- storage/innobase/dict/dict0stats.cc
- storage/xtradb/dict/dict0stats.cc

(Fixing both InnoDB and XtraDB)

Re-opening a TABLE object (after e.g. FLUSH TABLES or open table cache
eviction) causes ha_innobase to call
dict_stats_update(DICT_STATS_FETCH_ONLY_IF_NOT_IN_MEMORY).

Inside this call, the following is done:
  dict_stats_empty_table(table);
  dict_stats_copy(table, t);

On the other hand, commands like UPDATE make this call to get the "rows in
table" statistics in table->stats.records:

  ha_innobase->info(HA_STATUS_VARIABLE|HA_STATUS_NO_LOCK)

note the HA_STATUS_NO_LOCK parameter. It means, no locks are taken by
::info() If the ::info() call happens between dict_stats_empty_table
and dict_stats_copy calls, the UPDATE's optimizer will get an estimate
of table->stats.records=1, which causes it to pick a full table scan,
which in turn will take a lot of row locks and cause other bad
consequences.

On OS X with clang 7.3.0 running any test as embedded crashes. As
libfeedback was not compiled with embedded, we had 2 notions of THD,
with different number of members. When calling inline functions defined in
sql_class.h, the functions were defined for 1 THD but called on an
object that is different. This resulted in erroneous data being returned
and shortly after, a crash within the alloc_query() function.

Recompile the feedback plugin for the embedded server specifically to
not have such symbol conflicts.

Possible fix. Make the pluging more safe.

- Library path's are different now
- New dependency on Linux libuuid
- Add calls for SDK Initialization/shutdown
- Also add request_timeout parameter, default SDK HTTPs timeout appears
to be too short in my tests

Server uses gethostname() for the default base name for pid/log files.
If a character is not representable in current ANSI encoding, gethostname
replaces it with question mark. Thus, generated log file name would also
contain a question mark. However, Windows forbids certain characters in
filenames, among them '?'.

This is described in MSDN article https://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx
At attempts to create the file via freopen() fails, thus server would not
be able to start.

The fix is to verify hostname and fall back to  "mysql", if
invalid characters are found.

Problem was that test innodb.innodb_stats_fetch_corrupted will post
a error InnoDB: Error: Table "mysql"."innodb_index_stats" not found
to a error log and test ignores that message. However, following tests
might see this error also and they might not ignore this error.
Force a server restart after innodb.innodb_stats_fetch_corrupted
to get clean log.

Problem was that NULL-pointer was accessed inside a macro when
page read from tablespace is encrypted but decrypt fails because
of incorrect key file.

Removed unsafe macro using inlined function where used pointers
are checked.

Problem was that length of the filekeys file was not checked and if
length is less than OpenSSL_prefix_len uninitialized memory was
accessed.

        One line in mysql_install_db.sh was left unfixed.

on product builds.

* Out of tree build: mysql_install_db to see all .sql files

Since MDEV-7875 (da0991c6), not all sql source files are in the source
directory, maria_add_gis_sp_bootstrap.sql is in the build directory.

This corrects mysql_install_db{.sh} to be aware of the differing
locations.
Signed-off-by: Daniel Black <daniel.black@au.ibm.com>

* Out of tree build: scripts/mysql_install_db.pl.in
Signed-off-by: Daniel Black <daniel.black@au.ibm.com>

are not consistent on parallel mtr runs.

tablespace. Instead use fil_read() with syncronous setting.
Fix test failures and mask tablespace number as it could
change in concurrent mtr runs.

Analysis: By design InnoDB was reading first page of every .ibd file
at startup to find out is tablespace encrypted or not. This is
because tablespace could have been encrypted always, not
encrypted newer or encrypted based on configuration and this
information can be find realible only from first page of .ibd file.

Fix: Do not read first page of every .ibd file at startup. Instead
whenever tablespace is first time accedded we will read the first
page to find necessary information about tablespace encryption
status.

TODO: Add support for SYS_TABLEOPTIONS where all table options
encryption information included will be stored.

* Fix undefined database test error when running mysql_install_db

When using mariaDb in docker mode it can fail as it calls mysql_install_db but as we are going through a slightly different install process the test database has not been created, therefore we should fall back to the mysql database as per https://mariadb.com/kb/en/mariadb/mariadb-10112-mysql_install_db-aborts-on-unkown-file-test/

* Also fix mysql_install_db.pl.in

MDEV-10425 Assertion `collation.derivation == DERIVATION_IMPLICIT' failed in Item_func_conv_charset::fix_length_and_dec()
MDEV-10850 Wrong result for WHERE .. (f2=TO_BASE64('test') OR f2=TO_BASE64('TEST'))

Problem N1: MDEV-10425
Item_func_{md5|sha|sha2}::fix_length_and_dec() changed args[0]->collation
to force binary comparison in args[0]->eq().
It was done to treat e.g. MD5('a') and MD5('A') as different values.
It is wrong for a Item_func_xxx to modify its arguments.
Item_func_conv_charset did not expect that and crashed on assert.

Problem N2: MDEV-10850
Item_func_to_base64, Item_func_password, Item_func_hex are also case sensitive
hash functions, but they did not compare their arguments as binary.

Solution:
- Removing the code changing args[0]->collation
- Introducing Item_str_ascii_checksum_func as a common parent
  for Item_func_{md5|sha|sha2|password|hex|to_base64}
  and overriding its eq() method to compare arguments binary.

While copying the received state Id (uuid:seqno) to an uninitialized
buffer, it was not properly null-terminated.

When a deadlock kill is detected inside the storage engine, the kill
is not done immediately, to avoid calling back into the storage engine
kill_query method with various lock subsystem mutexes held. Instead the
kill is queued and done later by a slave background thread.

This patch in preparation for fixing TokuDB optimistic parallel
replication, as well as for removing locking hacks in InnoDB/XtraDB in
10.2.

Signed-off-by: Kristian Nielsen <knielsen at knielsen-hq.org>

without preceding sys_vars.wsrep_provider_basic

Once THDs have been added to the global "threads" list,
they must modify query_string only after acquiring per-
thread LOCK_thd_data mutex.

take into account all arguments of aggregate function

remove die() from BUILD/autorun.sh

Make aggregate function dependency visible.

Do not set 'optimized' flag until whole optimization procedure is finished.

List of unstable tests for 10.1

Activated by enabling wsrep_debug.

Fix an issue in Single_line_formatting_helper: flush_on_one_line()
didn't clean up the buffered items which could cause them to be
printed for the second time.

This can't be ever observed by a user (see MDEV text for details).