This reverts commit 13884cf2.

Fixed in 5.6 (merged in the next commit)

"x86_64" is "amd64" on FreeBSD

mysql.cc:
Unlike the main MYSQL structure, kill_mysql did not have MYSQL_OPT_PROTOCOL set.

Move all connection-related settings to a separate function and
use it both for the main MYSQL and for kill_mysql.

MDEV-9390 Function found_rows() gives incorrect result where the previous SELECT contains ORDER BY clause

use the raw found_rows value only when SQL_CALC_FOUND_ROWS was specified

following InnoDB's logic, altering a comment or a default
field's value needs "NO_LOCK", not EXCLUSIVE

Only set Alter_inplace_info::ALTER_COLUMN_VCOL flag if
a vcol might be affected by the ALTER TABLE statement

Followup fix for bug#35074: update the second location where
max_used_connections is calculated (to be like the
first one).

fix a few cases where a successful ALTER was not binlogged:
* on errors after the completed ALTER, binlog it, then return an error
* don't let thd->killed abort open_table() after completed
  online ALTER.

Fix the doubly questional fix for MySQL Bug#17250787:
* it detected autoinc index by looking for the first index
  that starts from autoinc column. never mind one column
  can be part of many indexes.
* it used autoinc_field->field_index to look up into internal
  innodb dictionary. But field_index accounts for virtual
  columns too, while innodb dictionary ignores them.

Find the index by its name, like elsewhere in ha_innobase.

don't use ipv6 is the check fails for any reason. for example, an
ancient solaris perl didn't have sockaddr_in6() in Socket at all.

it caused feedback plugin tests to fail

1. ignore generated dtrace files
2. don't rewrite probes_mysql_nodtrace.h in-place

1. check that unused inline functions are removed
2. only allow compilation if they are or if the check if overridden
3. with CMAKE_GENERATOR=Makefiles, use all flags when testing
   (e.g. both CMAKE_C_FLAGS and CMAKE_C_FLAGS_DEBUG if
   CMAKE_BUILD_TYPE=Debug). This is because
 - on Solaris with the SunPro compiler, default CMAKE_C_FLAGS_xxx
   values contain -xO2 (for Release and RelWithDebInfo)
   and -g (for RelWithDebInfo and Debug)
 - proper inlining only works at -xO4 without -g
 - so if CMAKE_C_FLAGS has -xO4, inlining would work in
   configure.cmake (before this fix) and fail during actual compilation

also remove the outdated check for inline from myu_global.h

1. unused static inline functions are only removed at -xO4,
   otherwise test binaries will depend on various mysys
   symbols that they don't use. Link test with libmysys.

2. Sphinx - don't instantiate (explicitly) templates before
   they're defined. Or, rather, don't instantiate them explicitly at
   all.

3. GIS - don't use anonymous unions and structs.

In case of missing includedir file, we would attempt to free a NULL
pointer. Make sure to guard against that.

In the case of error during my_load_defaults, we would not free the
args array.

dml_setup_instruments did not restore settings in setup_instruments table
after updating it as a part of the test flow

Documentation fixes

MDEV-9354: Debian: unmask the mysql.service on installation

Fix test failure when using maria small-block size. We need to query
the max_key_length and max_key_parts based on the the tmp table engine.

MDEV-7765: Crash (Assertion `!table || (!table->write_set || bitmap_is_set(table->write_set, field_index) || bitmap_is_set(table->vcol_set, field_index))' fails) on using function over not created table

Problem was that created table was not marked as used (not set query_id) and so opening tables for stored function pick it up (as opened place holder for it) and used changing TABLE internals.

The select mentioned in the bug attempted to create a temporary table
using the maria storage engine. The table needs to have primary keys such that
duplicates can be removed. Unfortunately this use case has a longer
than allowed key and the tmp table got created without a temporary key.
We must not allow materialization for the subquery if the total key
length and key parts is greater than what the storage engine supports.

5.5.47-37.7

test case for ALTER TABLE view CHECK PARTITION

test case

Correct fix. Generate the query in the client charset,
not in the system one.

reverted about half of commits as either not applicable or
outright wrong

        The http-proxy option to the FEEDBACK plugin backported.

When one evaluates row-based comparison like (X, Y) = (A,B), one should
first call bring_value() for the Item that returns row value. If you
don't do that and just attempt to read values of X and Y, you get stale
values.
Semi-join/Materialization can take a row-based comparison apart and
make ref access from it. In that case, we need to call bring_value()
to get the index lookup components.

Consider a query with subquery in form t.key=(select ...). Suppose, the
parent query uses this equality for ref access.
It will attempt to evaluate the subquery in get_best_combination(),
right before the join->join_tab[...] array is filled.  The problem was
that subquery optimization will attempt to look at parent's join->join_tab
to check how many times subquery will be executed (and crash).

Fixed by not doing that when the subquery is constant (non-constant
subqueries are only be evaluated during join execution, so they are not
affected)