cmake-wsrep: friendly error message about missing wsrep_api.h

This task involves the implementation for the optimizer trace.

This feature produces a trace for any SELECT/UPDATE/DELETE/,
which contains information about decisions taken by the optimizer during
the optimization phase (choice of table access method, various costs,
transformations, etc). This feature would help to tell why some decisions were
taken by the optimizer and why some were rejected.

Trace is session-local, controlled by the @@optimizer_trace variable.
To enable optimizer trace we need to write:
   set @@optimizer_trace variable= 'enabled=on';

To display the trace one can run:
   SELECT trace FROM INFORMATION_SCHEMA.OPTIMIZER_TRACE;

This task also involves:
    MDEV-18489: Limit the memory used by the optimizer trace
    introduces a switch optimizer_trace_max_mem_size which limits
    the memory used by the optimizer trace. This was implemented by
    Sergei Petrunia.

classes for Json_writer_object and Json_writer_array.
These classes will be used for the implementation
of the optimizer trace.

Without this commit the error message is:

$ cmake . -DWITH_WSREP=ON
-- Running cmake version 3.12.1
-- MariaDB 10.4.3
CMake Error at cmake/wsrep.cmake:36 (FILE):
  FILE STRINGS file
  ".../mariadb-server-10.4/wsrep-lib/wsrep-API/v26/wsrep_api.h"
  cannot be read.
Call Stack (most recent call first):
  CMakeLists.txt:147 (INCLUDE)

Having 'git config cmake.update-submodules yes' nicely updates the
build without error.

$ cmake . -DWITH_WSREP=ON
-- Running cmake version 3.12.1
-- MariaDB 10.4.3
Submodule path 'wsrep-lib': checked out 'e7d72ae7f6a6995a21d743389426a963429a1fff'
Submodule path 'wsrep-lib/wsrep-API/v26': checked out '75a5f452f2ba07b0f4a3a9a94825fccc71b27398'

Remove unused key_thread_bootstrap.
No more global bootstrap_error: it is return value of bootstrap() instead.

Optionally, restore the default root authentication, that is
mysql_native_password USING 'invalid' OR unix_socket

recognize new unix_socket protected accounts, and don't
scare the user into setting the password for them

Change the default authentication for root@localhost to

  IDENTIFIED VIA mysql_native_password USING 'invalid' OR unix_socket

which provides secure passwordless login, while still allowing
SET PASSWORD to work as expected.

Also create a second all-privilege account for the user that owns
datadir (and thus has full access to the data anyway).

Compile unix_socket plugin statically into the server.

cmake/wsrep: allow disabling -DWITH_WSREP=OFF

warning.

When importing a tablespace, we must initialize dummy DEFAULT NULL
values for any instantly added columns in order to avoid a debug
assertion failure when PageConverter::update_records() invokes
rec_get_offsets(). Finally, when the operation completes, we must
evict and reload the table definition, so that the correct
default values for instantly added columns will be loaded.

ha_innobase::discard_or_import_tablespace(): On successful
IMPORT TABLESPACE, evict and reload the table definition,
so that btr_cur_instant_init() will load the correct metadata.

PageConverter::update_index_page(): Fill in dummy DEFAULT NULL values
for instantly added columns. These will be replaced upon the
completion of the operation by evicting and reloading the metadata.

row_discard_tablespace(): Invoke dict_table_t::remove_instant().
After DISCARD TABLESPACE, the table is no longer in "instant ALTER"
format, because there is no data file attached.

MDEV-18178, mariabackup script timestamp change, galera suite fixed & re-recorded tests

Fixed use of uninitialized value in wsrep_TOI_end()

The if/else was a remnant from error voting code which was not
included in the release.

This is only a placeholder that allows an implementation later
during the development of MariaDB, so that downgrade to an earlier
version (with this code) will be possible.
We want to be able to zero out freed pages to reduce write amplification,
and to scrub old data. Zeroing out the pages is optional, not mandatory
for correctness. After all, the MLOG_INIT_FREE_PAGE record can only be
emitted for pages that are marked free in the allocation bitmap page.

Use ibuf_bitmap_page_init() only during recovery.

fsp_fill_free_list(): Initialize the FIL_PAGE_TYPE using MLOG_2BYTES.
The page contents will already have been zeroed out by
MLOG_INIT_FILE_PAGE2.

ibuf_bitmap_init_apply(): Replaces ibuf_parse_bitmap_init().

rec_init_offsets(): Relax a debug assertion that is too strict
during the execution of btr_cur_instant_init_low().

Re-recorded: galera.galera_gcache_recover_manytrx
Fixed: galera_3nodes.galera_evs_suspect_timeout

* Donor node will now provide binlog-index argument to wsrep_sst_rsync script if binlog is used.
* Write correct path and binlog file names into joiner binlog-index file

During parallel execution mariabackup script can fail when trying to create archive because of non unique name.  Extending archive timestamp with nanosecond.

Fix for failures on galera.galera_sst_mysqldump

Test galera_sst_mysqldump could fail occasionally due certification failure. Enforcing stricter causality checks by removing wsrep_sync_wait=0 from test configuration.

MySQL 5.7 introduced the class page_size_t and increased the size of
buffer pool page descriptors by introducing this object to them.

Maybe the intention of this exercise was to prepare for a future
where the buffer pool could accommodate multiple page sizes.
But that future never arrived, not even in MySQL 8.0. It is much
easier to manage a pool of a single page size, and typically all
storage devices of an InnoDB instance benefit from using the same
page size.

Let us remove page_size_t from MariaDB Server. This will make it
easier to remove support for ROW_FORMAT=COMPRESSED (or make it a
compile-time option) in the future, just by removing various
occurrences of zip_size.

instant_alter_column_possible(): allow non-rebuilding operations for
ROW_FORMAT=COMPRESSED

Updated wsrep-lib to galera cache encryption implementation

- Fixes building with galera and tokudb
- Added support for --without-wsrep BUILD script option

- Backport from 10.4

According to close(2) "Retrying the close() after a failure return is
the wrong thing to do"

Corrects 5c81cb88 in MDEV-15635

test a multi-auth with a missing auth plugin on the client

update (and simplify) plugin_auth.test to match

introduce the syntax

... IDENTIFIED { WITH | VIA }
      plugin [ { USING | AS } auth ]
 [ OR plugin [ { USING | AS } auth ]
 [ OR ... ]]

Server will try auth plugins in the specified order until the first
success. No protocol changes, server uses the existing "switch plugin"
packet.

The auth chain is stored in json as

  "auth_or":[{"plugin":"xxx","authentication_string":"yyy"},
             {},
             {"plugin":"foo","authentication_string":"bar"},
            ...],
  "plugin":"aaa", "authentication_string":"bbb"

Note:
* "auth_or" implies that there might be "auth_and" someday;
* one entry in the array is an empty object, meaning to take plugin/auth
  from the main json object. This preserves compatibility with
  the existing mysql.global_priv table and with the mysql.user view.
  This entry is preferrably a mysql_native_password plugin for a
  non-empty mysql.user.password column.

SET PASSWORD is supported and changes the password for the *first*
plugin in the chain that has a notion of a "password"