDEV-17032: Estimates are higher for partitions of a table with @@use_stat_tables= PREFERABLY

The problem here is EITS statistics does not calculate statistics for the partitions of the table.
So a temporary solution would be to not read EITS statistics for partitioned tables.

Also disabling reading of EITS for columns that participate in the partition list of a table.

Author: Kristian Nielsen
Author Date: 2018-10-14 18:41:49 UTC

Move deletion of old GTID rows to slave background thread

This patch changes how old rows in mysql.gtid_slave_pos* tables are deleted.
Instead of doing it as part of every replicated transaction in
record_gtid(), it is done periodically (every @@gtid_cleanup_batch_size
transaction) in the slave background thread.

This removes the deletion step from the replication process in SQL or worker
threads, which could speed up replication with many small transactions. It
also decreases contention on the global mutex LOCK_slave_state. And it
simplifies the logic, eg. when a replicated transaction fails after having
deleted old rows.

With this patch, the deletion of old GTID rows happens asynchroneously and
slightly non-deterministic. Thus the number of old rows in
mysql.gtid_slave_pos can temporarily exceed @@gtid_cleanup_batch_size. But
all old rows will be deleted eventually after sufficiently many new GTIDs
have been replicated.

Author: Sergei Golubchik
Author Date: 2018-12-06 00:17:44 UTC

MDEV-17898 FLUSH PRIVILEGES crashes server with segfault

merge_role_db_privileges() was remembering pointers into Dynamic_array
acl_dbs, and later was using them, while pushing more elements into the
array. But pushing can cause realloc, and it can invalidate all pointers.

Fix: remember and use indexes of elements, not pointers.

Author: Georg Richter
Author Date: 2018-12-05 16:36:34 UTC

Windows fixes

Author: Rasmus Johansson
Author Date: 2018-12-05 12:17:47 UTC

Merge pull request #995 from rasmushoj/10.4

MariaDB symlinks

Author: Alexander Barkov
Author Date: 2018-12-05 10:06:49 UTC

MDEV-17906 Class Binary_string

Author: Vladislav Vaintroub
Author Date: 2018-11-26 20:24:05 UTC

MDEV-15649 Speedup search in acl_users and acl_dbs array,
sorting them by usernames first, and then by get_sort() value.

Search functions now use binary search to find the the first entry with
given name. Then, linear search is done, until the first match.

Author: Sergey Vojtovich
Author Date: 2018-12-03 20:53:36 UTC

Enable main.connect-abstract

Author: Georg Richter
Author Date: 2018-12-03 12:55:20 UTC

MDEV-14101: Provide option to specify tls_version for client tools

Author: Vladislav Vaintroub
Author Date: 2018-12-01 22:54:15 UTC

MDEV-16266 : New command FLUSH SSL ro reload SSL acceptor context.

Reloads certificate, key, CA, CRL.

Author: Varun Gupta
Author Date: 2018-12-01 06:16:04 UTC

Adjusting the trace_info_function for ROR intersect

Author: Igor Babaev
Author Date: 2018-04-16 17:31:30 UTC

MDEV-15556 MariaDB crash with big_tables=1 and CTE

This bug manifested itself when the optimizer chose an execution plan with
an access of the recursive CTE in a recursive query by key and ARIA/MYISAM
temporary tables were used to store recursive tables.
The problem appeared due to passing an incorrect parameter to the call of
instantiate_tmp_table() in the function With_element::instantiate_tmp_tables().

Author: Sergei Golubchik
Author Date: 2018-11-28 18:19:16 UTC

mysqltest: use a dynamically growing command buffer

Author: Sachin Setiya
Author Date: 2018-11-27 13:24:22 UTC

Compile error solved

Author: Kristian Nielsen
Author Date: 2018-10-14 18:41:49 UTC

Move deletion of old GTID rows to slave background thread

This patch changes how old rows in mysql.gtid_slave_pos* tables are deleted.
Instead of doing it as part of every replicated transaction in
record_gtid(), it is done periodically (every @@gtid_cleanup_batch_size
transaction) in the slave background thread.

This removes the deletion step from the replication process in SQL or worker
threads, which could speed up replication with many small transactions. It
also decreases contention on the global mutex LOCK_slave_state. And it
simplifies the logic, eg. when a replicated transaction fails after having
deleted old rows.

With this patch, the deletion of old GTID rows happens asynchroneously and
slightly non-deterministic. Thus the number of old rows in
mysql.gtid_slave_pos can temporarily exceed @@gtid_cleanup_batch_size. But
all old rows will be deleted eventually after sufficiently many new GTIDs
have been replicated.

Author: Jan Lindström
Author Date: 2018-11-22 14:33:20 UTC

MDEV-17804: Galera tests cause mysql_socket.h:738: inline_mysql_socket_send: Assertion `mysql_socket.fd != -1' failed.

Do not do end of statement logic if thd is already killed as
socket is already closed.

Author: Thirunarayanan Balathandayuthapani
Author Date: 2018-11-20 12:26:16 UTC

MDEV-14398 When innodb_encryption_rotate_key_age=0 is set, server won't encrypt tablespaces

Problem:
========
 If innodb_encryption_rotate_key_age is zero then it disables key rotation
entirely. But it also prevents the server from encrypting unencrypted tablespaces
and also decrypting the encrypted tablespace.

Solution:
========
 Store the global encryption status of all tablespace in dictionary header
page. There can be 3 possible states for encryption status:
 i) ALL_ENCRYPTED
 ii) ALL_DECRYPTED
 iii) MIX_STATE.

Read the encryption status during startup and check whether both the
innodb_encrypt_tables and global encrypt status of all tablespace are
encrypted/decrypted if innodb_encryption_rotate_key_age is 0. In that case,
fil_crypt_thread can avoid traversing the space list.

Whenever newly created tablespace is created, change the global encrypt status
to MIXED STATE.

If innodb_encryption_rotate_key_age is 0 then it avoids the encryption of
already encrypted tablespace.

Author: Oleksandr Byelkin
Author Date: 2018-11-20 09:58:34 UTC

MDEV-15073: Generic UDAF parser code in server for windows functions

Added support for usual agreggate UDF (UDAF)
Added remove() call support for more efficient window function processing
Added example of aggregate UDF with efficient windows function support

Author: Jan Lindström
Author Date: 2018-11-20 05:49:46 UTC

Add missing .rdiff file to test galera_sst_xtrabackup-v2_data_dir
for debug build.

Author: Vladislav Vaintroub
Author Date: 2018-11-19 19:43:04 UTC

MDEV-17308 mariabackup to use the BACKUP statement instead of FTWRL

Author: Elena Stepanova
Author Date: 2018-11-18 15:38:48 UTC

MDEV-17507 Make MTR tests work for builds without Aria for temporary tables

Skip tests which expectedly fail when Aria is not used for temporary tables

Author: Elena Stepanova
Author Date: 2018-11-16 17:58:52 UTC

Updated list of unstable tests for 10.3.11 release

Author: Monty
Author Date: 2018-11-16 14:54:33 UTC

Max transid was not stored directly after Aria recovery

This caused ma_test_recovery.pl to fail

Other things:
- Fixed bug where "ma_test_recovert.pl --abort-on-error" didn't abort
  on error

Author: Jan Lindström
Author Date: 2018-11-16 08:21:11 UTC

MDEV-13881: galera.partition failed in buildbot with wrong result

Add proper wait conditions to second node before continuing test.

Author: Thirunarayanan Balathandayuthapani
Author Date: 2018-11-15 09:44:02 UTC

MDEV-16849 Extending indexed VARCHAR column should be instantaneous

Analysis:
========
Increasing the length of the indexed varchar column is not an instant operation for
innodb.

Fix:
===
- Introduce the new handler flag 'Alter_inplace_info::ALTER_COLUMN_INDEX_LENGTH' to
indicate the index length differs due to change of column length changes.

- InnoDB makes the ALTER_COLUMN_INDEX_LENGTH flag as instant operation.

This is a port of Mysql fix.

    commit 913071c0b16cc03e703308250d795bc381627e37
    Author: Nisha Gopalakrishnan <nisha.gopalakrishnan@oracle.com>
    Date: Wed May 30 14:54:46 2018 +0530

        BUG#26848813: INDEXED COLUMN CAN'T BE CHANGED FROM VARCHAR(15)
                      TO VARCHAR(40) INSTANTANEOUSLY

Author: varun
Author Date: 2018-11-14 07:34:25 UTC

MDEV-13170: Database service (MySQL) stops after update with trigger

For prepare statemtent/stored procedures we rollback the items to original
ones after prepare execution in the function reinit_stmt_before_use.
This rollback is done for group by, order by clauses but is not done for
the window specification containing the order by and partition by clause of
the window function.

Author: Oleksandr Byelkin
Author Date: 2018-11-14 19:32:17 UTC

Merge branch '10.2' into 10.3

Author: Jan Lindström
Author Date: 2018-11-12 10:05:42 UTC

MDEV-17229: Encryption threads ignore innodb_default_encryption_key_id

Background

  MariaDB identifies encryption keys using a positive 32-bit integer.
  This makes it easier to identify the key you want to use for particular
  tables, through the ENCRYPTION_KEY_ID table option. You can set the default
  key using the innodb_default_encryption_key_id system variable. If no
  value is set system default (1) is used.

Syntax
  SET [GLOBAL|SESSION] innodb-default-encryption-key-id=<key_id>;

Used default encryption key_id can't be set to value that does not exists
in the encryption plugin. e.g.

   SET GLOBAL innodb_default_encryption_key_id = 999;
   ERROR 42000: Variable 'innodb_default_encryption_key_id' can't be set to the value of '999'
   SHOW WARNINGS;
   Level Code Message
   Warning 1210 innodb_default_encryption_key_id=999 not available in the encryption plugin
   Error 1231 Variable 'innodb_default_encryption_key_id' can't be set to the value of '999'

If variable is set on configuration file (my.cnf or similar) server will refuse to start if used
key_id is not found in the encryption plugin. This can be seen on server error log e.g.

   [ERROR] innodb_default_encryption_key_id=999 is unavailable in the encryption plugin
   [ERROR] Plugin 'InnoDB' init function returned error.
   [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.

If user creates a table that is not encrypted e.g:
   create table t1(a int) encrypted=no engine=innodb;

and default encryption key_id is not same as system default (1), a warning
is issued and system default is used.

   SET GLOBAL innodb-default-encryption-key-id=1;
   CREATE TABLE t1 (a int) ENGINE=INNODB ENCRYPTED=NO ENCRYPTION_KEY_ID=10;
   Warnings:
   Warning 140 InnoDB: Ignored ENCRYPTION_KEY_ID=10 when encryption is disabled

Currently set default encryption key_id does not effect tables that
are created using specific encryption key_id e.g.
    create table t1(a int) engine=innodb encrypted=yes encryption_key_id=4;

Instead specified key_id (in above case 4) is used. Tables using encryption but no
specific encryption key_id e.g.
    create table t1(a int) engine=innodb encrypted=yes;

will use currently set default encryption key_id for that session or
if no key_id defined the set global encryption key_id or system default.

Tables using default table options e.g
    create table t1(a int) engine=innodb;

will use currently set global encryption key_id when encryption threads
does key rotation from unencrypted to encrypted state.

Changes

innodb-checksum-algorithm
innodb-encryption-alter
innodb-redo-badkey
innodb-redo-nokeys
 These test require small changes as now we do not allow default key_id
 that is not found from encryption plugin and tables using no encryption
 produce warning when nondefault encryption key_id is used.

fil_crypt_start_encrypting_space()
 Use global default encryption key_id when creating encryption meta data.

innodb_default_encryption_key_id_validate()
 New function to validate that used key_id is found from encryption plugin.

innodb_default_encryption_key_id()
 New accessessor function for default encryption key_id.

innobase_init()
 Do not allow server startup using encryption key_id that is not found
 from encryption plugin.

Author: Alexander Barkov
Author Date: 2018-11-13 14:59:38 UTC

MDEV-17278 CURSOR FOR LOOP - ERROR: unexpected end of stream, read 0 bytes (SERVER CRASH)

sp_instr_cursor_copy_struct::exec_core() created TYPELIBs on a wrong mem_root,
the one which is initialized in sp_head::execute(), this code:

  /* init per-instruction memroot */
  init_sql_alloc(&execute_mem_root, "per_instruction_memroot",
                 MEM_ROOT_BLOCK_SIZE, 0, MYF(0));

This memory root cleans up after every sp_instr_xxx executed, so later
sp_instr_cfetch::execute() tried to use already freed and trashed memory.

Changing sp_instr_cursor_copy_struct::exec_core() to call tmp.export_structure()
inside this block (not outside of it):
  thd->set_n_backup_active_arena(thd->spcont->callers_arena, &current_arena);
  ...
  thd->restore_active_arena(thd->spcont->callers_arena, &current_arena);

So now TYPELIBs created by sp_instr_cursor_copy_struct::exec_core() are
still available and valid when sp_instr_cfetch::execute() is called.
They are freed at the end of dispatch_command() corresponding to
the "CALL p1" statement.

Author: Oleksandr Byelkin
Author Date: 2018-11-13 09:10:09 UTC

MDEV-17124: mariadb 10.1.34, views and prepared statements: ERROR 1615 (HY000): Prepared statement needs to be re-prepared

The problem is that if table definition cache (TDC) is full of real tables which are in tables cache, view definition can not stay there so will me removed by its own underlying tables.
In situation above old mechanism of detection matching definition in PS and current version always require reprepare and so prevent executing the PS.
One work arount is to increase TDC, other - improve version check for views (which is done here).
Now in suspiciouse cases we check MD5 of the view to be sure that version really have chenged.

Author: Monty
Author Date: 2018-11-12 23:34:37 UTC

Added syntax and implementation for BACKUP STAGE's

Part of MDEV-5336 Implement LOCK FOR BACKUP

Author: Sergei Golubchik
Author Date: 2018-11-06 20:18:49 UTC

update results after CONC-351 fix

Author: Sergei Golubchik
Author Date: 2018-10-30 18:04:54 UTC

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

mysqltest.cc:
* make sure run_query_stmt() doesn't crash if there's
  no active connection (in wait_until_connected_again.inc)

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 the correct metadata when creating a column for ANALYZE SELECT.
  The actual value is always FLOAT(4,2), but the column was created
  as FLOAT(4,10), so --ps and text protocols were using different number
  of decimals.

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()

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

sql_tvc.cc:
* allocate Type_holder and items in the correct arena

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.

Author: Sergei Golubchik
Author Date: 2018-11-06 10:37:15 UTC

C/C 3.0.7

Author: Andrei Elkin
Author Date: 2018-11-07 13:24:30 UTC

MDEV-14528 followup.

There was a failure in rpl_delayed_slave after recent MDEV-14528 commit.
The parallel applier should not set its
Relay_log::last_master_timestamp from Format-descriptor log event.
The latter may reflect a deep past so Seconds-behind-master will be
computed through it all long time while the first group of events is
executed.
The main MDEV-14528 is refined, rpl_delayed_slave now passes also
in the parallel mode.

Author: varun
Author Date: 2018-05-31 10:21:59 UTC

MDEV-15253: Default optimizer setting changes for MariaDB 10.4

Author: Vladislav Vaintroub
Author Date: 2018-11-06 08:29:41 UTC

MDEV-14781 - threadpool slowdown with slow ssl handshake.

Fix is not to use blocking socket IO during SSL handshake.
With non-blocking socket IO, threadpool is able to utilize the wait
notification callbacks, that vio_io_wait() is calling whenever
socket would block.

Author: Alexey Botchkov
Author Date: 2018-11-02 22:18:27 UTC

MDEV-14567

Author: Alexey Botchkov
Author Date: 2018-11-02 20:01:50 UTC

MDEV-15890 X2

Author: Jan Lindström
Author Date: 2018-11-02 06:46:04 UTC

Merge remote-tracking branch 'origin/10.0' into 10.0-galera

Author: varun
Author Date: 2018-11-01 08:56:13 UTC

MDEV-12575: Server crash in AGGR_OP::put_record or in JOIN_CACHE::free or Invalid write in JOIN::make_aggr_tables_info

During the optimize state of a query, we come know that the result set
would atmost contain one row, then for such a query we don't need
to compute GROUP BY, ORDER BY and DISTINCT.

Author: Sergey Petrunia
Author Date: 2018-11-01 08:35:28 UTC

MDEV-17518: Range optimization doesn't use ON expressions from nested outer joins

Part#2: take into account that join nest that we are marking as constant
might already have constant tables in it. Don't count these tables twice.

Author: Vicențiu Ciorbaru
Author Date: 2018-11-01 08:19:20 UTC

update win result further

Author: Jan Lindström
Author Date: 2018-10-31 16:28:33 UTC

Fix merge error on tests.

Author: Daniel Bartholomew
Author Date: 2018-10-31 14:07:09 UTC

bump the VERSION

Author: Sergei Golubchik
Author Date: 2018-10-30 23:25:26 UTC

Merge branch '5.5' into 10.0

Author: Alexey Botchkov
Author Date: 2018-10-30 14:15:41 UTC

MDEV-17256 Decimal field multiplication bug.

We should clear trailing zeroes in frac part. Otherwise
that tail is growing quickly and forces unnecessary truncating of
arguments.

Author: Alexey Botchkov
Author Date: 2018-10-30 14:15:41 UTC

MDEV-17256 Decimal field multiplication bug.

We should clear trailing zeroes in frac part. Otherwise
that tail is growing quickly and forces unnecessary truncating of
arguments.

Author: Jan Lindström
Author Date: 2018-10-29 16:45:19 UTC

Merge tag 'mariadb-5.5.62' into 5.5-galera

Author: Sergey Vojtovich
Author Date: 2018-10-25 15:13:05 UTC

MDEV-17441 - InnoDB transition to C++11 atomics

ibuf_t::n_merges transition to simple_atomic_counter.

Author: Andrei Elkin
Author Date: 2018-07-27 19:55:18 UTC

MDEV-10963 Fragmented BINLOG query

The problem was originally stated in
  http://bugs.mysql.com/bug.php?id=82212
The size of an base64-encoded Rows_log_event exceeds its
vanilla byte representation in 4/3 times.
When a binlogged event size is about 1GB mysqlbinlog generates
a BINLOG query that can't be send out due to its size.

It is fixed with fragmenting the BINLOG argument C-string into
(approximate) halves when the base64 encoded event is over 1GB size.
The mysqlbinlog in such case puts out

    SET @binlog_fragment_0='base64-encoded-fragment_0';
    SET @binlog_fragment_1='base64-encoded-fragment_1';
    BINLOG @binlog_fragment_0, @binlog_fragment_1;

to represent a big BINLOG.
For prompt memory release BINLOG handler is made to reset the BINLOG argument
user variables in the middle of processing, as if @binlog_fragment_{0,1} = NULL
is assigned.

Notice the 2 fragments are enough, though the client and server still may
need to tweak their @@max_allowed_packet to satisfy to the fragment
size (which they would have to do anyway with greater number of
fragments, should that be desired).

On the lower level the following changes are made:

Log_event::print_base64()
  remains to call encoder and store the encoded data into a cache but
  now *without* doing any formatting. The latter is left for time
  when the cache is copied to an output file (e.g mysqlbinlog output).
  No formatting behavior is also reflected by the change in the meaning
  of the last argument which specifies whether to cache the encoded data.

Rows_log_event::print_helper()
  is made to invoke a specialized fragmented cache-to-file copying function
  which is

copy_cache_to_file_wrapped()
  that takes care of fragmenting also optionally wraps encoded
  strings (fragments) into SQL stanzas.

my_b_copy_to_file()
  is generalized to accepts more a limit argument to constraint the copying.
  The limit does not do any effect on the fully read cache.

Author: Marko Mäkelä
Author Date: 2018-10-19 16:00:25 UTC

Merge 10.4 into 10.4-mdev-15562

This branch exists just for the record, for preserving the
development history, should it ever be needed.

Author: Sergey Vojtovich
Author Date: 2018-10-18 18:48:28 UTC

MDEV-14815 - Server crash or AddressSanitizer errors or valgrind warnings
             in thr_lock / has_old_lock upon FLUSH TABLES

Explicit partition access under LOCK TABLES may cause subsequent
statements to crashes the server, deadlock, trigger valgrind warnings or
ASAN errors. Freed memory was being used due to incorrect cleanup.

Some unused code removed.

Author: Larysa Sherepa
Author Date: 2018-10-18 16:46:00 UTC

CTE tests

Author: Dyre Tjeldvoll
Author Date: 2015-04-30 10:56:33 UTC

BUG#19988193: ASSERTION `(*TABLES)->REGINFO.LOCK_TYPE >= TL_READ'
FAILED IN LOCK_EXTERNAL
BUG#21198646: ASSERTION FAILED: (*TABLES)->REGINFO.LOCK_TYPE >=
TL_READ FILE LOCK.CC, LINE 356

This patch addresses two related issues: Calling a procedure which
creates a view from a trigger (BUG#19988193), and creating a function
calling a procedure doing RENAME TABLE (BUG#21198646), could both, in
certain circumstances, trigger an assert.

Root cause was that prelocking of tables with lock_type==TL_IGNORE is
not supported, and so triggers an assert. TL_IGNORE is only used for
source tables in CREATE VIEW statements and the table of a RENAME TABLE
statement. It is very unusual for these statements to be part of
prelocking analysis, as both are implicit commit statements which are
not permitted in triggers and stored functions/procedures. But as the test
cases show; it is possible to have such statements contribute to the
prelocking set, but in both cases the statement is "meaningless", in
the sense that it will trigger an error during execution.

Fix: In mysql_make_view(), avoid adding the backing tables to view_ref
if view_ref->prelocking_placeholder==true and lock_type==TL_IGNORE. In
sp_head::add_used_tables_to_table_list() skip SP_TABLES which have
lock_type=TL_IGNORE.

Test: New test cases added to tablelock.test

Author: Sachin Setiya
Author Date: 2018-10-14 17:46:53 UTC

MDEV-15919 lower_case_table_names does not behave as expected(nor...
 consistently) on Replication Slave

lower_case_table_names 0 -> 1 replication works, it's safe as long as
 mixed case names mapping to the lower case ones is one-to-one

Author: Thirunarayanan Balathandayuthapani
Author Date: 2018-10-16 06:08:11 UTC

MDEV-12547: InnoDB FULLTEXT index has too strict innodb_ft_result_cache_limit max limit

- Removed f_n_char from fts_string_t. Instead of that, InnoDB calculate
the number of chars when it is needed.

Author: Vladislav Vaintroub
Author Date: 2018-10-15 10:06:00 UTC

MDEV-14429 sql_safe_updates in my.cnf not work

Author: Vicențiu Ciorbaru
Author Date: 2018-10-14 17:54:52 UTC

Merge branch 'MDEV1732_check_constraint_10.2_v1' of https://github.com/an3l/server into an3l-MDEV1732_check_constraint_10.2_v1

Author: Sachin Setiya
Author Date: 2018-10-13 19:12:24 UTC

MDEV-17119 replicate_rewrite_db does not work for single char database name

Fixed issue in logic.

Author: Oleksandr Byelkin
Author Date: 2018-10-12 18:47:22 UTC

MDEV-16207: Sig 11 on RHEL7 on Azure

The problem was that Item_subselect::const_item was not checked correctly and can differ from used_tables result.
Which lead to the situation when Item_direct_view_ref pointing on subselect was constant but the subquery return
FALSE on const_item() request (which was an error).

Now both methods report the same.

Author: Vladislav Vaintroub
Author Date: 2018-10-11 18:36:59 UTC

fix test suite after MDEV-15438

Author: Andrei Elkin
Author Date: 2018-10-10 14:36:09 UTC

MDEV-14431 binlog.binlog_flush_binlogs_delete_domain failed in buildbot

The test and also rpl_gtid_delete_domain failed on PPC64 platform
due to an incorrectly specified actual key for searching
in a gtid domain system hash. While the correct size is 32 bits
the supplied value was 8 bytes of long int size on the platform.
The problem became evident thanks to the big endiness which
cut off the *least* significant part of the value field.

Fixed with correcting a dynamic array initialization to hold
now uint32 values as well as the values extraction for
searching in the gtid domain system hash.
A new added test ensures no overflowed values are accepted
for deletion which prevents inadvertent action. Notice though

MariaDB [test]> set @@session.gtid_domain_id=(1 << 32) + 1;
MariaDB [test]> show warnings;
+---------+------+--------------------------------------------------------+
| Level | Code | Message |
+---------+------+--------------------------------------------------------+
| Warning | 1292 | Truncated incorrect gtid_domain_id value: '4294967297' |
+---------+------+--------------------------------------------------------+
MariaDB [test]> select @@session.gtid_domain_id;
+--------------------------+
| @@session.gtid_domain_id |
+--------------------------+
| 4294967295 |
+--------------------------+

Author: Jan Lindström
Author Date: 2018-10-10 15:25:53 UTC

MDEV-17230: encryption_key_id from alter is ignored by encryption threads

Background: Used encryption key_id is stored to encryption metadata
i.e. crypt_data that is stored on page 0 of the tablespace of the
table. crypt_data is created only if implicit encryption/not encryption
is requested i.e. ENCRYPTED=[YES|NO] table option is used
fil_create_new_single_table_tablespace on fil0fil.cc.

Later if encryption is enabled all tables that use default encryption
mode (i.e. no encryption table option is set) are encrypted with
default encryption key_id that is 1. See fil_crypt_start_encrypting_space on
fil0crypt.cc.

ha_innobase::check_table_options()
 If default encryption is used and encryption is disabled, you may
        not use nondefault encryption_key_id as it is not stored anywhere.

Author: Jan Lindström
Author Date: 2018-10-10 15:12:27 UTC

MDEV-17229: Encryption threads ignore innodb_default_encryption_key_id

Background: Used encryption key_id is stored to encryption metadata
i.e. crypt_data that is stored on page 0 of the tablespace of the
table. crypt_data is created only if implicit encryption/not encryption
is requested i.e. ENCRYPTED=[YES|NO] table option is used
fil_create_new_single_table_tablespace on fil0fil.cc.

innodb_default_encryption_key_id setting should effect only tables created
with ENCRYPTED=YES table option. Similarly ENCRYPTION_KEY_ID table option
should be allowed only for tables using ENCRYPTED=YES table option.

Later if encryption is enabled all tables that use default encryption
mode (i.e. no encryption table option is set) are encrypted with
default encryption key_id that is 1. See fil_crypt_start_encrypting_space on
fil0crypt.cc.

Author: Sachin Setiya
Author Date: 2018-10-10 13:24:39 UTC

dummy commit 2