Issue:
------
When a subquery contains UNION the count of the number of
subquery columns is calculated incorrectly. Only the first
query block in the subquery's UNION is considered and an
array indexing goes out-of-bounds, and this is caught by an
assert.

Solution:
---------
Sum up the columns from all query blocks of the query
expression.

Change specific to 5.6/5.5:
---------------------------
The "child" points to the last query block of the UNION
(as opposed to 5.7+ where it points to the first member of
UNION). So "child->master_unit()->first_select()" is used
to reach the first query block of UNION.

PROBLEM
-------
Memory sanitizer reports uninitialized comparisons
in log_in_use(), because strings are compared with
memcmp() instead of strncmp.

FIX
---
Use strncmp() to compare strings

fil_extend_space_to_desired_size(): Invoke fsync() after posix_fallocate()
in order to durably extend the file in a crash-safe file system.

Avoid introducing new dependencies or new syntax.
That is, don't use $(...) and don't assume dirname is present.

And remove unsighty /foo/bar/../xyz from the path. Use dirname
instead of ../

Closes #1086

don't shortcut trying to test for openssl version,
test what is actually needed for a code to compile

Patch by Eugene Kosov.

Signed-off-by: Alexander Kuleshov <kuleshovmail@gmail.com>

Fix the off-by-one overflow which was introduced with commit
b0fd06a6 (MDEV-15670 - unit.my_atomic failed in buildbot with
Signal 11 thrown)

Closes #1098.

The problem was in calculating of the mask to clear unused null bits in case of using full byte.

mysql_install_db.exe should not remove datadir, if it was not created by
it.

Workaround glibc bug: https://sourceware.org/bugzilla/show_bug.cgi?id=20116
by making unittest threads joinable. It makes code better anyway.

This reverts part of c5427172

Forbid ALTER DATABASE under read_only.

Closes #3

The fix for "MDEV-17698 MEMORY engine performance regression"
previously fixed this problem.

- Adding the test for MDEV-17724
- Re-recording wrong results for tests:
  * engines/iuds/r/insert_number
  * engines/iuds/r/update_delete_number
  which started to fail since MDEV-17698

Also, backporting a part of:
  MDEV-11485 Split Item_func_between::val_int() into virtual methods in Type_handler
for easier merge to 10.3.

my_read fixed as in higher versions.
my_pread made as my_read aware of partial read of huge chunks of files
MY_FULL_IO enabled for file operations

followup for c32f7ed2

reset lex->many_values for LOAD DATA, as it's used for
auto-inc range size estimation.

--gdb now accepts an argument, it will be passed to gdb as a command.
multiple commands can be separated by a (non-standard and not escapable)
delimiter - semicolon (;).

Old usage with a bare --gdb continues to work too, of course.

Cherry-picked c47c0ca5 5441bbd3 339b9055

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

This reverts commit ebaacf07.

Pushed by mistake.
Broke tokudb on trusty-amd64.

test case

This reverts commit dc3a20b1.

It requires GET_BIT in include/my_getopt.h, which is available
only in 10.3+

mysql_upgrade used to convert all columns of mysql.db to
utf8_general_ci and then back to utf8_bin. In two separate ALTER's.

This failed if UNIQUE indexes in mysql.db contained entries
that differ only in the letter case.

 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

CMP0075 Include file check macros honor CMAKE_REQUIRED_LIBRARIES