- The reason the test failed was competition between 3+ QEPs with identical
  costs. Before, two plans were competing, and that was addressed by using 
  --sorted_result on the EXPLAIN output because they were different only in 
  join order.
  Now we've got a 3rd plan which differs with "Using where" and that doesn't 
  work anymore.
- This patch fixes it by removing 'Using where' from EXPLAIN output. Test coverage
  is somewhat reduced but probably still ok as PBXT and nested outer join processing
  have no interaction and we don't expect any bugs here.

Also some small fixes to make the PBXT testsuite work in --embedded.

config/ac-macros/plugins.m4:
  MYSQL_PLUGIN_DEPENDS_ON_MYSQL_INTERNALS macro extended to support multiple files.
mysql-test/std_data/pbxt_load_unique_error1.inc:
  Move file to be accessible also for testing embedded server.
mysql-test/suite/pbxt/r/pbxt_bugs.result:
  Fix LOAD DATA LOCAL INFILE path so it works also for testing embedded server.
mysql-test/suite/pbxt/t/pbxt_bugs.test:
  Fix LOAD DATA LOCAL INFILE path so it works also for testing embedded server.
mysql-test/suite/pbxt/t/pbxt_locking.test:
  Disable for embedded, as it needs SHOW PROCESSLIST functionality not available there.
mysql-test/suite/pbxt/t/pbxt_transactions.test:
  Disable test for embedded, as it needs ability to connect from outside (mysqldump).
mysql-test/suite/pbxt/t/ps_1general.test:
  Fix replace_result for new mysql-test-run.pl.
sql/sql_plugin.cc:
  Remove hack that disables PBXT in embedded.
storage/pbxt/plug.in:
  Fix crash in PBXT in embedded server.
storage/pbxt/src/Makefile.am:
  Remove not needed CFLAGS/CXXFLAGS (they cause autotools to generate different object names, which in turn cause the MYSQL_PLUGIN_DEPENDS_ON_MYSQL_INTERNALS replacement of object files inside library files not to work).
storage/pbxt/src/ha_pbxt.cc:
  Ugly hack to allow more threads in embedded server (need a better fix I think).
storage/pbxt/src/table_xt.cc:
  Use stderr for logging not stdout (prevent spamming --embedded test suite output with stray messages).
storage/pbxt/src/thread_xt.cc:
  Use stderr for logging not stdout (prevent spamming --embedded test suite output with stray messages).
storage/pbxt/src/trace_xt.cc:
  Use stderr for logging not stdout (prevent spamming --embedded test suite output with stray messages).

Modified RN291: Fixed bug #489088, I use plugin_unlock to remove the reference from the THD, as it is not required

    Bug #494255
    Cannot run RQG due to errors in mysql-test-run.pl v1

Problem was missing variable declaration. Most likely due to a merge issue.

This solves a core dump in MariaDB when one sets the GLOBAL.DEBUG variable in mysql-test-run when other threads are checking the keyword list

dbug/dbug.c:
  Protect stack->keywords with THR_LOCK_dbug

Reviewer: knielsens


sql/slave.cc:
  For 4.0 server (with no time_zone variable), don't stop replication but give a warning

Fix bad merge causing error when specifying test case in non-default suite for mysql-test-run.
Implement the ability to add extra flags and configure options when running BUILD/xxx scripts.
Improve unclear help text in mysql-test-run

BUILD/FINISH.sh:
  Implement the ability to add extra flags and configure options when running BUILD/xxx scripts.
BUILD/SETUP.sh:
  Implement the ability to add extra flags and configure options when running BUILD/xxx scripts.
extra/yassl/taocrypt/include/block.hpp:
  Fix some compiler warnings.
mysql-test/lib/mtr_cases.pm:
  Fix bad merge causing error when specifying test case in non-default suite for mysql-test-run.
  Also remove some non-essential differences to mysql version to simplify future merges.
mysql-test/mysql-test-run.pl:
  Improve help texts.
plugin/fulltext/plugin_example.c:
  Fix some compiler warnings.
vio/viosslfactories.c:
  Fix some compiler warnings.

(-1 doesn't make sence as a row number and when doing insert / load data, first row is 1, so 0 is free to use)


mysql-test/r/warnings.result:
  Added test for warnings with row number
mysql-test/suite/funcs_1/r/innodb_func_view.result:
  Updated test results
mysql-test/suite/funcs_1/r/memory_func_view.result:
  Updated test results
mysql-test/suite/funcs_1/r/myisam_func_view.result:
  Updated test results
mysql-test/t/warnings.test:
  Added test for warnings with row numberAdded test for warnings with row number
sql/my_decimal.cc:
  Don't use -1 as row number
sql/share/errmsg.txt:
  Make row numbers unsigned

Fixed coredump in sql_plugin.cc:intern_plugin_lock() on mysqld start with PBXT


sql/mysqld.cc:
  Fixed coredump in sql_plugin.cc:intern_plugin_lock() on mysqld start with PBXT
sql/share/errmsg.txt:
  Row numbers are always positive
sql/sql_base.cc:
  Fixed race condition in lock tables when killing insert_delayed thread.
  This fixes Bug#47017 rpl_timezone fails on PB-2 with mismatch error
  (Note that the patch only adds a continue; The rest is (required) indentation changes)
sql/sql_class.cc:
  Fixed wrong output for high end machines in outfile_loaddata.
  (Problem was that ER_TRUNCATED_WRONG_VALUE_FOR_FIELD expects ulong, not ulonglong)
sql/sql_insert.cc:
  Ensure that if we get a lock problem with delayed_insert, the error is logged.

Ensure that mysql_get_server_version() also works if there is a non numerical prefix before the version number

Author: Alexander Barkov
License: GPL

mysql-test/t/ctype_ucs.test:
  Added test case for Croatina character set

Author: Antony Curtis
License: BSD


client/mysql.cc:
  Reset variable if CTRL-C was used to kill running query, so that the user can do it again
client/mysqlcheck.c:
  Indentation fix
client/mysqlslap.c:
  Indentation fixes
client/mysqltest.cc:
  Make testing of commands safer by also testing length
  Removed not used variable
  Fixed indentation to be as it was before last patch
mysql-test/lib/mtr_report.pm:
  Fixed typo
mysql-test/mysql-test-run.pl:
  Merge (Align code with default mysql-tes-run.pl)
mysql-test/suite/federated/disabled.def:
  Removed test case
storage/federatedx/ha_federatedx.cc:
  Removed my changes and applied Antony's instead

cmd-line-utils/readline/config_readline.h:
  Add back defines needed for linux to get rid of compiler warnings, but hide them for now until we have a solution for compile failure on MacOSX
strings/ctype-uca.c:
  Always define UCA contraction functions used by ctype-mb.c

Also merge charset patch.

- Removed some not needed casts
- Change plugin.h to be 'binary compatible' with old versions
- Added mysql_ft_size_t typedef to plugin.h to make it trivial to change string lengths to size_t on next ABI change
- Made some fixes suggested by Kristian to make things more portable and future safe (when it comes to strict aliasing)

include/ft_global.h:
  Introduced FT_WEIGTH, to handle fulltext weights in a slightly more portable manner
include/mysql/plugin.h:
  Change plugin.h to be 'binary compatible' with old versions
  Added mysql_ft_size_t typedef to plugin.h to make it trivial to change string lengths to size_t on next ABI change
  Changed flags to unsigned (as flags should always be unsigned)
mysql-test/t/information_schema.test:
  Fixed typo
sql/sp_head.cc:
  Removed cast
sql/sql_select.cc:
  Removed cast
sql/table.cc:
  Removed cast
storage/maria/ma_ft_boolean_search.c:
  Use mysql_ft_size_t instead of size_t for plugin.h code
  Changed some other string lengths to size_t
storage/maria/ma_ft_nlq_search.c:
  Use FT_WEIGTH to make code more portable
storage/maria/ma_ft_parser.c:
  Use mysql_ft_size_t instead of size_t for plugin.h code
  Changed some other string lengths to size_t
storage/maria/ma_ftdefs.h:
  Changed some string lengths to size_t
storage/maria/maria_ftdump.c:
  Use FT_WEIGTH to make code more portable
storage/myisam/ft_boolean_search.c:
  Use mysql_ft_size_t instead of size_t for plugin.h code
storage/myisam/ft_nlq_search.c:
  Use FT_WEIGTH to make code more portable
storage/myisam/ft_parser.c:
  Use mysql_ft_size_t instead of size_t for plugin.h code
storage/myisam/myisam_ftdump.c:
  Use FT_WEIGTH to make code more portable

Added support for Croatian sorting orders utf8_croatian_ci and ucs2_croatian_ci.
Patch done by Alexander Barkov. See http://www.collation-charts.org/articles/croatian.htm

mysql-test/r/ctype_uca.result:
  Added testing of Croatian sort order
mysql-test/t/ctype_uca.test:
  Added testing of Croatian sort order

Don't use static link by default (in compile-pentium) as some new systems doesn't have all static libraries available
Change type for functions in plugin.h:str_mysql_ftparser_param() to const unsigned char and string lengths to size_t.
One effect of the above change is that one needs to include mysql_global.h or define size_t before including plugin.h
This fixes a case where mysql_client_test failed with newer gcc that enables strict-aliasing by default


BUILD/compile-pentium:
  Don't use static link by default as some new systems doesn't have all static libraries available
client/mysql_upgrade.c:
  Remove not used variable
cmd-line-utils/readline/config_readline.h:
  Define some constants to get rid of compiler warnings on Linux
cmd-line-utils/readline/display.c:
  Get rid of compiler warnings
cmd-line-utils/readline/history.c:
  Got rid of compiler warnings:
  - Defining some strings as const
  - Added cast
cmd-line-utils/readline/rlmbutil.h:
  Added cast to get rid of compiler warnings
cmd-line-utils/readline/text.c:
  Remove not needed initialization to get rid of compiler warnings
cmd-line-utils/readline/xmalloc.c:
  Changed types to 'const char* to get rid of compiler warnings
configure.in:
  Ensure that we use MariaDB as suffix
include/mysql/plugin.h:
  Changed types to 'const unsigned char* to get rid of compiler warnings (in other parts of the code)
  Change length for not \0 terminated string to size_t
include/mysql/plugin.h.pp:
  Update related to plugin.h
libmysql/libmysql.c:
  Fixed bug that caused core dump with newer gcc when strict aliasing is not turned off
mysql-test/t/information_schema.test:
  Test is depending on innodb
mysql-test/t/not_partition.test:
  Fixed wrong directory name
  (Not noticed before as we don't ususally run this test)
mysys/lf_hash.c:
  Got rid of compiler warnings from -Wstrict-aliasing
mysys/my_redel.c:
  Removed not used variable
regex/engine.c:
  Changed types to 'const char* to get rid of compiler warnings
regex/engine.ih:
  Changed types to 'const char* to get rid of compiler warnings
sql/sp_head.cc:
  Got rid of compiler warning from -Wstrict-aliasing
sql/sql_base.cc:
  Got rid of compiler warnings from -Wstrict-aliasing
  (The original code was probably wrong as nj_col->table_field was
sql/sql_builtin.cc.in:
  plugin.h needs to have size_t defined
sql/sql_parse.cc:
  Remove used variable
sql/sql_select.cc:
  Got rid of compiler warnings from -Wstrict-aliasing
sql/sql_show.cc:
  Added #ifdef to get rid of compiler warning when not using partition engine
sql/table.cc:
  Got rid of compiler warning from -Wstrict-aliasing
storage/maria/ha_maria.cc:
  Got rid of compiler warnings from -Wstrict-aliasing:
  - Use the thd_killed() API function
storage/maria/lockman.c:
  Got rid of compiler warnings from -Wstrict-aliasing
storage/maria/ma_check.c:
  Got rid of compiler warnings from -Wstrict-aliasing
  Change to use new version of _ma_killed_ptr; Don't call it as often as before
storage/maria/ma_check_standalone.h:
  Update to compatible _ma_killed_ptr() from ha_maria.cc
storage/maria/ma_ft_boolean_search.c:
  Changed pointers from char -> const char* and length to size_t (to get rid of compiler warnings and casts)
storage/maria/ma_ft_nlq_search.c:
  Got rid of compiler warnings from -Wstrict-aliasing
  Ensure that 'subkeys' is 32 bit
storage/maria/ma_ft_parser.c:
  Changed pointers from char -> const char* and length to size_t (to get rid of compiler warnings and casts)
storage/maria/ma_ftdefs.h:
  Changed pointers from char -> const char* and length to size_t (to get rid of compiler warnings and casts)
storage/maria/ma_sort.c:
  Change to use new version of _ma_killed_ptr; Don't call it as often as before
storage/maria/ma_state.c:
  Got rid of compiler warnings from -Wstrict-aliasing
storage/maria/maria_def.h:
  Redefine ma_killed_ptr()
storage/maria/maria_ftdump.c:
  Got rid of compiler warnings from -Wstrict-aliasing
storage/maria/trnman.c:
  Got rid of compiler warnings from -Wstrict-aliasing
storage/myisam/ft_boolean_search.c:
  Changed pointers from char -> const char* and length to size_t (to get rid of compiler warnings and casts)
storage/myisam/ft_nlq_search.c:
  Got rid of compiler warnings from -Wstrict-aliasing
storage/myisam/ft_parser.c:
  Changed pointers from char -> const char* and length to size_t (to get rid of compiler warnings and casts)
storage/myisam/ft_stopwords.c:
  Changed pointers from char -> const char* and length to size_t (to get rid of compiler warnings and casts)
storage/myisam/ftdefs.h:
  Changed pointers from char -> const char* and length to size_t (to get rid of compiler warnings and casts)
storage/myisam/ha_myisam.cc:
  Got rid of compiler warnings from -Wstrict-aliasing:
  - Use the thd_killed() API function
storage/myisam/mi_check.c:
  Use new killed_ptr() function
storage/myisam/myisam_ftdump.c:
  Got rid of compiler warnings from -Wstrict-aliasing
storage/myisam/myisamchk.c:
  Update to compatible killed_ptr() from ha_myisam.cc
storage/myisam/myisamdef.h:
  Redefine killed_ptr()
storage/myisam/myisamlog.c:
  Got rid of compiler warnings from -Wstrict-aliasing
storage/myisam/sort.c:
  Change to use new version of killed_ptr; Don't call it as often as before
storage/xtradb/fil/fil0fil.c:
  Fixedc ompiler warning
storage/xtradb/trx/trx0i_s.c:
  Include mysql_plugin.h later to ensure that size_t is defined

mysql-test/mysql-test-run.pl:
  Manually apply similar patch to the one in Bug#47983.
mysql-test/suite/rpl/r/rpl_temporary_errors.result:
  Fix wrong failure with warning in error log due to per-test suppressions
  not being active during server shutdown.
mysql-test/suite/rpl/t/rpl_temporary_errors.test:
  Fix wrong failure with warning in error log due to per-test suppressions
  not being active during server shutdown.
sql/sql_select.cc:
  Manually cherry-pick fix for Bug#45989.
  For some reason, that fix was reverted in MySQL 5.1, even though it is
  necessary to plug the memory leak.