trx_rollback_active(): When aborting the rollback, free the query graph.

MDEV-12323 Rollback progress log messages during crash recovery are intermixed with unrelated log messages

trx_roll_must_shutdown(): During the rollback of recovered transactions,
report progress and check if the rollback should be interrupted because
of a pending shutdown.

trx_roll_max_undo_no, trx_roll_progress_printed_pct: Remove, along with
the messages that were interleaved with other messages.

row_undo_step(), trx_rollback_active(): Abort the rollback of a
recovered ordinary transaction if fast shutdown has been initiated.

trx_rollback_resurrected(): Convert an aborted-rollback transaction
into a fake XA PREPARE transaction, so that fast shutdown can proceed.

MDEV-13797 InnoDB may hang if shutdown is initiated soon after startup while rolling back recovered incomplete transactions

trx_rollback_resurrected(): If shutdown was initiated, fake all
remaining active transactions to XA PREPARE state, so that shutdown
can proceed. Also, make the parameter "all" an output that will be
assigned to FALSE in this case.

trx_rollback_or_clean_recovered(): Remove the shutdown check
(it was moved to trx_rollback_resurrected()).

trx_undo_free_prepared(): Relax assertions.

row_quiesce_table_start(), row_quiesce_table_complete():
Use the more appropriate predicate srv_undo_sources for skipping
purge control. (This change alone is insufficient; it is possible
that this predicate will change during the call to trx_purge_stop()
or trx_purge_run().)

trx_purge_stop(), trx_purge_run(): Tolerate PURGE_STATE_EXIT.
It is very well possible to initiate shutdown soon after the statement
FLUSH TABLES FOR EXPORT has been submitted to execution.

srv_purge_coordinator_thread(): Ensure that the wait for purge_sys->event
in trx_purge_stop() will terminate when the coordinator thread exits.

The loop in read_xml_field(), unlike the same loop in read_sep_field(),
cannot end with item<>NULL, as it does not have any "break" statements.
The entire block "if (item) {...}" was a dead code.

On Windows, sometimes more files stay open:

[Warning] InnoDB: innodb_open_files=13 is exceeded (15 files stay open)

The fixes for these bugs:

Bug#27586 Wrong autoinc value assigned by LOAD DATA in the NO_AUTO_VALUE_ON_ZERO mode
Bug#22372 Disable spatial key, load data, enable spatial key, crashes table

fixed only LOAD DATA INFILE, but did not fix LOAD XML INFILE.

This patch does for LOAD XML FILE what patches for Bug#27586 and Bug#22372
earlier did for LOAD DATA INFILE.

1. Fixing the auto_increment problem:
   a. table->auto_increment_field_not_null is not set to TRUE
      anymore when a column does not have a corresponding XML tag.
   b. Adding "table->auto_increment_field_not_null= false"
      in the end of read_xml_field().
   These two changes resemble the patch for Bug#27586.

2. Fixing the GEOMETRY problem:
   The result for "reset()" was not tested for errors in read_xml_field(),
   which made it possible for empty string to sneak into a "GEOMETRY NOT NULL"
   column when this column does not have a corresponding XML tag with data.
   After this patch the result of reset() is tested and and an error is
   returned in such cases.
   This change effectively resembles the patch for Bug#22372

3. Spliting the code into a new virtual method Field::load_data_set_null().

   Rationale:
   a. To avoid duplicate code in read_sep_field() and read_xml_field():
      Changes #1 and #2 made the code handling NULL values for Field
      exactly the same in read_sep_field() and read_xml_field().

  b. To avoid tests for field_type(), which is not friendly to
     upcoming data type plugins.
     This change makes it possible for data type plugins
     to implement their own special way for handling NULL values in LOAD DATA
     by overriding Field_xxx::load_data_set_null(),
     like Field_geom and Field_timestamp do.

Instead of updating global counter, calculate Threads_running on the fly.
All threads having command != COM_SLEEP are included.

Behaviour changes:

Previously SHOW STATUS and SHOW GLOBAL STATUS returned the same values
representing global status. Now SHOW STATUS always returns 1 indicating that
current session has 1 thread running.

Previously only threads that were executing dispatch_command() or running events
were accounted by Threads_running. Now it is rough equivalent of
SELECT COUNT(*) FROM INFORMATION_SCHEMA.PROCESSLIST WHERE state!='Sleep'

ha_print_info(): Remove.

srv_printf_innodb_monitor(): Do not acquire btr_search_latches[]
Add the equivalent functionality that was part of the non-debug
version of ha_print_info().

If connection was not blocked, the 'db' was not properly reset
This bug didn't affect information_schema

The problem was that multi_range_read_info_const() called
multi_range_key_create_key() which changed m_part_spec.start_part,
while there was an activ table scan ongoing.

Fixed by copying and restoring m_part_spec around
multi_range_key_create_calls()

Part #2: Take into accont the fact that three binary collations that
we handle are technically NOPAD collations

Disallow use of NOPAD collations in indexed columns.

first, and we do not not care whether client has received all data.

This is a TCP optimization to avoid TIME_WAIT in TCP connection teardown.

This patch would abort connection on timeout, which usually happens when
client reads a large result set, at slower pace then the server can
write.

The patch also cleans up socket timeout handling, so that Windows
is consistent with another platforms (using nonblocking socket IO
 + waiting in poll/select on single socket, rather than setsockopt).
This makes identifying timeouts easier.


Also removed the superficial shutdown() before closesocket() in a few
places where it was used, because it  was never needed , and
reportedly breaks SO_LINGER on Windows.

When the transaction isolation level is SERIALIZABLE, or when
a locking read is performed in the REPEATABLE READ isolation level,
InnoDB must lock delete-marked records in order to prevent another
transaction from inserting something.

However, at READ UNCOMMITTED or READ COMMITTED isolation level or
when the parameter innodb_locks_unsafe_for_binlog is set, the
repeatability of the reads does not matter, and there is no need
to lock any records.

row_search_mvcc(): Skip locks on delete-marked committed records upfront,
instead of invoking row_unlock_for_mysql() afterwards. The unlocking
never worked for secondary index records.

dict_stats_rename_table(): After DB_LOCK_WAIT_TIMEOUT
or DB_DUPLICATE_KEY, reset the trx->error_state before retrying.
Also, properly treat DB_DEADLOCK as a hard error.