GitLab

impossible view security".

We should not expose names of tables which are explicitly or implicitly (via
routine or trigger) used by view even if we find that they are missing.
So during building of list of prelocked tables for statement we track which
routines (and therefore tables for these routines) are used from views. We
mark elements of LEX::routines set which correspond to routines used in views
by setting Sroutine_hash_entry::belong_to_view member to point to TABLE_LIST
object for topmost view which uses routine. We propagate this mark to all
routines which are used by this routine and which we add to this set. We also
mark tables used by such routine which we add to the list of tables for
prelocking as belonging to this view.


mysql-test/r/sp-error.result:
  Added test for bug #11555 "Stored procedures: current SP tables locking make 
  impossible view security".
mysql-test/r/view.result:
  We should not expose tables which are expicitly/implicitly used in view in
  check table statement.
mysql-test/t/sp-error.test:
  Added test for bug #11555 "Stored procedures: current SP tables locking make 
  impossible view security".
mysql-test/t/view.test:
  Removed comment obsoleted by bugfix.
sql/sp.cc:
  We should not expose names of tables which are explicitly or implicitly
  (via routine or trigger) used by view even if we find that they are missing.
  So during building of list of prelocked tables for statement we track which
  routines (and therefore tables for these routines) are used from views. We
  mark elements of LEX::routines set which correspond to routines used in views
  by setting Sroutine_hash_entry::belong_to_view member to point to TABLE_LIST
  object for topmost view which uses routine. We propagate this mark to all
  routines which are used by this routine and which we add to this set. We also
  mark tables used by such routine which we add to the list of tables for
  prelocking as belonging to this view.
sql/sp.h:
  sp_cache_routines_and_add_tables_for_view()/for_triggers():
    To be able to determine correctly uppermost view which uses this view/table
    with trigger we have to pass pointer to TABLE_LIST object instead of pointer
    to view's LEX or to Table_triggers_list object.
sql/sp_head.cc:
  sp_head::add_used_tables_to_table_list():
    Added new argument which allows to mark tables which are added to table
    list for prelocking as belonging to view (this allows properly hide names
    of tables which are used in routines used by views).
sql/sp_head.h:
  sp_head::add_used_tables_to_table_list():
    Added new argument which allows to mark tables which are added to table
    list for prelocking as belonging to view (this allows properly hide names
    of tables which are used in routines used by views).
sql/sql_base.cc:
  open_tables():
    sp_cache_routines_and_add_tables_for_view()/for_triggers() now accept
    pointer to table list element as last argument, this allows them to determine
    correctly uppermost view which uses this view/table with trigger.
sql/sql_trigger.h:
  Table_triggers_list:
    sp_cache_routines_and_add_tables_for_triggers() now accept pointer to table
    list element as last argument, this allows to determine correctly uppermost
    view which uses this table with trigger.

Post-review fixes that simplify the way access rights
are checked during name resolution and factor out all
entry points to check access rights into one single
function.


sql/item.cc:
  Simplfied find_field_in_table - factored out all acces right checks
  into a separate function.
sql/mysql_priv.h:
  Simplified the way we control whether to perform access right
  checks for columns.
sql/sql_acl.cc:
  - Added new functon check_column_grant_in_table_ref that serves as a single
    point of entry to check access rights during name resolution for different
    kinds of table references.
  - Moved check_grant_column_in_sctx to sql_acl.cc where it
    logically belongs.
  - Removed the parameter check_grants - it is checked before
    calling the function.
sql/sql_acl.h:
  - Added new function check_column_grant_in_table_ref.
  - Made check_grant_column_in_sctx available to other modules.
sql/sql_base.cc:
  - Factored out all code that check access rights for columns during
    name resolution into one function - check_column_grant_in_table_ref.
  - Moved check_grant_column_in_sctx to sql_acl.cc where it
    logically belongs.
  - Removed the parameter check_grants - it is checked before
    calling the function.
sql/table.cc:
  Removed code that duplicates the functionality of
  check_column_grant_in_table_ref, and called directly
  that function.
sql/table.h:
  check_grants method is replaced by more general
  check_column_grant_in_table_ref.

Post-review fixes according to Monty's review.


sql/item.h:
  Unite all code that stores and restores the state of a name resolution context
  into a class to represent the state, and methods to save/restore that
  state.
sql/mysql_priv.h:
  Reorder parameters so that length is after the name of a field,
  and database is before table name.
sql/sql_acl.cc:
  Reorder parameters so that length is after the name of a field,
  and database is before table name.
sql/sql_base.cc:
  * Reorder parameters so that length is after the name of a field,
    and database is before table name.
  * Added new method - Field_iterator_table_ref::get_natural_column_ref
    to avoid unnecessary code when it is knwon that no new columns will
    be created when accessing natural join columns.
sql/sql_insert.cc:
  Unite all code that stores and restores the state of a name resolution context
  into a class to represent the state, and methods to save/restore that
  state.
sql/sql_lex.cc:
  Removed obsolete comment.
sql/sql_lex.h:
  Return error from push_contex() if there is no memory.
sql/sql_list.h:
  Extended base_list_iterator, List_iterator, and List_iterator_fast with an
  empty constructor, and init() methods, so that one doesn't have to construct
  a new iterator object every time one needs to iterate over a new list.
sql/sql_parse.cc:
  Moved common functionality from the parser into one function, and renamed
  the function to better reflect what it does.
sql/sql_yacc.yy:
  Moved common functionality from the parser into one function, and renamed
  the function to better reflect what it does.
sql/table.cc:
  * Extended base_list_iterator, List_iterator, and List_iterator_fast with an
    empty constructor, and init() methods, so that one doesn't have to construct
    a new iterator object every time one needs to iterate over a new list.
  * Added new method Field_iterator_table_ref::get_natural_column_ref to be
    used in cases when it is known for sure that no new columns should be
    created.
sql/table.h:
  - column_ref_it no longer allocated for each new list of columns
  - new method get_natural_join_column for faster/simpler access
    to natural join columns.

  Post-review version. Some minor review fixes, but also changed the way
  some errors are handled: Don't return specific parse errors; instead
  always use the more general "table corrupt" error (amended accordingly).


mysql-test/r/sp-destruct.result:
  Updated results.
mysql-test/r/sp-error.result:
  Updated for fully qualified name in "no return" error message.
mysql-test/t/sp-destruct.test:
  Adopted the more consistent error handling for a corrupted mysql.proc table.
  (No more "parse error" et al).
sql/share/errmsg.txt:
  Changed ER_SP_PROC_TABLE_CORRUPT to be more explicit.
sql/sp.cc:
  Review fixes.
  Changed the handling of parse errors, and added the routine name to the "table corrupt" error message.
sql/sql_base.cc:
  Review changes: Change error tests and added comments.
sql/sql_parse.cc:
  Mored ER_SP_NORETURN test of functions to sql_yacc.yy for more general error handling.
sql/sql_yacc.yy:
  Mored ER_SP_NORETURN test of functions from sql_parse.cc for more general error handling.

Larger stack size neaded for open table on x86 64 bit
Fix failing test cases
Deleted symlink from bk


BitKeeper/etc/ignore:
  added libmysqld/ha_blackhole.cc
BitKeeper/deleted/.del-ha_blackhole.cc~727c69ef7846623a:
  Delete: libmysqld/ha_blackhole.cc
include/my_global.h:
  Don't use PATH_MAX for FN_REFLEN as this uses too much stack space.
  (With a PATH_MAX of 4096, we use 80K for opening a table as there is several objects of size FN_REFLEN on stack)
mysql-test/r/federated.result:
  Update results after error message changes
mysql-test/r/grant.result:
  Update results after error message changes
mysql-test/r/grant2.result:
  Update results after error message changes
sql/ha_federated.cc:
  Fix error messages to be more consistent
sql/mysql_priv.h:
  Stack size to have when opening a table
  (This was needed on x86 64 bit Linux)
sql/share/errmsg.txt:
  Remove quotes around error string for federated as two quotes in the output looks stran...

client/mysqltest.c:
  An expected error messages hiding from the log if disable_result_log is in force.
mysql-test/r/sp-dynamic.result:
  The test expanded for case of allowed/disalowed recursion.
mysql-test/r/sp-error.result:
  Error messages changed.
  Test of bug11394() made with allowed recursion.
mysql-test/r/sp.result:
  Tests for recursion.
mysql-test/r/trigger.result:
  Check that triggers are not affected by this patch.
mysql-test/r/variables.result:
  Test of max_sp_recursion_depth variable.
mysql-test/t/sp-dynamic.test:
  The test expanded for case of allowed/disalowed recursion.
mysql-test/t/sp-error.test:
  Error messages changed.
  Test of bug11394() made with allowed recursion.
mysql-test/t/sp.test:
  Tests for recursion.
mysql-test/t/trigger.test:
  Check that triggers are not affected by this patch.
mysql-test/t/variables.test:
  Test of max_sp_recursion_depth variable.
sql/item_func.cc:
  sp_find_function() and sp_find_procedure() joined to sp_find_routine()
    function as it was mentioned in TODO.
sql/mysqld.cc:
  max_sp_recursion_depth variable added.
sql/set_var.cc:
  max_sp_recursion_depth variable added.
sql/share/errmsg.txt:
  An error message changed.
  An error message added.
sql/sp.cc:
  sp_find_function() and sp_find_procedure() joined to sp_find_routine()
    function as it was mentioned in TODO.
  Temory LEX is allocated on a stack, not on a heap.
  Recursion support added for stored procedures.
sql/sp.h:
  sp_find_function() and sp_find_procedure() joined to sp_find_routine()
    function as it was mentioned in TODO.
sql/sp_head.cc:
  Initialization of new sp_head fields to get correct list of instances
    contained one instance only.
  Stack requirement for SP instruction is increased.
  Stack free space is checked before mem root initialisation to avoid
    memory leak.
  Pointer to the free instance management added before and after
    SP execution.
sql/sp_head.h:
  New sp_head variables added to support inst of instances of SP
    for recursion and pointer on ths first free to use instance.
sql/sql_base.cc:
  open_table() consume a lot of stack space so we check free stack space before it.
sql/sql_class.h:
  max_sp_recursion_depth variable added.
sql/sql_parse.cc:
  sp_find_function() and sp_find_procedure() joined to sp_find_routine()
    function as it was mentioned in TODO.

mysql-test/r/view_grant.result:
  BUG#14726 test suite.
mysql-test/t/view_grant.test:
  BUG#14726 test suite.

Version for 5.0.
It fixes three problems:
1. The cause of the bug was that we did not check the table version for
 the HANDLER ... READ commands. We did not notice when a table was
 replaced by a new one. This can happen during ALTER TABLE, REPAIR
 TABLE, and OPTIMIZE TABLE (there might be more cases). I call the fix
 for this problem "the primary bug fix".
2. mysql_ha_flush() was not always called with a locked LOCK_open.
 Though the function comment clearly said it must.
 I changed the code so that the locking is done when required. I call
 the fix for this problem "the secondary fix".
3. In 5.0 (not in 4.1 or 4.0) DROP TABLE had a possible deadlock flaw in
 concur with FLUSH TABLES WITH READ LOCK. I call the fix for this
 problem "the 5.0 addendum fix".


include/my_pthread.h:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  Added a new macro for the 5.0 addendum fix.
mysql-test/r/handler.result:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  The test result.
mysql-test/t/handler.test:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  The test case.
sql/lock.cc:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  Changed a comment which did confuse me and which is not fully
  correct anymore after the 5.0 addendum fix.
  Added an assertion which would fire without the 5.0 addendum fix.
sql/mysql_priv.h:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  Changed a definition for the secondary fix.
sql/sql_base.cc:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  Changed function calls for the secondary fix.
sql/sql_class.cc:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  Changed a function call for the secondary fix.
sql/sql_handler.cc:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  The first two diffs make the primary bug fix.
  The rest is for the secondary fix.
sql/sql_table.cc:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  The first diff (four changed places) make the 5.0 addendum fix.
  The other three are changed function calls for the secondary fix.

Version for 4.0.
It fixes two problems:
1. The cause of the bug was that we did not check the table version for
   the HANDLER ... READ commands. We did not notice when a table was
   replaced by a new one. This can happen during ALTER TABLE, REPAIR
   TABLE, and OPTIMIZE TABLE (there might be more cases). I call the fix
   for this problem "the primary bug fix".
2. mysql_ha_flush() was not always called with a locked LOCK_open.
   Though the function comment clearly said it must.
   I changed the code so that the locking is done when required. I call
   the fix for this problem "the secondary fix".


mysql-test/r/handler.result:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  The test result.
mysql-test/t/handler.test:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  The test case.
sql/mysql_priv.h:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  Changed a definition for the secondary fix.
sql/sql_base.cc:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  Changed function calls for the secondary fix.
sql/sql_class.cc:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  Changed a function call for the secondary fix.
sql/sql_handler.cc:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  The first two diffs make the primary bug fix.
  The rest is for the secondary fix.
sql/sql_table.cc:
  Bug#14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
  Changed function calls for the secondary fix.

that in mysql_rm_table_part2_with_lock() previously we needed to open
same file twice. Now once is enough.


sql/mysql_priv.h:
  Merged functions get_table_type() and mysql_frm_type() into one,
  using the name from latter one.
sql/sql_base.cc:
  Changed get_table_type() to mysql_frm_type()
sql/sql_delete.cc:
  Changed get_table_type() to mysql_frm_type()
sql/sql_rename.cc:
  Changed get_table_type() to mysql_frm_type()
sql/sql_show.cc:
  Changed get_table_type() to mysql_frm_type()
sql/sql_table.cc:
  Changed get_table_type() to mysql_frm_type()
sql/sql_view.cc:
  Merged code from get_table_type() and mysql_frm_type() into the latter one.
sql/sql_view.h:
  Function prototype changes.
sql/table.cc:
  No longer needed.

large table gives server crash": make sure that when a MyISAM temporary
table is created for a cursor, it's created in its memory root,
not the memory root of the current query.


mysql-test/r/sp.result:
  Test results fixed: a test case for Bug#14210
mysql-test/t/sp.test:
  A test case for Bug#14210 "Simple query with > operator on large table 
  gives server crash"
sql/handler.cc:
  - rewrite get_new_handler to accept a memory root and use it for
sql/handler.h:
  - get_new_handler declaration changed
sql/opt_range.cc:
  - get_new_handler declaration changed
sql/sql_base.cc:
  - get_new_handler declaration changed
sql/sql_select.cc:
  - the actual fix for Bug#14210. In create_myisam_from_heap we should
  create the new table handler in TABLE::mem_root, not in THD::mem_root:
  the latter is freed shortly after cursor is open.
  - adjust create_tmp_table to explicitly supply &table->mem_root
  to get_new_handler when creating a handler for a new temporary table
sql/sql_table.cc:
  - get_new_handler declaration changed
sql/table.cc:
  - get_new_handler declaration changed
sql/unireg.cc:
  - get_new_handler declaration changed
tests/mysql_client_test.c:
  A test case for Bug#14210 "Simple query with > operator on large table
   gives server crash": a C API test case is worth adding because of different
  memory allocation/freeing patterns in handling of C API and SP cursors

Fix for bug #14536: SELECT @A,@A:=... fails with prepared statements


mysql-test/r/func_sapdb.result:
  Correct tests after reverting patch for BUG #14009 (use of abs() on null value causes problems with filesort)
mysql-test/r/type_newdecimal.result:
  Correct tests after reverting patch for BUG #14009 (use of abs() on null value causes problems with filesort)
mysql-test/r/user_var.result:
  More test with SELECT @A:=
mysql-test/t/disabled.def:
  Enable user_var.test for
mysql-test/t/user_var.test:
  More test with SELECT @A:=
sql/item.cc:
  Simple optimization during review of new code
sql/item_func.cc:
  Reverting patch for BUG #14009 (use of abs() on null value causes problems with filesort)
sql/item_timefunc.h:
  timediff() can return NULL for not NULL arguments
sql/sql_base.cc:
  Remove usage of current_thd() in mysql_make_view()
sql/sql_lex.h:
  Remove usage of current_thd() in mysql_make_view()
sql/sql_select.cc:
  Fix for bug #14536: SELECT @A,@A:=... fails with prepared statements
sql/sql_view.cc:
  Remove usage of current_thd() in mysql_make_view()
  Simple optimization of new code
sql/sql_view.h:
  Remove usage of current_thd() in mysql_make_view()
sql/table.cc:
  Simple optimization of new code

sql/item_func.cc:
  some optimisation
sql/sql_acl.cc:
  some optimisation
sql/sql_base.cc:
  some optimisation
sql/sql_parse.cc:
  some optimisation
sql/table.cc:
  some optimisation

mysql-test/r/information_schema.result:
  error message changed
mysql-test/r/sp.result:
  error message changed
mysql-test/r/sql_mode.result:
  fixed test suite
mysql-test/r/view.result:
  error message changed
mysql-test/r/view_grant.result:
  test of underlying view tables check
mysql-test/t/sql_mode.test:
  fixed test suite
mysql-test/t/view_grant.test:
  test of underlying view tables check
sql/item.cc:
  check of underlying tables privilege added
sql/item.h:
  Name the resolution context points to the security  context of view (if item belong to the view)
sql/item_func.cc:
  a view error hiding for execution of prepared function belonged to a view
  fixed checking privileges if stored functions belonds to some view
sql/mysql_priv.h:
  refult of derived table processing functions changed to bool
  Security_context added as an argument to find_field_in_table()
sql/share/errmsg.txt:
  error message fixed
sql/sq...

  Added error checking for errors when attempting to use stored procedures
  after the mysql.proc table has been dropped, corrupted, or tampered with.
  Test cases were put in a separate file (sp-destruct.test).


mysql-test/t/sp.test:
  Added comment.
sql/share/errmsg.txt:
  New error message for corrupted mysql.proc table.
sql/sp.cc:
  Check and return error code when caching stored routines.
  In the case when no error message has been set, set one.
sql/sp.h:
  Return error code from stored routine cache function.
sql/sql_base.cc:
  Check for error from sp_cache_routines_* calls.
sql/sql_trigger.h:
  Updated friend declaration for sp_cache_routines*.
mysql-test/r/sp-destruct.result:
  New test file for destruction of the mysql.proc table.
mysql-test/t/sp-destruct.test:
  New result file for destruction of the mysql.proc table.

mysql-test/r/sp-error.result:
  Results for the test case for BUG#13037.
mysql-test/t/sp-error.test:
  Test case for BUG#13037.
sql/sql_base.cc:
  Polishing: use constant.
sql/sql_class.cc:
  Reset THD::where in THD::cleanup_after_query();
  Polishing: use the constant (THD::DEFAULT_WHERE).
sql/sql_class.h:
  Introduce a constant for the default value of THD::where.

droping trigger on InnoDB table".

Deadlock occured in cases when we were trying to create two triggers for
the same InnoDB table concurrently and both threads were able to reach
close_cached_table() simultaneously. Bugfix implements new approach to
table locking and table cache invalidation during creation/dropping
of trigger.

No testcase is supplied since bug was repeatable only under high concurrency.


sql/mysql_priv.h:
  reopen_name_locked_table():
    Changed function signature to make it more robust against erroneous usage.
sql/sql_base.cc:
  reopen_name_locked_table():
    Changed function signature to make it more robust against erroneous usage.
    Obtaining LOCK_open lock is now resposibility of caller.
    When searching for the table to open we should not prefer connection's current
    database over database which was explicitly specified in TABLE_LIST::db member
    (even if database is not explicitly specified for t...

- CHAR() now returns binary string as default
- CHAR(X*65536+Y*256+Z) is now equal to CHAR(X,Y,Z) independent of the character set for CHAR()
- Test for both ETIMEDOUT and ETIME from pthread_cond_timedwait()
  (Some old systems returns ETIME and it's safer to test for both values
   than to try to write a wrapper for each old system)
- Fixed new introduced bug in NOT BETWEEN X and X
- Ensure we call commit_by_xid or rollback_by_xid for all engines, even if one engine has failed
- Use octet2hex() for all conversion of string to hex
- Simplify and optimize code


client/mysqldump.c:
  Simple optimizations of new code
  Indentation fixes
client/mysqltest.c:
  Removed not needed variable
include/mysql_com.h:
  Made octec2hex() more usable
mysql-test/r/ctype_utf8.result:
  CHAR() now returns binary string as default
mysql-test/r/func_str.result:
  CHAR() now returns binary string as default
mysql-test/r/range.result:
  Added test to verify new introduced bug in NOT BETWEEN X and X
mysq...

The problem was then when a column reference was resolved to a view column, the new Item
created for this column contained the name of the view, and not the view alias.


mysql-test/r/view.result:
  Additional test for BUG#13410.
mysql-test/t/view.test:
  Additional test for BUG#13410.
sql/item.cc:
  Correctly cast 'cur_field' to Item_ident because if the original item is
  an Item_field, the cur_field is either an Item_field or an Item_ref.
  Thus we have to cast cur_field to a common super-class of both.
sql/item.h:
  - real_item() may be called before Item_ref::ref is set
    (i.e. the Item_ref object was resolved).
  - To avoid a crash and to return some meaningful value
    in such cases we return 'this'.
sql/sql_base.cc:
  - 'item' may be an Item_ref, so we test for the type of the actual
    referenced item.
  - Correctly cast 'cur_field' to Item_ident because if the original
    item is an Item_field, the cur_field is either an Item_field or an
    Item_ref. Thus we have to cast cur_field to a common super-class
    of both.
sql/table.cc:
  - When creating a new Item for a reference to a view column, use the view alias,
    and not the real view name.
  - Removed old code

"SELECT ... FOR UPDATE executed as consistent read inside LOCK TABLES"
Do not discard lock_type information as handler::start_stmt() may require knowledge.
(fixed by Antony)

sql/sql_base.cc:
  fix for bug #13545: Server crash caused by select query.
  - compare table qualifier only with tables which are not nested joins.
  - perform recursion accordingly.
  - DBUG_ASSERT(table_list->table); added before the find_field_in_table() call.

  Fixed bug #13411.
  Fixed name resolution for non-qualified reference to a view column
  in the HAVING clause.
view.result, view.test:
  Added a test case for bug #13411.


mysql-test/t/view.test:
  Added a test case for bug #13411.
mysql-test/r/view.result:
  Added a test case for bug #13411.
sql/item.cc:
  Fixed bug #13411.
  Fixed name resolution for non-qualified reference to a view column
  in the HAVING clause.
sql/sql_base.cc:
  Fixed bug #13411.
  Fixed name resolution for non-qualified reference to a view column
  in the HAVING clause.

cmd-line-utils/readline/complete.c:
  Added a cast.
dbug/my_main.c:
  Added an include to avoid implicit declaration of
  my_thread_global_init()
include/my_global.h:
  undef cannot be used on this predefined name.
  Since it is a custom fix for gcc 2.8.0, let's make it only
  effective in that case.
include/my_sys.h:
  Added a new type, TYPE_NOT_SET.
myisam/ft_boolean_search.c:
  Added casts.
myisam/mi_key.c:
  Added cast.
myisam/mi_open.c:
  Added cast.
  Changed function types.
myisam/mi_test1.c:
  Added cast.
myisam/myisamchk.c:
  Added cast.
myisam/myisamdef.h:
  Changed function type.
myisam/myisampack.c:
  Added casts.
myisam/sp_key.c:
  Added cast.
mysys/mf_iocache.c:
  Fixed invalid use of 0 to info->type. According to comment it
  should not have been set, but in earlier code by setting it
  to 0 would have been same as setting it to READ_CACHE. This
  probably was not desired, potential bug.
server-tools/instance-manager/instance_options.cc:
  Fixed a typo.
server-tools/instance-manager/protocol.cc:
  Changed enum to int.
  Changed char to uchar.
  Added casts.
sql/mysql_priv.h:
  Bit overflow.
sql/sql_base.cc:
  Removed unused label. The code below label was unused too, because
  there is a return just before.
sql/sql_parse.cc:
  Removed unneccessary extra argument.

Fixed problems in test suite where some test failed
Fixed access to not initialized memory in federated
Fixed access to not initialized memory when using BIT fields in internal temporary tables


BitKeeper/etc/ignore:
  added libmysqld/sql_cursor.h
mysql-test/r/information_schema.result:
  Change view names to 'v#' to not cause massive conflict with other tests if test dies in the middlecd
mysql-test/r/information_schema_inno.result:
  Remove used tables at start
mysql-test/r/multi_statement.result:
  Remove used tables at start
mysql-test/r/temp_table.result:
  Change view names to 'v#' to not cause massive conflict with other tests if test dies in the middle
mysql-test/r/type_bit.result:
  More tests
mysql-test/t/information_schema.test:
  Change view names to 'v#' to not cause massive conflict with other tests if test dies in the middle
mysql-test/t/information_schema_inno.test:
  Remove used tables at start
mysql-test/t/multi_statement.test:
  Remove used tables at star...

The problem was in the way table references are pre-filtered when
resolving a qualified field. When resolving qualified table references
we search recursively in the operands of the join. If there is
natural/using join with a merge view, the first call to find_field_in_table_ref
makes a recursive call to itself with the view as the new table reference
to search for the column. However the view has both nested_join and
join_columns != NULL so it skipped the test whether the view name matches
the field qualifier. As a result the field was found in the view since the
view already has a field with the same name. Thus the field was incorrectly
resolved as the view field.


mysql-test/r/select.result:
  Test for BUG#13127.
mysql-test/t/select.test:
  Test for BUG#13127.
sql/sql_base.cc:
  The patch contains two independent changes:
  
  - When searching for qualified fields, include merge views and information schema tables
  used with SHOW (both using TABLE_LIST::field_translation to represent result fields) in
  the test that compares the qualifying table with the name of the table reference being
  searched. This change fixes the bug.
  
  - Do not search the materialized list of columns of a NATURAL/USING join if 'table_list'
  is a stored table or merge view. Instead search directly in the table or view as if it
  is not under a natural join. This is a performance improvement since if 'table_list'
  is a stored table, then the search can utilize the name hash for table names.

libmysqld/lib_sql.cc:
  changed securety context switching
mysql-test/r/rpl_sp.result:
  now it show real information from changed security context of SP (checked)
sql/ha_innodb.cc:
  changed securety context switching
sql/item.cc:
  changed securety context switching
sql/item_func.cc:
  changed securety context switching
sql/item_strfunc.cc:
  changed securety context switching
sql/log.cc:
  changed securety context switching
sql/mysql_priv.h:
  changed securety context switching
sql/mysqld.cc:
  changed securety context switching
sql/repl_failsafe.cc:
  changed securety context switching
sql/set_var.cc:
  changed securety context switching
sql/slave.cc:
  changed securety context switching
sql/sp.cc:
  changed securety context switching
sql/sp_head.cc:
  changed securety context switching
  in case of inability to switch context  we return error now
sql/sp_head.h:
  changed securety context switching
sql/sql_acl.cc:
  changed securety context switching
sql/sql_acl.h:
  changed securety context switching
sql/sql_base.cc:
  changed securety context switching
sql/sql_class.cc:
  changed securety context switching
sql/sql_class.h:
  changed securety context switching
sql/sql_db.cc:
  changed securety context switching
sql/sql_insert.cc:
  changed securety context switching
sql/sql_parse.cc:
  changed securety context switching
sql/sql_show.cc:
  changed securety context switching
sql/sql_trigger.cc:
  changed securety context switching
sql/sql_view.cc:
  changed securety context switching
sql/sql_yacc.yy:
  changed securety context switching

This bug occurs when some trigger for table used by DML statement is created
or changed while statement was waiting in lock_tables(). In this situation
prelocking set which we have calculated becames invalid which can easily lead
to errors and even in some cases to crashes.

With proposed patch we no longer silently reopen tables in lock_tables(),
instead caller of lock_tables() becomes responsible for reopening tables and
recalculation of prelocking set.


mysql-test/t/trigger.test:
  Added tests for bug #12704 "Server crashes during trigger execution".
  Unfortunately these tests rely on the order in which tables are locked
  by statement so they are non-determenistic and therefore should be disabled.
sql/lock.cc:
  mysql_lock_tables():
    Now instead of always reopening altered or dropped tables by itself
    mysql_lock_tables() can notify upper level and rely on caller doing this.
sql/mysql_priv.h:
  Now mysql_lock_tables() can either reopen deleted or altered tables by itself
  or notify caller about such situation through 'need_reopen' argument and rely
  on it in this.
  Also lock_tables() has new 'need_reopen' out parameter through which it
  notifies caller that some tables were altered or dropped so he needs to reopen
  them (and rebuild prelocking set some triggers may change or simply appear).
sql/sp.cc:
  sp_add_used_routine():
    To be able to restore LEX::sroutines_list to its state right after parsing
    we now adjust  LEX::sroutines_list_own_last/sroutines_list_own_elements when
    we add directly used routine.
  sp_remove_not_own_routines():
    Added procedure for restoring LEX::sroutines/sroutines_list to their state
    right after parsing (by throwing out non-directly used routines).
  sp_cache_routines_and_add_tables_for_view()/sp_update_stmt_used_routines():
    We should use LEX::sroutines_list instead of LEX::sroutines as source of
    routines used by view, since LEX::sroutines is not availiable for view
    on second attempt to open it (see comment in open_tables() about it).
sql/sp.h:
  sp_remove_not_own_routines():
    Added procedure for restoring LEX::sroutines/sroutines_list to their state
    right after parsing (by throwing out non-directly used routines).
sql/sp_head.cc:
  Removed assert which is no longer always true.
sql/sql_base.cc:
  reopen_table():
    When we re-open table and do shallow copy of TABLE object we should adjust
    pointers to it in associated Table_triggers_list object. Removed nil
    operation.
  open_tables():
    Now this function is able to rebuild prelocking set for statement if it is
    needed. It also correctly handles FLUSH TABLES which may occur during its
    execution.
  lock_tables():
    Instead of allowing mysql_lock_tables() to silently reopen altered or dropped
    tables let us notify caller and rely on that it will do reopen itself.
    This solves the problem when trigger suddenly appears or changed during
    mysq_lock_tables().
  close_tables_for_reopen():
    Added routine for properly preparing for reopening of tables and recalculation
    of set of prelocked tables.
sql/sql_handler.cc:
  Here we let mysql_lock_tables() to reopen dropped or altered tables by itself.
sql/sql_insert.cc:
  Here we let mysql_lock_tables() to reopen dropped or altered tables by itself.
sql/sql_lex.cc:
  LEX:
    Added 'sroutines_list_own_last' and 'sroutines_list_own_elements' members
    which are used for keeping state in which 'sroutines_list' was right after
    statement parsing (and for restoring of this list to this state).
sql/sql_lex.h:
  LEX:
    Added 'sroutines_list_own_last' and 'sroutines_list_own_elements' members
    which are used for keeping state in which 'sroutines_list' was right after
    statement parsing (and for restoring of this list to this state).
    Added chop_off_not_own_tables() method to simplify throwing away list
    of implicitly used (prelocked) tables.
sql/sql_prepare.cc:
  Now instead of silently reopening altered or dropped tables in
  lock_tables() we notify caller and rely on that the caller will
  reopen tables.
sql/sql_table.cc:
  Here we let mysql_lock_tables() to reopen dropped or altered tables by itself.
sql/sql_trigger.cc:
  Added Table_triggers_list::set_table() method to adjust Table_triggers_list
  to new pointer to TABLE instance.
sql/sql_trigger.h:
  Added Table_triggers_list::set_table() method to adjust Table_triggers_list
  to new pointer to TABLE instance.
sql/sql_update.cc:
  Now instead of silently reopening altered or dropped tables in
  lock_tables() we notify caller and rely on that the caller will
  reopen tables.

Solution for 4.0 and 4.1.
If the caller cannot re-open table(s), it gives a NULL
'refresh' argument to open_table(). We used to ignore
flushes then. Now we ignore drops too.

Solution for 5.0.
Changed calls to open_table(). Requested to ignore
flush at places where the command did already lock tables.
This could happen in CREATE ... SELECT and ALTER TABLE.

No test case. The bug can only be triggered by true concurrency.
The stress test suite provides a test case for this.


sql/sql_base.cc:
  Bug#12845 - Stress test: Server crashes on  CREATE .. SELECT statement
  Solution for 5.0.
  Changed open_table() so that ignoring flush requests
  is dependend on the 'flags' parameter as ignoring drop
  request was already. This aims for consistent behaviour.
sql/sql_table.cc:
  Bug#12845 - Stress test: Server crashes on  CREATE .. SELECT statement
  Solution for 5.0.
  Changed calls to open_table(). Requested to ignore
  flush at places where the command did already lock tables.
  This could happen in CREATE ... SELECT and ALTER TABLE.

mysql-test/r/select.result:
  Test for BUG#13067
mysql-test/t/select.test:
  Test for BUG#13067
sql/sql_base.cc:
  Correctly compare field names with respect to case sensitivity.

When view column aliased in subselect alias is set on ref which represents
field. When tmp table is created for subselect, it takes name of original field
not ref. Because of this alias on view column in subselect is lost. Which
results in reported error.

Now when alias is set on ref, it's set on ref real item too.


mysql-test/r/view.result:
  Test case for bug #12993  View column rename broken in subselect
mysql-test/t/view.test:
  Test case for bug #12993  View column rename broken in subselect
sql/sql_base.cc:
  Fix bug #12993 View column rename broken in subselect
  Now when alias is set on ref, it's set on ref real item too.

are thus dangling later)


mysql-test/r/ps.result:
  test for bug #12651
  (data allocated on thd's arena but not on permanent arena)
mysql-test/t/ps.test:
  test for bug #12651
  (data allocated on thd's arena but not on permanent arena)
sql/sql_base.cc:
  if there is tree transformation then backup the current arena
  and use permanent one (for PS) otherwise the data will be deallocated
  after the prepare process is finished.
  this bug was introduced with the recent natural join patch

mysql-test/r/select.result:
  Test for BUG#12977.
mysql-test/t/select.test:
  Test for BUG#12977.
sql/sql_base.cc:
  - Compare table qualifier of qualified fields only with tables that
    are not natural joins or their operands.
  - For qualified fields perform recursive search in all operands of
    natural joins that are nested joins.
  - Symmetrically detect ambiguous columns for both operands of
    NATURAL/USING joins.

- current_arena to stmt_arena: the thread may have more than one
'current' arenas: one for runtime data, and one for the parsed 
tree of a statement. Only one of them is active at any moment.
- set_item_arena -> set_query_arena, because Item_arena was renamed to 
Query_arena a while ago
- set_n_backup_item_arena -> set_n_backup_active_arena;
the active arena is the arena thd->mem_root and thd->free_list
are currently pointing at.
- restore_backup_item_arena -> restore_active_arena (with the same
rationale)
- change_arena_if_needed -> activate_stmt_arena_if_needed; this
method sets thd->stmt_arena active if it's not done yet.


sql/item.cc:
  Rename.
sql/item_cmpfunc.cc:
  Rename.
sql/item_func.cc:
  Rename.
sql/item_subselect.cc:
  Rename.
sql/item_subselect.h:
  Remove an unused forward declaration.
sql/item_sum.h:
  Remove an unused forward declaration.
sql/mysql_priv.h:
  Remove an unused forward declaration.
sql/sp.cc:
  Rename.
sql/sp_head.cc:
  Rename.
sql/sql_base.cc:
  Rename.
sql/sql_class.cc:
  Rename.
sql/sql_class.h:
  Rename.
sql/sql_lex.cc:
  Rename.
sql/sql_parse.cc:
  Rename.
sql/sql_prepare.cc:
  Rename.
sql/sql_select.cc:
  Rename.
sql/sql_show.cc:
  Rename.
sql/sql_union.cc:
  Rename.
sql/sql_view.cc:
  Rename.
sql/table.cc:
  Rename.

mysql-test/t/lowercase_view.test:
  error code changed after merege
mysql-test/t/view.test:
  error code changed after merege
sql/item.cc:
  inline function used instead of frequently used expression
  removed old TODO
sql/sql_base.cc:
  layoutfixed
  inline function used instead of frequently used expression
sql/sql_insert.cc:
  inline function used instead of frequently used expression

Removed ASSERT that can obviously never be wrong

sql/sql_base.cc:
  Fixed a bug in PS (prepared statements) mode - the "always TRUE" constant
  Item created for outer joins with no common columns was not created in
  the PS memory.
  Now it is created in the right memory.

mysql-test/r/select.result:
  Test for BUG#6276.
mysql-test/t/select.test:
  Test for BUG#6276.
sql/sql_base.cc:
  Add a true ON condition for outer joins without common columns.

* Provide backwards compatibility extension to name resolution of
  coalesced columns. The patch allows such columns to be qualified
  with a table (and db) name, as it is in 4.1.
  Based on a patch from Monty.

* Adjusted tests accordingly to test both backwards compatible name
  resolution of qualified columns, and ANSI-style resolution of
  non-qualified columns.
  For this, each affected test has two versions - one with qualified
  columns, and one without. 


mysql-test/include/ps_query.inc:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/bdb.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/innodb.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/join.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/join_nested.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/join_outer.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/null_key.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/order_by.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/ps_2myisam.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/ps_3innodb.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/ps_4heap.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/ps_5merge.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/ps_6bdb.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/ps_7ndb.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/select.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/subselect.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/r/type_ranges.result:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/t/bdb.test:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/t/innodb.test:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/t/join.test:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/t/join_nested.test:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/t/join_outer.test:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/t/null_key.test:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/t/order_by.test:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/t/select.test:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/t/subselect.test:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
mysql-test/t/type_ranges.test:
  Put back old tests to test that coalesced columns of natural joins can be qualified.
sql/sql_base.cc:
  * Applied Monty's patch for backwards compatible name resolution
    of qualified columns. The idea is:
    - When a column is qualified, search for the column in all
      tables/views underlying each natural join. In this case
      natural joins are *not* considered leaves.
    - If a column is not qualified, then consider natural joins
      as leaves, thus directly search the result columns of
      natural joins.
  * Simplified 'find_field_in_tables()' - unified two similar
    loops into one.
sql/table.cc:
  - Removed method & members not needed after Monty's patch.
sql/table.h:
  - Removed method & members not needed after Monty's patch.
tests/mysql_client_test.c:
  Put back old tests to test that coalesced columns of natural joins can be qualified.