MDEV-14269 errors.test fails with valgrind (Conditional jump or move depends on uninitialised value)

in joined table + GROUP BY + GROUP_CONCAT + HAVING + ORDER BY
[by field from HAVING] + 1 row expected

The fix is actually a port of the fix for bug #17055185 from
mysql code line (see commit f289aeeef0743508ff87211084453b3b88a6d017
by Mithun C Y into mysql-5.6). The test case for the bug #17055185
was also ported.

- The valgrind warning came from JOIN::optimize() (sql_select.cc:1123)

- Allocate only enough memory that we need for key status
  - Saves ~100K temp usage on maria_open

The background is that one user had a lot of views and using some complex
queries on information schema temporary memory of more than 2G was used.

- Added new element 'total_alloc' to MEM_ROOT for easier debugging.
- Added MAX_MEMORY_USED to information_schema.processlist.
- Added new status variable "Memory_used_initial" that shows how much MariaDB
  uses at startup. This gives the base value for "Memory_used".
- Reuse memory continuously for information schema queries instead of
  only freeing memory at query end.

Other things
- Removed some not needed set_notnull() calls for not null columns.

os_file_set_size(): If posix_fallocate() returns EINVAL, fall back
to writing zero bytes to the file. Also, remove some error log output,
and make it possible for a server shutdown to interrupt the fall-back
code.

MariaDB 10.2 used to handle the EINVAL return value from posix_fallocate()
before commit b731a5bc
which refactored os_file_set_size() to try posix_fallocate().

Why is the EINVAL returned? The GNU posix_fallocate() function
would first try the fallocate() system call, which would return
-EOPNOTSUPP for many file systems (notably, not ext4). Then, it
would fall back to extending the file one block at a time by invoking
pwrite(fd, "", 1, offset) where offset is 1 less than a multiple of
the file block size. This would fail with EINVAL if the file is in
O_DIRECT mode, because O_DIRECT requires aligned operation.

The failing test was main.gis-json

- Expand paths also for jemalloc
- Test also if tcmalloc or jemalloc is in /usr/lib64
- Take into account that .so has a version
- Remove automatic adding of flavors ( _minial, _debug). Better to
  have user specify these directly
- Changed documentation link to MariaDB
- Don't give extra error if mysqld_safe_helper doesn't exist

Also don't abort if global_status_var.global_memory_used != 0 on not
normal exit as then it's likely that not all memory is released.

MDEV-13568: Purge is not blocked by LOCK TABLES. Should it?

MDEV-13893/MDEV-12699: Crash recovery of corrupted or non-decryptable
pages is broken.

Signed-off-by: Daniel Black <daniel@linux.vnet.ibm.com>

When MariaDB 10.1.0 introduced table options for encryption and
compression, it unnecessarily changed
ha_innobase::check_if_supported_inplace_alter() so that ALGORITHM=COPY
is forced when these parameters differ.

A better solution is to move the check to innobase_need_rebuild().
In that way, the ALGORITHM=INPLACE interface (yes, the syntax is
very misleading) can be used for rebuilding the table much more
efficiently, with merge sort, with no undo logging, and allowing
concurrent DML operations.

If InnoDB or XtraDB recovered committed transactions at server
startup, but the processing of recovered transactions was
prevented by innodb_read_only or by innodb_force_recovery,
an assertion would fail at shutdown.

This bug was originally reproduced when Mariabackup executed
InnoDB shutdown after preparing (applying redo log into) a backup.

trx_free_prepared(): Allow TRX_STATE_COMMITTED_IN_MEMORY.

trx_undo_free_prepared(): Allow any undo log state. For transactions
that were resurrected in TRX_STATE_COMMITTED_IN_MEMORY
the undo log state would have been reset by trx_undo_set_state_at_finish().

MDEV-13997 Change Item_bool_rowready_func2 to cache const items at fix time rather than evaluation time

Side effect: the second debug Note in cache_temporal_4265.result disappeared.

Before this change:
- During JOIN::cache_const_exprs(),
  Item::get_cache() for Item_date_add_interval() was called.
  The data type for date_add('2001-01-01',interval 5 day) is VARCHAR,
  because the first argument is VARCHAR (not temporal).
  Item_get_cache() created Item_cache_str('2001-01-06').
- During evaluate_join_record(), get_datetime_value() was called,
  which called Item::get_date() for Item_cache_str('2001-01-06').
  This gave the second Note. Then, get_datetime_value() created
  a new cache, now Item_cache_temporal for '2001-01-06', so not
  further str_to_datetime() happened.

After this change:
- During tem_bool_rowready_func2::fix_length_and_dec(),
  Arg_comparator::set_cmp_func_datetime() is called,
  which immediately creates an instance of Item_cache_date for
  the result of date_add('2001-01-01',interval 5 day).
  So later no str_to_datetime happens any more,
  neither during JOIN::cache_const_exprs(),
  nor during evaluate_join_record().

Forgot to put the updated rocksdb.result in

Downscale rocksdb.add_index_inplace_sstfilewriter to be 10x smaller

Fix a race condition in the testcase.

ReadFile/WriteFile operations.

Innodb opens files with FILE_FLAG_OVERLAPPED. lpNumberOfBytesRead/Written
are documented to be potentially inaccurate in this case,
(possibly even if async operations complete synchronously?)

The fix is to always pass NULL for the correspondng parameters,
as recommended by  MSDN. Read the actual counts with
GetQueuedCompletionStatus() or GetOverlappedResult().

os_file_get_size(): Use fstat() instead of calling lseek() 3 times.
In this way, concurrent calls to this function should not interfere
with each other.

Suggested by Vladislav Vaintroub.

The predicate page_is_root() would not hold if btr_create() fails
before the root page is fully initialized. Move the debug assertion
from btr_free_root_invalidate() to its other caller,
btr_free_if_exists(). In that caller, we actually already checked
for page_is_root().