when it should use index

Sometimes the LEFT/RIGHT JOIN with an empty table caused an
unnecessary filesort.

Sample query, where t1.i1 is indexed and t3 is empty:

  SELECT t1.*, t2.* FROM t1 JOIN t2 ON t1.i1 = t2.i2
                       LEFT JOIN t3 ON t2.i2 = t3.i3
    ORDER BY t1.i1 LIMIT 5;

The server erroneously used an item of empty outer-joined
table as a common constant of a Item_equal (multi-equivalence
expression).
By the fix for the bug 16590 the constant status of such
an item has been propagated to st_table::const_key_parts
map bits related to other Item_equal argument-related
key parts (those are obviously not constant in our case).
As far as test_if_skip_sort_order function skips constant
prefixes of testing keys, this caused an ignorance of
available indices, since some prefixes were marked as
constant by mistake.


mysql-test/r/order_by.result:
  Test case for bug #38745.
mysql-test/t/order_by.test:
  Test case for bug #38745.
sql/item.h:
  Bug #38745: MySQL 5.1 optimizer uses filesort for ORDER BY
              when it should use index
  
  Item::is_outer_field() has been added and overloaded for
  Item_field and Item_ref classes.
sql/item_cmpfunc.cc:
  Bug #38745: MySQL 5.1 optimizer uses filesort for ORDER BY
              when it should use index
  
  Item_equal::update_const() and Item_equal::update_used_tables()
  have been updated to not take into account the constantness
  of outer-joined table items.

Backport from mysql-pe (of those parts which have not been upmerged from 5.1)



sql/field.cc:
  Local scope variable or method argument same as class attribute.
sql/item.cc:
  Rename auto variable to avoid name clash.
sql/item.h:
  Item_ref::basic_const_item had wrong signature (missing const)
  and was thus never called.
sql/partition_info.cc:
  Rename, to avoid name clashes.
sql/sql_load.cc:
  Rename, to avoid name clashes.

Reorganize function to reflect that it does not actually uses curses
and that there is no curses variant.

client/mysql.cc:
  Remove unused functions.
client/sql_string.cc:
  Remove unused functions.
include/my_pthread.h:
  Remove unused prototype.
mysys/my_pthread.c:
  Remove unused function.
sql/lock.cc:
  Remove unused function.
sql/lock.h:
  Remove unused and duplicated prototypes.
sql/sql_class.h:
  Removed unused variables.
sql/sql_const.h:
  Remove unused defines.
sql/sql_priv.h:
  Remove unused defines.
sql/sql_string.cc:
  Remove unused prototype.
sql/thr_malloc.cc:
  Remove unused function.
sql/thr_malloc.h:
  Remove unused prototype.
storage/myisam/CMakeLists.txt:
  Remove obsolete fulltext file.
storage/myisam/Makefile.am:
  Remove obsolete files (were already commented out).
storage/myisam/ft_eval.c:
  Remove obsolete fulltext file.
storage/myisam/ft_eval.h:
  Remove obsolete fulltext file.
storage/myisam/ft_stem.c:
  Remove obsolete fulltext file.
storage/myisam/ft_test1.c:
  Remove obsolete fulltext file.
storage/myisam/ft_test1.h:
  Remove obsolete fulltext file.

Formatting cleanup: removing tabs and trailing spaces
to conform to MySQL Internals Coding Guidelines.

The problem is that if a NULL is stored in an Item_cache_decimal object,
the associated my_decimal object is not initialized.  However, it is still
accessed when val_int() is called. The fix is to check for null_value
within val_int(), and return without accessing the my_decimal object when
the cached value is NULL.

Bug#52122 reports the same issue for val_real(), and this patch also includes
fixes for val_real() and val_str() and corresponding test cases from that
bug report.  

Also, NULL is returned from val_decimal() when value is null. This will
avoid that callers access an uninitialized my_decimal object.

Made similar changes to all other Item_cache classes.  Now all val_*
methods should return a well defined value when actual value is NULL.

mysql-test/r/type_decimal.result:
  Updated result file with test cases for Bug#52168 and Bug#52122.
mysql-test/t/type_decimal.test:
  Added test cases for Bug#52168 and Bug#52122.
sql/item.cc:
  In Item_cache_*::val_* methods, return a well defined value
  when actual value is NULL.
  
  This is especially important for Item_cache_decimal since
  otherwise one risk accessing an uninitialized my_decimal object.
sql/item.h:
  Added method Item_cache::has_value() which returns TRUE if cache 
  object contains a non-null value.

is not needed any more with current versions of FreeBSD.
(Entries 52410 and 52412 in the Bug DB)

That change is based on Dan Nelson replying on the
FreeBSD mailing list, freebsd-questions@freebsd.org
in a thread running from 2010-Apr-29 to 2010-May-1 titled
     "Need info about FreeBSD and interrupted system
      calls for MySQL code"

Also, ensure the cmake settings correspond to the autotools
ones: Add "HAVE_BROKEN_REALPATH" to cmake.


cmake/os/FreeBSD.cmake:
  With current versions of FreeBSD, threading is not done on
  the user-level any more, so the huge "net retry count"
  is not needed any more.
  However, it is still being kept as a comment for now,
  so that it can be re-activated should the need arise.
  
  In addition, set "HAVE_BROKEN_REALPATH" to completely
  mirror the autotools path.
configure.in:
  With current versions of FreeBSD, threading is not done on
  the user-level any more, so the huge "net retry count"
  is not needed any more.
  This definitely holds for FreeBSD version 6 and up.
  
  Adapt the informational messages to the path taken.

For BINARY(N) strip trailing zeroes to make the error message nice-looking


mysql-test/r/errors.result:
  test case
mysql-test/r/type_binary.result:
  result fix
mysql-test/t/errors.test:
  test case
sql/key.cc:
  For BINARY(N) strip trailing zeroes to make the error message nice-looking

We should avoid any SHARE fields assignments as
this is shared structure and assignments may
affect other therads. To avoid this
copy of SHARE struct is created and
stored into TABLE struct which is
used in get_schema_coulumns_record later.



mysql-test/r/mdl_sync.result:
  test case
mysql-test/t/mdl_sync.test:
  test case
sql/sql_show.cc:
  We should avoid any SHARE fields assignments as
  this is shared structure and assignments may
  affect other therads. To avoid this
  copy of SHARE struct is created and
  stored into TABLE struct which is
  used in get_schema_coulumns_record later.

Post-push fix: -Wall implies -Wunused on some platforms,
which will generate thousands of warnings about unused parameters.


configure.cmake:
  Do not warn about unused parameters in C++

Add -Wall to gcc/g++
Fix most warnings reported in dbg and opt mode.


cmd-line-utils/libedit/filecomplete.c:
  Remove unused auto variables.
configure.cmake:
  Add -Wall to gcc.
extra/comp_err.c:
  Cast to correct type.
extra/perror.c:
  Fix segfault (but warnings about deprecated features remain)
extra/yassl/taocrypt/include/runtime.hpp:
  Comparing two literals was reported as undefined behaviour.
include/my_global.h:
  Add a template for aligning character buffers.
mysys/lf_alloc-pin.c:
  Initialize pointer.
sql/mysqld.cc:
  Use UNINIT_VAR rather than LINT_INIT.
sql/partition_info.cc:
  Use UNINIT_VAR rather than LINT_INIT.
sql/rpl_handler.cc:
  Use char[] rather than unsigned long[] array for placement buffer.
sql/spatial.cc:
  Use char[] rather than unsigned void*[] array for placement buffer.
sql/spatial.h:
  Use char[] rather than unsigned void*[] array for placement buffer.
sql/sql_partition.cc:
  Initialize auto variable.
sql/sql_table.cc:
  Initialize auto variables.
  Add parens around assignment within if()
sql/sys_vars.cc:
  Use UNINIT_VAR.
storage/innobase/os/os0file.c:
  Init first slot in auto variable.
storage/myisam/mi_create.c:
  Use UNINIT_VAR rather than LINT_INIT.
storage/myisam/mi_open.c:
  Remove (wrong) casting.
storage/myisam/mi_page.c:
  Remove (wrong) casting.
storage/myisam/mi_search.c:
  Cast to uchar* rather than char*.
strings/ctype-ucs2.c:
  Use UNINIT_VAR rather than LINT_INIT.
  Add (uchar*) casting.

Remove make-ccc files.

in mysql-trunk-merge).

There were two problems:
  - a mistake during merge of a patch for Bug 52629 from 5.1;
  - MTR treated auxilary output of newer valgrind as an error.

The fixes are:
  - Fix merge error;
  - Teach MTR to skip 'HEAP summary' section of valgrind output.

Also moved HA_DATA_PARTITION from ha_partition.cc to table.h.

Conflicts:

   conflict      Makefile.am
   conflict      mysql-test/suite/rpl/r/rpl_stm_mixing_engines.result
   conflict      mysql-test/suite/rpl/t/rpl_tmp_table_and_DDL.test
   conflict      sql/opt_sum.cc
   conflict      sql/set_var.cc
   conflict      sql/sql_base.cc
   conflict      sql/sql_priv.h
   conflict      sql/sql_show.cc

            data directory name command

The check_db_name function has been modified to validate tails of
#mysql50#-prefixed database names for compliance with MySQL 5.0
database name encoding rules (the check_table_name function call
has been reused).


mysql-test/r/renamedb.result:
  Updated test case.
mysql-test/r/upgrade.result:
  Test case for bug #53804.
mysql-test/t/renamedb.test:
  Updated test case.
mysql-test/t/upgrade.test:
  Test case for bug #53804.
sql/mysql_priv.h:
  Bug #53804: serious flaws in the alter database .. upgrade
              data directory name command
  
  The check_mysql50_prefix has been added.
sql/sql_table.cc:
  Bug #53804: serious flaws in the alter database .. upgrade
              data directory name command
  
  - The check_mysql50_prefix has been added.
  - The check_n_cut_mysql50_prefix function has been refactored
  to share code with new check_mysql50_prefix function.
sql/table.cc:
  Bug #53804: serious flaws in the alter database .. upgrade
              data directory name command
  
  The check_db_name function has been modified to validate tails of
  #mysql50#-prefixed database names for compliance with MySQL 5.0
  database name encoding rules.

to crash) from mysql-6.0-codebase.

The thing is that on some platforms (e.g. Mac OS X) sockaddr_in / sockaddr_in6
contain a non-standard field (sin_len / sin6_len), that must be set.

The problem was that only standard fields were set, thus getnameinfo() returned
EAI_SYSTEM instead of EAI_NONAME.

The fix is to introduce configure-time checks (for GNU auto-tools and CMake) for
those additional fields and to set them if they are available.

FOR UPDATE is causing a lock".
 
This patch tries to address problems which were exposed 
during backporting of original patch to 5.1 tree.
 
- It ensures that we don't change locking behavior of simple
  SELECT statements on InnoDB tables when they are executed
  under LOCK TABLES ... READ and with @@innodb_table_locks=0.
  Also we no longer pass TL_READ_DEFAULT/TL_WRITE_DEFAULT 
  lock types, which are supposed to be parser-only, to 
  handler::start_stmt() method.
- It makes check_/no_concurrent_insert.inc auxiliary scripts 
  more robust against changes in test cases that use them
  and also ensures that they don't unnecessarily change 
  environment of caller.

mysql-test/include/check_concurrent_insert.inc:
  Reset DEBUG_SYNC facility before and after using it in
  auxiliary script. This makes this script more robust against
  changes in test cases calling it. It also ensures that script
  does not unnecessarily change environment of caller.
mysql-test/include/check_no_concurrent_insert.inc:
  Reset DEBUG_SYNC facility before and after using it in
  auxiliary script. This makes this script more robust against
  changes in test cases calling it. It also ensures that script
  does not unnecessarily change environment of caller.
mysql-test/r/innodb-lock.result:
  Added coverage for LOCK TABLES ... READ behavior in
  @@innodb_table_locks = 0 mode. This test also checks
  that an appropriate type of lock is passed to
  handler::start_stmt() method.
mysql-test/t/innodb-lock.test:
  Added coverage for LOCK TABLES ... READ behavior in
  @@innodb_table_locks = 0 mode. This test also checks
  that an appropriate type of lock is passed to
  handler::start_stmt() method.
sql/sql_base.cc:
  Since we no longer set TL_READ as lock type for tables used  
  in simple SELECT right in the parser, in order to preserve  
  behavior for such statements on InnoDB tables when in  
  LOCK TABLES mode with @innodb_table_locks = 0,  
  check_lock_and_start_stmt() had to be changed to convert  
  TL_READ_DEFAULT to an appropriate type of read lock before  
  passing it to handler::start_stmt() method.  
  We do similar thing for TL_WRITE_DEFAULT as this lock type  
  is also supposed to be parser-only type.  
  As consequence read_lock_type_for_table() had to be  
  adjusted to behave properly when it is called from  
  check_lock_and_start_stmt() in prelocked mode.

Problem: The test case mysqldump reads a file that must
be world-readable. The test did not force the file to be
world-readable, so if the tree was branched with a umask
of 0077, the test would fail.
Fix: chmod the file.


mysql-test/t/mysqldump.test:
  Added chmod so that the file is guaranteed to be world-readable.

Conflicts:
  - mysql-test/r/partition.result
  - mysql-test/r/variables_debug.result
  - mysql-test/t/partition.test
  - mysql-test/t/variables_debug.test