Cherry-pick fix for Bug#53371, security hole with bypassing grants using special path in db/table names.
Bump MariaDB version for security fix release.

The fix actually reverts the change introduced by the patch
for bug 51494
The fact is that the patch for bug 52177 fixes bug 51194 as well.

Fix missing bounds check in string conversion.
Bump version number for security fix release.

Bump version number for security fix release.

After-merge fixes for MySQL 5.1.46 merge into MariaDB: result file update and compiler warning removals.

the previous attempt broke build on Debian4.

Still two test failures to be solved: main.myisam and main.subselect.

sql/sql_base.cc:
  Replace strmov() with strnmov() to remove the possibility for buffer overflow.
sql/sql_parse.cc:
  Reject COM_FIELD_LIST with too-big table or wildcard argument.
  (libmysqlclient doesn't allow sending too long arguments anyway, but we
  need this to protect against buffer overflow exploits).

Fixed race condition in safe_process.cc which may have caused some mysqltests to be reported as failed even if they succeded.

mysql-test/lib/My/SafeProcess/safe_process.cc:
  Fixed race condition when getting a signal while calling waitpid()
  Don't send kill on killed processes
  Return different exit codes depending on how we die.

mysql-test/r/func_time.result:
  Updated result
mysql-test/t/func_time.test:
  sysdate() could be same even with one sleep(1) in between.
storage/maria/ma_loghandler.c:
  Fixed syntax error

Fixed some timing issues in test suite



mysql-test/mysql-test-run.pl:
  Added suppression of errors that comes when master is killed before slave has connect (like with rpl_udf.test)
mysql-test/suite/funcs_1/datadict/processlist_priv.inc:
  Wait for check thread to end
mysql-test/suite/pbxt/t/status.test:
  Added longer sleep to avoid timing issue
support-files/compiler_warnings.supp:
  Fixed wrong suppression

Removed random failures from test suite

mysql-test/extra/rpl_tests/rpl_insert_id_pk.test:
  Make test predicatable.
mysql-test/include/maria_empty_logs.inc:
  We can't use 'Threads_connected' for syncronization, as the 'check_warnings' thread that just quit may still be counted in 'Threads_connected'
  Now we just wait until MySQLD answers again, which should be good enough for our purposes
mysql-test/suite/binlog/r/binlog_index.result:
  Updated results file
mysql-test/suite/binlog/t/binlog_index-master.opt:
  Added option file to not get stack traces in .err file.
mysql-test/suite/binlog/t/binlog_index.test:
  Added 'flush tables' to remove warning about crashed suppression file from logs
mysql-test/suite/pbxt/r/multi_statement.result:
  Updated results
mysql-test/suite/pbxt/t/multi_statement-master.opt:
  Added options so that slow query testing makes sense
sql/events.cc:
  Don't write Event Scheduler startup message if warnings are turned off.
sql/handler.cc:
  Removed compiler warning
sql/log.cc:
  Removed compiler warning
sql/mysqld.cc:
  Added option 'test-expect-abort'; If this is set, we don't write message to log in case of 'DBUG_ABORT'.
  (Gives us smaller, easier to read log files)
sql/set_var.cc:
  Removed compiler warning
sql/slave.cc:
  Removed compiler warning
sql/sql_plugin.cc:
  Don't write warnings about disabled plugin if using --log_warnings=0
storage/xtradb/include/ut0lst.h:
  Removed compiler warning
support-files/compiler_warnings.supp:
  Supress warning from xtradb

The problem is that we can not use make_cond_for_table().
This function relies on used_tables() condition
which is not set properly for subqueries.
As result subquery is not filtered out.
The fix is to use remove_eq_conds() function instead
of make_cond_for_table() func. 'remove_eq_conds()'
algorithm relies on const_item() value and it allows
to handle subqueries in right way.


mysql-test/r/having.result:
  test case
mysql-test/t/having.test:
  test case
sql/sql_select.cc:
  The fix is to use remove_eq_conds() function instead
  of make_cond_for_table() function.