lp:maria

Author: Marko Mäkelä
Author Date: 2019-12-10 14:56:52 UTC

Cleanup test sys_vars.innodb_buffer_pool_size_basic

When using huge pages, the innodb_buffer_pool_size cannot necessarily
be restored. Simplify things by restarting the server.

Author: Marko Mäkelä
Author Date: 2019-12-10 14:18:30 UTC

MDEV-16678: Fix a problem with duplicate #sql2 table names

row_drop_table_for_mysql(): If a #sql2 table is open in another
thread (purge) while we attempting to drop it, rename it to #sql-ib
name to hide it from the SQL layer.

Adjust tests accordingly to hide #sql-ib tables, which can
continue to exist until the background DROP TABLE completes.

Author: Elena Stepanova
Author Date: 2019-12-10 13:43:56 UTC

List of unstable tests for 10.4.11 release

Author: THIRUNARAYANAN BALATHANDAYUTHAPANI
Author Date: 2019-12-10 07:17:49 UTC

MDEV-19176 Reduce the memory usage during recovery

- Moved the recv_sys->heap memory condition inside recv_parse_log_recs().
So that, InnoDB can mark the status as STORE_NO earlier.
- InnoDB uses one third of buffer pool chunk size for reading the redo
log records. In that case, we can avoid the scenario where buffer ran
out of memory issue during recovery.

Author: varun
Author Date: 2019-12-04 14:34:45 UTC

MDEV-20900: IN predicate to IN subquery conversion causes performance regression

Disable the IN predicate to IN subquery conversion when the types on the left and
right hand side of the IN predicate are not of comparable type.

Author: Anel Husakovic
Author Date: 2019-12-09 21:53:57 UTC

Backporting from 10.2 PR #1127 and PR #1150

    - Delete `is_check_constraint.test` and rebase
    `is_check_constraints.test` from `10.2`
    - Adjust granting wrt changes in `10.3` for `table_acl_check`

    Backport from 10.2:
    PR#1127: Fix is_check_constraints.result to be compatibile with 10.3

    The patch is done according to the original patch for MDEV-14474
    1edd09c325525cba33152 and not one which is merged on server
    d526679efd108478cc2af07578.
    This patch includes:
    - Rename from `is_check_constraint` to `is_check_constraints` to tests
    and results
    - Per review, change the order of fields in IS check_constraints table by adding
    the column `table_name` before `constraint_name`. According to the standard
    2006 there is no `table_name` column.
    - Original patch and one in `10.3` supports embedded server this patch doesn't
    support. After the merge `10.3` will not support also.
    - Don't use patch c8b8b01b61 to change the length of `CHECK_CLAUSE` field

    PR#1150: MDEV-18440: Information_schema.check_constraints possible data leak

    This patch is extension of PR 1127 and includes:
    - Check for table grants
    - Additional test according to the MDEV specification

Author: Anel Husakovic
Author Date: 2019-01-24 11:06:56 UTC

PR #1127 and PR #1150

PR#1127: Fix is_check_constraints.result to be compatibile with 10.3

The patch is done according to the original patch for MDEV-14474
1edd09c325525cba33152 and not one which is merged on server
d526679efd108478cc2af07578.
This patch includes:
- Rename from `is_check_constraint` to `is_check_constraints` to tests
and results
- Per review, change the order of fields in IS check_constraints table by adding
the column `table_name` before `constraint_name`. According to the standard
2006 there is no `table_name` column.
- Original patch and one in `10.3` supports embedded server this patch doesn't
support. After the merge `10.3` will not support also.
- Don't use patch c8b8b01b61 to change the length of `CHECK_CLAUSE` field

PR#1150: MDEV-18440: Information_schema.check_constraints possible data leak

This patch is extension of PR 1127 and includes:
- Check for table grants
- Additional test according to the MDEV specification

Author: Sergey Petrunia
Author Date: 2019-12-09 20:25:13 UTC

MDEV-21243: Join buffer: condition is checked in wrong place for range access

In this scenario:
- There is a possible range access for table T
- And there is a ref access on the same index which uses fewer key parts
- The join optimizer picks the ref access (because it is cheaper)
- make_join_select applies this heuristic to switch to range:
  /* Range uses longer key; Use this instead of ref on key */

Join buffer will be used without having called
JOIN_TAB::make_scan_filter(). This means, conditions that should be
checked when reading table T will be checked after T is joined with the
contents of the join buffer, instead.

Fixed this by adding a make_scan_filter() check.
(updated patch after backport to 10.3)
(Backported to 10.3.10)

Author: Eugene
Author Date: 2019-12-09 13:42:58 UTC

WIP2

Author: Teemu Ollakka
Author Date: 2019-12-09 05:45:40 UTC

MDEV-21229 Update wsrep-lib to deal with libstdc++ vector assertion (#1423)

The assertion was caused by use of operator[] for an empty vector.
The wsrep-lib update changes vector operations which access
the underlying data array to use data() method which is valid
for empty vectors.

Author: Vladislav Vaintroub
Author Date: 2019-12-08 16:05:57 UTC

CONC-447 Bad certificate validation on client

Enable CRL client test on Windows

Author: Alexey Botchkov
Author Date: 2019-12-08 10:52:53 UTC

MDEV-18463 Don't allow multiple table CONSTRAINTs with the same name.

Add necessary checks.

Author: midenok
Author Date: 2019-12-07 20:24:11 UTC

FOREIGN_KEY_INFO refactored to FK_info holding Lex_cstring

Author: Sergey Petrunia
Author Date: 2019-12-07 15:21:36 UTC

MDEV-21243: Join buffer: condition is checked in wrong place for range access

In this scenario:
- There is a possible range access for table T
- And there is a ref access on the same index which uses fewer key parts
- The join optimizer picks the ref access (because it is cheaper)
- make_join_select applies this heuristic to switch to range:
  /* Range uses longer key; Use this instead of ref on key */

Join buffer will be used without having called
JOIN_TAB::make_scan_filter(). This means, conditions that should be
checked when reading table T will be checked after T is joined with the
contents of the join buffer, instead.

Fixed this by adding a make_scan_filter() check.
(updated patch after backport to 10.3)

Author: Alexander Barkov
Author Date: 2019-12-07 14:17:08 UTC

A cleanup for MDEV-8844: Fixing compilation failure on Windows

Fixing lossy type conversions:
- from int64 to int
- from size_t to uint

Author: Sujatha
Author Date: 2019-12-06 14:26:23 UTC

MDEV-20959: binlog.binlog_parallel_replication_marks_row fails in buildbot with wrong result

Problem:
=======
Test "binlog.binlog_parallel_replication_marks_row" fails sporadically due to
result length mismatch.

Analysis:
=========
Test generates a binary log and it looks for certain words within the binary
log file and prints them. For example word like "GTID,BEGIN,COMMIT ...".
Binary log output contains base64 encoded characters. Occasionally the encoded
characters match with the above words and results in test failure.

+XwoFWxMBAAAALgAAAGEDAAAAAB8AAAAAAAEABHRlc3QAAnQxAAIDAwACFGTIDQ==
+AAAAAAAAAAAEEwQADQgICAoKCgGTIDw9

Fix:
===
Improve the regular expression to match exact words.

Author: Anel Husakovic
Author Date: 2019-01-22 08:53:57 UTC

MDEV-18323 Convert MySQL JSON type to MariaDB TEXT in mysql_upgrade

Author: Eugene
Author Date: 2019-12-03 06:09:04 UTC

MDEV-20950 Reduce size of record offsets

offset_t: this is a type which represents one record offset.
It's uint16_t.

a lot of functions: replace ulint with offset_t

btr_pcur_restore_position_func(),
page_validate(),
row_ins_scan_sec_index_for_duplicate(),
row_upd_clust_rec_by_insert_inherit_func(),
row_vers_impl_x_locked_low(),
trx_undo_prev_version_build():
  allocate record offsets on the stack instead of waiting for rec_get_offsets()
  to allocate it from mem_heap_t. So, reducing memory allocations.

RECORD_OFFSET, INDEX_OFFSET:
  now it's less convenient to store pointers in offset_t*
  array. One pointer occupies now several offset_t. And those constant are start
  indexes into array to places where to store pointer values

REC_OFFS_HEADER_SIZE: adjusted for the new reality

REC_OFFS_NORMAL_SIZE:
  increase size from 100 to 300 which means less heap allocations.
  And sizeof(offset_t[REC_OFFS_NORMAL_SIZE]) now is 600 bytes which
  is smaller than previous 800 bytes.

REC_OFFS_SEC_INDEX_SIZE: adjusted for the new reality

rem0rec.h, rem0rec.ic, rem0rec.cc:
  various arguments, return values and local variables types were changed to
  fix numerous integer conversions issues.

enum field_type_t:
  offset types concept was introduces which replaces old offset flags stuff.
  Like in earlier version, 2 upper bits are used to store offset type.
  And this enum represents those types.

REC_OFFS_SQL_NULL, REC_OFFS_MASK: removed

get_type(), set_type(), get_value(), combine():
  these are convenience functions to work with offsets and it's types

rec_offs_base()[0]:
  still uses an old scheme with flags REC_OFFS_COMPACT and REC_OFFS_EXTERNAL

rec_offs_base()[i]:
  these have type offset_t now. Two upper bits contains type.

Author: Nikita Malyavin
Author Date: 2019-11-20 02:48:16 UTC

key_copy: use from_record argument to copy the data from

Author: Kentoku SHIBA
Author Date: 2019-12-05 19:32:56 UTC

fix compiler warnings

Author: Eugene
Author Date: 2019-11-04 19:30:12 UTC

MDEV-20950 Reduce size of record offsets

offset_t: this is a type which represents one record offset.
It's uint16_t.

a lot of functions: replace ulint with offset_t

btr_pcur_restore_position_func(),
page_validate(),
row_ins_scan_sec_index_for_duplicate(),
row_upd_clust_rec_by_insert_inherit_func(),
row_vers_impl_x_locked_low(),
trx_undo_prev_version_build():
  allocate record offsets on the stack instead of waiting for rec_get_offsets()
  to allocate it from mem_heap_t. So, reducing memory allocations.

RECORD_OFFSET, INDEX_OFFSET:
  now it's less convenient to store pointers in offset_t*
  array. One pointer occupies now several offset_t. And those constant are start
  indexes into array to places where to store pointer values

REC_OFFS_HEADER_SIZE: adjusted for the new reality

REC_OFFS_NORMAL_SIZE:
  increase size from 100 to 300 which means less heap allocations.
  And sizeof(offset_t[REC_OFFS_NORMAL_SIZE]) now is 600 bytes which
  is smaller than previous 800 bytes.

REC_OFFS_SEC_INDEX_SIZE: adjusted for the new reality

rem0rec.h, rem0rec.ic, rem0rec.cc:
  various arguments, return values and local variables types were changed to
  fix numerous integer conversions issues.

enum field_type_t:
  offset types concept was introduces which replaces old offset flags stuff.
  Like in earlier version, 2 upper bits are used to store offset type.
  And this enum represents those types.

REC_OFFS_SQL_NULL, REC_OFFS_MASK: removed

get_type(), set_type(), get_value(), combine():
  these are convenience functions to work with offsets and it's types

rec_offs_base()[0]:
  still uses an old scheme with flags REC_OFFS_COMPACT and REC_OFFS_EXTERNAL

rec_offs_base()[i]:
  these have type offset_t now. Two upper bits contains type.

Author: Marko Mäkelä
Author Date: 2019-12-05 13:28:59 UTC

Fix a race condition

dict_table_t::parse_name(): Convert an InnoDB table name to MDL.
Use dict_sys.mutex to protect dict_table_t::name against renaming.

Author: Eugene
Author Date: 2019-12-04 15:29:33 UTC

MDEV-21014 MTR does not detect {A,M,T,L,UB}SAN errors which happen upon server shutdown

Let MTR check for error existence after running a test and return it back to user.

Error reporting itset might be much better, but first of all we need to see that
something went wrong.

Author: Anel Husakovic
Author Date: 2019-12-05 12:23:32 UTC

After review fix

- Delete byte at the end of key/values
- Add space after colon for keys and comma for values
- Escaping character `'` should be ignored

Author: Marko Mäkelä
Author Date: 2019-12-04 16:10:31 UTC

MDEV-21172 Memory leak after failed ADD PRIMARY KEY

Author: Eugene
Author Date: 2019-12-05 07:51:55 UTC

MDEV-21223 innodb_fts.sync_ddl fails in buildbot, server crashed in que_thr_step

FreeState(): replace pointer to freed memory with NULL. This actually fixes a crash
which is use-after-free as reported by ASAN

DbugParse(): unconditionally lock mutex because we're touching shared init_settings.keywords

Author: Vicențiu Ciorbaru
Author Date: 2019-12-04 21:37:04 UTC

cleanup

Author: varun
Author Date: 2019-12-04 14:34:45 UTC

MDEV-20900: IN predicate to IN subquery conversion causes performance regression

Disable the IN predicate to IN subquery conversion when the types on the left and
right hand side of the IN predicate are not of comparable type.

Author: Larysa Sherepa
Author Date: 2019-12-04 13:19:12 UTC

MDEV-20959 binlog.binlog_parallel_replication_marks_row fails with wrong results

Author: Alexander Barkov
Author Date: 2019-12-04 11:49:31 UTC

A cleanup for MDEV-17088 Provide tools to encode/decode mysql-encoded file system names

- Load and convert the entire input file at once,
  rather than reading string-by-string using fgets().
  This change makes it possible to convert from UCS2, UTF16, UTF32 data.
- Adding the --delimiter command, to treat the specified
  characters as delimiters rather than data to convert.
  Useful in combination with `-f filename` or `-t filename`.
  The delimiter characters are not converted,
  they are copied from the input to the output as is.

Author: Sujatha
Author Date: 2019-12-04 10:32:04 UTC

MDEV-18046: Assortment of crashes, assertion failures and ASAN errors in mysql_show_binlog_events

Problem:
========
SHOW BINLOG EVENTS FROM <pos> reports following assert when ASAN is enabled.

Query_log_event::Query_log_event(const char*, uint,
    const Format_description_log_event*, Log_event_type):
  Assertion `(pos) + (6) <= (end)' failed

Fix:
===
**Part9: Removed additional DBUG_ASSERT**

Author: midenok
Author Date: 2019-12-03 18:26:44 UTC

MDEV-18929 2nd execution of SP does not detect ER_VERS_NOT_VERSIONED

Related to 87731177 and Bug#47649.

Don't set Query_arena::STMT_EXECUTED for any error
code. skip_setup_conds() depends on STMT_INITIALIZED_FOR_SP, but
sp_lex_keeper::reset_lex_and_exec_core() sets it to STMT_EXECUTED on
ER_TABLE_NOT_LOCKED_FOR_WRITE. There are other error codes that can
break skip_setup_conds() (ER_IT_IS_A_VIEW, ER_NON_UPDATABLE_TABLE,
etc).

Author: midenok
Author Date: 2019-12-03 18:26:44 UTC

MDEV-18929 2nd execution of SP does not detect ER_VERS_NOT_VERSIONED

Related to 87731177 and Bug#47649.

Don't set Query_arena::STMT_EXECUTED for any error
code. skip_setup_conds() depends on STMT_INITIALIZED_FOR_SP, but
sp_lex_keeper::reset_lex_and_exec_core() sets it to STMT_EXECUTED on
ER_TABLE_NOT_LOCKED_FOR_WRITE. There are other error codes that can
break skip_setup_conds() (ER_IT_IS_A_VIEW, ER_NON_UPDATABLE_TABLE,
etc).

Author: Jan Lindström
Author Date: 2019-12-03 13:02:17 UTC

Merge branch 'codership-10.4-MDEV-20848' into 10.4

Author: midenok
Author Date: 2019-12-03 12:28:41 UTC

MDEV-21147 Assertion `marked_for_read()' upon UPDATE on versioned table via view

Unit prepare prematurely fixed field which must be fixed via
setup_conds() to correctly update table->covering_keys.

Call vers_setup_conds() directly instead, because actually everything
else is not needed.

Author: Vladislav Vaintroub
Author Date: 2019-11-26 08:35:09 UTC

Part2: MDEV-12518 Unify sql_yacc.yy and sql_yacc_ora.yy

Generate sql_yacc_ora.yy from sql_yacc.yy , by using cmake script
which does some simple string(REPLACE)

Author: Oleksandr Byelkin
Author Date: 2019-12-03 08:39:53 UTC

Merge branch '5.5' into 10.1

Author: Anel Husakovic
Author Date: 2019-11-28 14:08:29 UTC

Using `variables` instead of `values` in mysqld --help documentation would be more accurate

Author: midenok
Author Date: 2019-12-02 18:43:38 UTC

Versioning partition tests reorganize

Moved partition_rotation, partition_innodb to partition.test

Author: varun
Author Date: 2019-11-15 14:10:47 UTC

MDEV-6915: Allow packed keys and packed values of non-sorted fields in the sort buffer

Author: Jan Lindström
Author Date: 2019-11-30 07:48:15 UTC

MDEV-21182: Galera test failure on MW-284

galera_2nodes.cnf did not contain wsrep_on=1 on correct places. Fixed
restart options to use correct configuration.

Author: Vladislav Vaintroub
Author Date: 2019-11-29 22:26:04 UTC

MDEV-16264 - some improvements

- wait notification, tpool_wait_begin/tpool_wait_end - to notify the
threadpool that current thread is going to wait

Use it to wait for IOs to complete and also when purge waits for workers.

Author: varun
Author Date: 2019-11-29 17:16:16 UTC

Refactoring

Author: varun
Author Date: 2019-11-28 15:25:26 UTC

Refactoring

Author: Kentoku SHIBA
Author Date: 2019-11-28 23:22:13 UTC

MDEV-18973 CLIENT_FOUND_ROWS wrong in spider

Get count from last_used_con->info
Contributed by willhan at Tencent Games

Author: Vlad Lesin
Author Date: 2019-11-20 18:27:30 UTC

MDEV-18310: Aria engine: Undo phase failed with "Got error 121 when
executing undo undo_key_delete" upon startup on datadir restored from
incremental backup

aria_log* files were not copied on --prepare --incremental-dir step from
incremental to destination backup directory.

Author: Sachin Setiya
Author Date: 2019-11-28 18:09:22 UTC

more info

Author: Sachin Setiya
Author Date: 2019-11-28 18:08:36 UTC

12

Author: Vlad Lesin
Author Date: 2019-11-27 22:16:22 UTC

Add some extra logging to catch the case when page is written with wrong offset.

Author: Vlad Lesin
Author Date: 2019-11-27 22:16:22 UTC

Add some extra logging to catch the case when page is written with wrong offset.

Author: Sergei Golubchik
Author Date: 2019-11-28 13:01:46 UTC

Fix upgrade errors on eoan

mariadb packages conflict with mysql-8.0

Author: Vlad Lesin
Author Date: 2019-11-27 22:16:22 UTC

Add some extra logging to catch the case when page is written with wrong offset.

Author: Sergei Golubchik
Author Date: 2019-11-27 13:36:21 UTC

fix ssl_crl test failures on newer OpenSSL

generalize the replacement

Author: Vladislav Vaintroub
Author Date: 2019-11-27 19:44:14 UTC

MDEV-19669
proxy_protocol_networks does not evaluate/parse properly

Prior to this fix, when matching addresses using mask,
extra bits could be used for comparison, e.g to
match with "a.b.c.d/24" , 27 bits were compared rather than 24.

The patch fixes the calculation.

Author: Nikita Malyavin
Author Date: 2019-10-31 23:38:22 UTC

write_record: move gotos to funcs

Author: Vicențiu Ciorbaru
Author Date: 2019-11-26 17:52:37 UTC

MDEV-13288: Proper fix for cracklib-runtime

The required dependencies should be added through the autobake script,
to also cover distributions that do not support libcrack2.

Author: Rasmus Johansson
Author Date: 2019-11-17 08:39:47 UTC

MDEV-19781 Create MariaDB named commands on Windows

Added CreateSymlinks and DeleteSymlinks functions to
CustomAction.cpp. Extra.wxs.in calls them.

Author: Nikita Malyavin
Author Date: 2019-10-31 23:38:22 UTC

move gotos to funcs

Author: Nikita Malyavin
Author Date: 2019-08-16 12:12:05 UTC

check referenced table period name in FK

Author: Vladislav Vaintroub
Author Date: 2019-11-25 13:41:13 UTC

Fix shutdown hang in dict_stats , caused by MDEV-16264

dict_stats_shutdown() can hang, waiting for timer callback to finish.
This happens because locks the same mutex, which can also used inside
timer callback, within dict_stats_schedule() function.

Fix is to make dict_stats_schedule() use mutex.try_lock() instead of
mutex.lock().

In the unlikely case of simultaneous dict_stats_schedule() setting
different timer delays, now the first one would win, which is fine.
Important is that shutdown won't hang.

Author: willhan
Author Date: 2019-04-16 14:10:05 UTC

fix bug for spider where using "not like" (#890)

test case:
t1 is a spider engine table;
  CREATE TABLE `t1` (
  `id` int(11) NOT NULL DEFAULT '0',
  `name` char(64) DEFAULT NULL,
  PRIMARY KEY (`id`)
) ENGINE=SPIDER

query: "select * from t1 where name not like 'x%' " would dispatch "select xxx name name like 'x%' " to remote mysqld, is wrong

Author: Jan Lindström
Author Date: 2019-11-25 12:46:48 UTC

Merge branch 'codership-bb-10.1-MDEV-19572' into 10.1

Author: THIRUNARAYANAN BALATHANDAYUTHAPANI
Author Date: 2019-11-23 07:21:53 UTC

MDEV-20190 Instant operation fails when add column and collation change on non-indexed column

- Relax the debug assertion when doing instant operation on collation
changed column.

Author: Sergei Golubchik
Author Date: 2019-11-22 17:52:40 UTC

Merge tag 'mariadb-5.5.66' into 10.0

Author: Sergei Golubchik
Author Date: 2019-11-22 17:52:40 UTC

Merge tag 'mariadb-5.5.66' into 10.0

Author: Jan Lindström
Author Date: 2019-11-22 06:59:01 UTC

Merge branch 'codership-bb-10.4-MDEV-21096' into 10.4

Author: Nikita Malyavin
Author Date: 2019-09-03 18:29:03 UTC

MDEV-16039 Crash when selecting virtual columns generated using functions with DAYNAME()

Due to collation specifics an Item tree could change while fixing. Virtual fields.

This should be done on the correct query arena, which is table->expr_arena.

Author: Anel Husakovic
Author Date: 2019-11-18 18:01:20 UTC

MDEV-20787: Script dgcov.pl does not work

Let's change CMakeList with `--coverage` flag as an alias for
`-fprofile-arcs -ftest-coverage -lgcov` in addition.
When the server is compiled with `-DENABLE_GCOV=ON`, from object files are generated
`.gcno` and `.gcda` files.
`./mtr --gcov is_check_constraint` is invoking the script calls
`./dgcov.pl --purge`, `./mtr is_check_constraint`,
`./dgcov.pl --generate>/var/last_changes.dgcov`.
The `purge` flag is clearing `.gcda` files (and others extensions),
while running the test new `.gcda` files are obtained.
With `generate` flag, `gcov -i` (`intermediate format`) is called
on obtained `<object-files-name>.gcda` files (`dbug.c.gcda` e.g.).
The patch is tested on `gcov 6.3` and `gcov 7.4` versions
and can be seen that resulting `.gcov` file for `6.3` creates
`<full path>.gcov` (`dbug.c.gcda.gcov` e.g.) file,
where `gcov 7.4` is still creating `object-file-names.gcov`(`dbug.c.gcov`) files
as `gcov` in general is doing.

Author: Sachin Setiya
Author Date: 2019-11-18 14:35:28 UTC

MDEV-21063 Very many test failures on big-endian PowerPC

Memcpy is endian dependent, So instead of memcpy we will use int2store.
And m_metadata_size can be 0,1 or 2, And int2store will take care of it.

Author: midenok
Author Date: 2019-10-22 16:47:10 UTC

More dberr_t

Author: Sergey Vojtovich
Author Date: 2018-05-23 14:52:55 UTC

Use std::isfinite in C++ code

This is addition to parent revision fixing build failures.

(cherry picked from commit 54999f4e75f42baca484ae436b382ca8817df1dd)

Author: Sergey Petrunia
Author Date: 2019-11-16 10:18:24 UTC

Fix compile failure on Windows: use explicit type casts

Author: Vlad Lesin
Author Date: 2019-11-15 12:15:10 UTC

MDEV-21050: warning: 'finite' is deprecated: first deprecated in macOS
10.9

The issue is not only in is_finite() deprecation.

On some platforms(like Ubuntu 16.04) math.h contains isnan/isinf
definition only for C99 standard. If some code is compiled with C++11
standard, math.h does not contain isnan/isinf definitions, it's supposed
cmath must be used instead of math.h for C++11.

At the other hand, CHECK_SYMBOL_EXISTS(isnan math.h HAVE_ISNAN) returns
true because it uses try_compile() cmake function, which does not apply
CMAKE_CXX_FLAGS or CMAKE_CXX_STANDARD in cmake <= 3.7(default cmake
version for Ubuntu 16.04 is 3.5), and the code, passed as an argument to
try_compile(), is compiled under default C99 standard(if the
correspondent compilation flag is not set explicitly in CMAKE_REQUIRED_FLAGS).

So on Ubuntu 16.04 we have HAVE_ISNAN/HAVE_ISINF set to true, but
isnan/isinf are not defined in math.h if the code is compiled under C++11
standard, what is true for ES and 10.4+ CS.

To avoid this, we use std::isnan/std::isinf if the code is compiled with
C++11 standard.

To use std::isnan/std::isinf we must include cmath. It was included in
my_global.h only if "isfinite" macro is not defined and C++11 is used.
But there is another issue. On some platforms like MacOS X, "isfinite"
is defined for both C99 and C++11 standards, so on MacOS X cmath was not
included in my_global.h, what caused compilation errors for the code which used
std::inan/std::isinf. Besides "isfinite" is deprecated since MacOS 10.9,
what led to a lot of warnins during compilation.

That is why we include cmath and use std::isfinite on C++11
standard, even if "isfinite" is defined.

Besides, C++11 standard contains the following information(C++11
17.6.1.2/4):

In the C++ standard library, however, the declarations (except for names
which are defined as macros in C) are within namespace scope (3.3.6) of
the namespace std. It is unspecified whether these names are first
declared within the global namespace scope and are then injected into
namespace std by explicit using-declarations (7.3.3).

So <cmath> can contain isfinite() declaration, that is why macro
definition for "isfinit" is used instead of inline function declaration.

Author: Sergey Vojtovich
Author Date: 2019-11-14 13:45:19 UTC

Don't use plugin->data for storage engine plugins

Use plugin->plugin->info->hton instead.
plugin_data() replaced with plugin_hton().
plugin_hton() must never return NULL anymore and is only good to be called
against plugins in PLUGIN_IS_READY state.

Part of
MDEV-20044 - Replace dynamic storage engine initialisation with declarative
             approach

Author: Vladislav Vaintroub
Author Date: 2019-11-13 20:12:48 UTC

MDEV-21062 Buildbot, Windows - sporadically missing lines from mtr's "exec"

Provide own version of popen/pclose, in attempt to workaround
sporadic erratic behavior of UCRT's one.

Author: varun
Author Date: 2019-11-15 06:51:26 UTC

MDEV-21044: Wrong result when using a smaller size for sort buffer

Make sure that the sort buffers can store atleast one sort key.
This is needed to make sure that all merge buffers are read else
with no sort keys some merge buffers are skipped because the code
makes a conclusion there is no data to be read.

Author: Oleksandr Byelkin
Author Date: 2019-11-06 11:35:19 UTC

MDEV-20076: SHOW GRANTS does not quote role names properly

Quotes added to output.

Author: Sujatha
Author Date: 2019-10-14 06:38:28 UTC

MDEV-20707: Missing memory barrier in parallel replication error handler in wait_for_prior_commit()

revision-id: 673e253724979fd9fe43a4a22bd7e1b2c3a5269e
Author: Kristian Nielsen

Fix missing memory barrier in wait_for_commit.

The function wait_for_commit::wait_for_prior_commit() has a fast path where it
checks without locks if wakeup_subsequent_commits() has already been called.
This check was missing a memory barrier. The waitee thread does two writes to
variables `waitee' and `wakeup_error', and if the waiting thread sees the
first write it _must_ also see the second or incorrect behavior will occur.
This requires memory barriers between both the writes (release semantics) and
the reads (acquire semantics) of those two variables.

Other accesses to these variables are done under lock or where only one thread
will be accessing them, and can be done without barriers (relaxed semantics).

Author: THIRUNARAYANAN BALATHANDAYUTHAPANI
Author Date: 2019-11-13 08:10:50 UTC

MDEV-16678 Use MDL for innodb background threads instead of dict_operation_lock

- Remove index_fts_syncing and its related functions

Author: Eugene
Author Date: 2019-11-12 11:15:26 UTC

cleanup: replace List_iterator(_fast) in handler0alter.cc

Basically, use more List<T>::iterator. This patch required adding two more
overloads to new iterator for convenience.

Author: Sergey Petrunia
Author Date: 2019-11-09 18:03:23 UTC

MDEV-20854: ANALYZE for statements: not clear where the time is spent

Count the "gap" time between table accesses and display it as
r_other_time_ms in the "table" element.

* The advantage of this approach is that it doesn't add any new
  my_timer_cycles() calls.
* The disadvantage is that the definition of what is done during
  "other time" is not that clear: it includes checking the WHERE
  (for this table), constructing index lookup tuple (for the next table)
  writing to GROUP BY temporary table (as we dont account for that time
  separately [yet], etc)

Author: Sujatha
Author Date: 2019-11-11 16:55:35 UTC

MDEV-20923:UBSAN: member access within address … which does not point to an object of type 'xid_count_per_binlog'

Problem:
-------
Accessing a member within 'xid_count_per_binlog' structure results in
following error when 'UBSAN' is enabled.

member access within address 0xXXX which does not point to an object of type
'xid_count_per_binlog'

Analysis:
---------
The problem appears to be that no constructor for 'xid_count_per_binlog' is
being called, and thus the vtable will not be initialized.

Fix:
---
Defined a parameterized constructor for 'xid_count_per_binlog' class.

Author: Andrei Elkin
Author Date: 2019-11-10 20:23:36 UTC

merge 10.2->10.3 with conflict resolutions

Author: Kentoku SHIBA
Author Date: 2019-11-08 01:27:07 UTC

MDEV-18973 CLIENT_FOUND_ROWS wrong in spider

Get count from last_used_con->info
Contributed by willhan at Tencent Games

Author: Vladislav Vaintroub
Author Date: 2019-11-04 17:09:09 UTC

Fix race condition between purge_sys::stop() and purge_sys::resume().

Found by Matthias' tests.
Add latch lock that was missing in resume()

Author: midenok
Author Date: 2019-07-22 12:38:37 UTC

MDEV-20015 Assertion `!in_use->is_error()' failed in TABLE::update_virtual_field

Preserve and restore statement DA.

update_virtual_field() is called as part of index rebuild in
ha_myisam::repair() (MDEV-5800) which is done on bulk INSERT finish.

Assertion in update_virtual_field() was put as part of MDEV-16222
because update_virtual_field() returns in_use->is_error(). The idea:
wrongly mixed semantics of error status before update_virtual_field()
and the status returned by update_virtual_field(). The former can
falsely influence the latter.

Preserve global error status and run update_virtual_field() with clear
DA since no matter how SQL command is finished it must update the
index after bulk INSERT.

Author: varun
Author Date: 2019-11-06 07:54:29 UTC

Extend the join prefix to ensure all tables in the duplicate range either are inside or outside the sort nest

Author: THIRUNARAYANAN BALATHANDAYUTHAPANI
Author Date: 2019-11-06 04:50:32 UTC

MDEV-20987 InnoDB fails to start when fts table has FK relation

Problem:
========
  FK tables are non-evictable from dict_sys cache. But if the fts table
has FK relation then InnoDB tries to make it non-evictable again. So
it leads to failure.

Fix:
===
 InnoDB shouldn't make the FTS table as non-evictable if it is already
non-evictable table when FK is involved.

Author: Oleksandr Byelkin
Author Date: 2019-07-10 11:40:54 UTC

MDEV-18027: Running out of file descriptors and eventual crash

For automatic number of opened files limit take into account number of table instances for table cache

Author: Sergei Golubchik
Author Date: 2019-11-04 17:30:48 UTC

MDEV-20971 ASAN heap-use-after-free in list_delete / heap_close

Don't save/restore HP_INFO as it could be changed by a concurrent thread.
different parts of HP_INFO are protected by different mutexes and
the mutex that protect most of the HP_INFO does not protect its open_list
data.

As a bonus, make heap_check_heap() to take const HP_INFO* and not
make any changes there whatsoever.

Author: Sergei Golubchik
Author Date: 2019-11-03 11:22:55 UTC

Fix linking errors on Windows

followup to 00c3a28820c

Author: pkubaj
Author Date: 2019-10-09 10:01:19 UTC

Fix build on !glibc/powerpc*

Do the same that newer branches do and don't include glibc-related headers on non-glibc environment.

Author: Sergey Vojtovich
Author Date: 2019-11-01 15:18:47 UTC

MDEV-20867 - Perform careful review of "Server crashes with BACKUP STAGE and FLUSH TABLE table_name"

Reverted original patch (c2e0a0b).

For consistency with "LOCK TABLE <table_name> READ" and "FLUSH TABLES
WITH READ LOCK", which are forbidden under "BACKUP STAGE", forbid "FLUSH
TABLE <table_name> FOR EXPORT" and "FLUSH TABLE <table_name> WITH READ
LOCK" as well.

It'd allow consistent fixes for problems like MDEV-18643, which make
BACKUP hardly useful.

Author: Marko Mäkelä
Author Date: 2019-11-01 12:28:33 UTC

MDEV-19457 sys_vars.wsrep_provider_basic failed

wsrep_init_provider_status_variables(): Always set wsrep_inited
to ensure that the memory will be freed.

The initial patch was provided by Julius Goryavsky.

Author: THIRUNARAYANAN BALATHANDAYUTHAPANI
Author Date: 2019-11-01 01:27:20 UTC

Double free of dict_foreign_t for instant table

- Don't evict and reload the dict_foreign_t for instant table if the
foriegn key info is being dropped as a part of alter statement.

Author: Alexey Botchkov
Author Date: 2019-10-31 22:33:25 UTC

MDEV-18244

Author: THIRUNARAYANAN BALATHANDAYUTHAPANI
Author Date: 2019-10-31 12:14:05 UTC

MDEV-16678 Use MDL for innodb background threads instead of dict_operation_lock

- Basically it sorts the undo log records based on table id.
- All table entries in one batch assigned by the same purge thread
- It avoids contention on dict_index_t::lock

Author: Robert Bindar
Author Date: 2019-10-10 09:58:29 UTC

MDEV-17896 Assertion `pfs->get_refcount() > 0' failed

Unfortunate DROP TEMPORARY..IF EXISTS on a regular table may allow
subsequent CREATE TABLE statements to steal away the PFS_table_share
instance from the dropped table.

Author: varun
Author Date: 2019-10-31 10:02:08 UTC

Fix for mdev-20424

Author: Sergei Golubchik
Author Date: 2019-10-30 18:42:16 UTC

cleanup

* use compile_time_assert instead of DBUG_ASSERT

* don't use thd->clear_error(), because
  * the error was already consumed by the error handler, so there is
     nothing to clear
  * it's dangerous to clear errors indiscriminately, if the error came
    from outside of read_statistics_for_tables() it must not be cleared