crashes server (prepared statements)": the bug was that all boolean
items always recovered its original arguments at statement cleanup
stage.
This collided with Item_subselect::select_transformer, which tries to
permanently change the item tree to use a transformed subselect instead of
original one.
So we had this call sequence for prepare:
mysql_stmt_prepare -> JOIN::prepare ->
Item_subselect::fix_fields -> the item tree gets transformed ->
Item_bool_rowready_func2::cleanup, item tree is recovered to original
state, while it shouldn't have been;
mysql_stmt_execute -> attempts to execute a broken tree -> crash.
Now instead of bluntly recovering all arguments of bool functions in
Item_bool_rowready_func2::cleanup, we recover only those
which were changed, and do it in one place.
There still would exist a possibility for a collision with subselect
tranformation, if permanent and temporary changes were performed at the
same stage.
But fortunately subselect transformation is always done first, so it
doesn't conflict with the optimization done by propogate_cond_constants.
Now we have:
mysql_stmt_prepare -> JOIN::prepare -> subselect transformation
permanently changes the tree -> cleanup doesn't recover anything,
because nothing was registered for recovery.
mysql_stmt_execute -> JOIN::prepare (the tree is already transformed,
so it doesn't change), JOIN::optimize ->
propogate_cond_constants -> temporary changes the item tree
with constants -> JOIN::execute -> cleanup ->
the changes done by propogate_cond_constants are recovered, as
they were registered for recovery.
mysql-test/r/ps.result:
Bug#5987: test results fixed.
mysql-test/t/ps.test:
A test for bug#5987 "subselect in bool function crashes server
(prepared statements)"
sql/item.cc:
resolve_const_item is now responsible to register all changes of the
item tree for recovery
sql/item.h:
resolve_const_item signagture changed
sql/item_cmpfunc.h:
Arguments of boolean functions are now recovered using the
centralized registry of THD.
sql/sql_class.cc:
It's crucial to add new items to the beginning of the recovery list,
so that the recovery is performed in LIFO mode: otherwise if we
change one node of a tree twice, it will be recovered to some intermediate
state.
sql/sql_select.cc:
change_cond_ref_to_const and propogate_cond_constants are now responsible
to register all changes of the item tree for recovery.
The recovery is done using the centralized THD registry of
changed tree items.
crashes mysqld": implementation for a generic item tree modifications
registry. Every item tree modification which should be rolled back for
subsequent execution of a prepared statement or stored procedure should
be saved in the registry. All such modifications are rolled back at once
during cleanup stage of PS.
Actual fix for the bug just adds a call to register modifications to
convert_constant_item.
Post review fixes implemented.
mysql-test/r/ps.result:
A fix for bug#5748, test results fixed.
mysql-test/t/ps.test:
A test case for Bug#5748 "Prepared statement with BETWEEN and bigint
values crashes mysqld"
sql/item.cc:
Fix for Bug#5748 "Prepared statement with BETWEEN and bigint values
crashes mysqld":
First step in removing up item-specific cleanups: now all such
tree modifications should be done using the genericm mechanism implemented
in this changeset.
sql/item.h:
Fix for Bug#5748 "Prepared statement with BETWEEN and bigint values
crashes mysqld": no need for an item-specific change record any more.
sql/item_cmpfunc.cc:
A fix for Bug#5748 "Prepared statement with BETWEEN and bigint
values crashes mysqld": register item tree transformation performed by
convert_constant_item.
sql/sql_class.cc:
Implementation for item tree transformations registry.
sql/sql_class.h:
Declarations, necessary for the tree transformations registry.
sql/sql_parse.cc:
Assert that the item tree transformations registry is not used for
conventional execution.
sql/sql_prepare.cc:
Use of the item tree modifications registry in prepared statements:
rollback all modifications in the end of statement prepare and execute.
Also we now always set thd->current_arena to be able to determine that
this is an execution of prepared statement inside the registry code.
tests/client_test.c:
A typo fixed.
mysql-test/r/ps.result:
Test results fixed: the test case for bug#5688 "Upgraded 4.1.5 Server
seg faults"
mysql-test/t/ps.test:
Test case for bug#5688 "Upgraded 4.1.5 Server seg faults"
sql/item_cmpfunc.cc:
A fix for bug#5688 "Upgraded 4.1.5 Server seg faults":
fix just another place where we use wrong memory root for an Item
in statement prepare.
In addition, make the check for charsets in Item_bool_func2
more generic (fixes the test case when we use LIKE to compare BLOBs
with TEXT data).
key Column Fails".
mysql-test/r/ps.result:
Test results updated: a test case for Bug#5510 "inserting Null in
AutoIncrement primary key Column Fails".
mysql-test/r/ps_2myisam.result:
Bug#5510: a different warning in case of NULL->default truncation.This is
OK, the new warning is the same as produced by conventional execution.
mysql-test/r/ps_3innodb.result:
Bug#5510: a different warning in case of NULL->default truncation.This is
OK, the new warning is the same as produced by conventional execution.
mysql-test/r/ps_4heap.result:
Bug#5510: a different warning in case of NULL->default truncation.This is
OK, the new warning is the same as produced by conventional execution.
mysql-test/r/ps_5merge.result:
Bug#5510: a different warning in case of NULL->default truncation.This is
OK, the new warning is the same as produced by conventional execution.
mysql-test/r/ps_6bdb.result:
Bug#5510: a different warning in case of NULL->default truncation.This is
OK, the new warning is the same as produced by conventional execution.
mysql-test/t/ps.test:
A test case for Bug#5510 "inserting Null in AutoIncrement primary key Column
Fails".
sql/item.cc:
A fix for bug#5510 "inserting Null in AutoIncrement primary key Column
Fails": use proper Field API function for NULL placholeders.
with a fixed constant to pass on QNX:
max_data_file_length is calculated depending on max possible value of
record pointer. It's different on QNX, so the difference in results is OK.
mysql-test/t/ps.test:
A quick fix for bug#5556: replace a column of SHOW TABLE STATUS result
with a fixed constant to pass on QNX:
max_data_file_length is calculated depending on max value of record
pointer. It's different on QNX, so the result difference is intentional.
server': the bug occurs when arguments of LIKE function are in
differentcharacter sets. If these character sets are compatible,
we create an item-converter. In prepared mode, this item
needs to be created in memory of current prepared statement.
mysql-test/r/ps.result:
Test for Bug#4368 added.
mysql-test/t/ps.test:
A test case for bug#4368 '"like" fails in PreparedStatement,
crashes server'.
execute crashes server": we were deleting lex->result
after each execute, but prepared statements assumed that
it's left intact.
The fix adds cleanup() method to select_result hierarchy,
so that result objects can be reused.
Plus we now need to delete result objects more wisely.
mysql-test/r/ps.result:
Test results fixed: test case for bug#5034
mysql-test/t/ps.test:
A test case for bug#5034, few followups
sql/sql_class.cc:
- fix warning in THD::THD
- implementation of cleanup() for select_result hierarchy
- select_export::send_eof was identical to
select_dump::send_eof: moved to the base class select_to_file.
- Statement::end_statement() to end lex, free items, and
delete possible select_result
sql/sql_class.h:
- select_result::cleanup() declaration
-
sql/sql_insert.cc:
- implementation of select_insert::cleanup(): currently
we always create a new instance of select_insert/
select_create on each execute.
sql/sql_lex.cc:
- with more complicated logic of freeing lex->result it's
easier to have it non-zero only if it points to a valid
result.
sql/sql_lex.h:
Now st_lex::st_lex is not empty.
sql/sql_parse.cc:
mysql_execute_command():
- delete select_result *result only if it was created in
this function.
- use end_statement() to cleanup lex and thd in the end of
each statement.
- no need to save THD::lock if this is explain. This save
apparently left from times when derived tables were
materialized here, not in open_and_lock_tables.
sql/sql_prepare.cc:
- call result->cleanup() in reset_stmt_for_execute
- now Statement is responsible for freeing its lex->result.
sql/sql_select.cc:
handle_select():
- don't delete result, it might be needed
for next executions
- result is never null
a second time". The bug was caused by incompatibility of
negations elimination algorithm and PS: during first statement
execute a subtree with negation was replaced with equivalent
subtree without NOTs.
The problem was that although this transformation was permanent,
items of the new subtree were created in execute-local memory.
The patch adds means to check if it is the first execute of a
prepared statement, and if this is the case, to allocate items
in memory of the prepared statement.
The implementation:
- backports Item_arena from 5.0
- adds Item_arena::is_stmt_prepare(),
Item_arena::is_first_stmt_execute().
- deletes THD::allocate_temporary_pool_for_ps_preparing(),
THD::free_temporary_pool_for_ps_preparing(); they
were redundant.
and adds a few invariants:
- thd->free_list never contains junk (= freed items)
- thd->current_arena is never null. If there is no
prepared statement, it points at the thd.
The rest of the patch contains mainly mechanical changes and
cleanups.
mysql-test/r/ps.result:
Test results updated (test case for Bug#4912)
mysql-test/t/ps.test:
A test case for Bug#4912 "mysqld crashs in case a statement is
executed a second time"
sql/item_cmpfunc.cc:
current_statement -> current_arena
sql/item_subselect.cc:
Statement -> Item_arena, current_statement -> current_arena
sql/item_subselect.h:
Item_subselect does not need to save thd->current_statement.
sql/item_sum.cc:
Statement -> Item_arena
sql/item_sum.h:
Statement -> Item_arena
sql/mysql_priv.h:
Statement -> Item_arena
sql/sql_base.cc:
current_statement -> current_arena
sql/sql_class.cc:
- Item_arena
- convenient set_n_backup_statement, restore_backup_statement
(nice idea, Sanja)
sql/sql_class.h:
- Item_arena: backport from 5.0
- allocate_temporary_pool_for_ps_preparing,
free_temporary_pool_for_ps_preparing removed.
sql/sql_derived.cc:
current_statement -> current_arena
sql/sql_lex.cc:
current_statement -> current_arena
sql/sql_parse.cc:
Deploy invariant that thd->free_list never contains junk items
(backport from 5.0).
sql/sql_prepare.cc:
- backporting Item_arena
- no need to allocate_temporary_pool_for_ps_preparing().
sql/sql_select.cc:
Fix for bug#4912 "mysqld crashs in case a statement is
executed a second time": if this is the first execute of
a prepared statement, negation elimination is
done in memory of the prepared statement.
sql/sql_union.cc:
Backporting Item_arena from 5.0.
mysql-test/r/func_group.result:
fixed
mysql-test/r/ps.result:
SHOW TABLE STATUS test
mysql-test/t/ps.test:
SHOW TABLE STATUS test
sql/sql_parse.cc:
Use lex->describe instead of lex->select_lex.options
sql/sql_yacc.yy:
removed unneed initialization of lex->select_lex.options
mysql-test/r/ps.result:
parameters from variables (for field creation)
mysql-test/t/ps.test:
parameters from variables (for field creation)
sql/item.cc:
type of parameter assignment for parameters from variables
sql/item.h:
removed spaces at line ends in Item_param definition
mysql-test/r/ps.result:
test of prepared EXPLAIN
mysql-test/t/ps.test:
fix layout and table removing of previous tests
test of prepared EXPLAIN
sql/sql_prepare.cc:
explain pave to be prepared, too
cleanup options after explain execution
sql/sql_select.cc:
removed unneed assignment
mysql-test/r/ps.result:
New test case for BUG#3486.
mysql-test/t/ps.test:
New test case for BUG#3486.
sql/item_create.cc:
Create an Item_func_found_rows() at parse time, not an Item_int.
sql/item_func.cc:
Added val_int() method for new Item_func_found_rows class.
sql/item_func.h:
New class Item_func_found_rows for FOUND_ROWS() function.
sql/sql_select.cc:
Don't reset thd->limit_found_rows too early, or FOUND_ROWS() wont work.
character set introducer": add new item type to be returned before from
Item_param until it's value is set.
This way items like Item_bool_func2 and udf_handler won't treat this item
as constant literal when statement is prepared.
mysql-test/r/ps.result:
Test results fixed (test case for bug #4105)
mysql-test/t/ps.test:
Followup to bug #4105: a test case.
sql/item.cc:
Fix for bug#4105 "Server crash on attempt to prepare a statement with
character set introducer": add new item type to be returned before from
Item_param until it's value is set.
sql/item.h:
Fix for bug#4105 "Server crash on attempt to prepare a statement with
character set introducer": add new item type to be returned before from
Item_param until it's value is set.
mysql-test/r/ps.result:
Added tests for PREPARE stmt1 FROM @var syntax
mysql-test/t/ps.test:
Added tests for PREPARE stmt1 FROM @var syntax
mysys/my_error.c:
Added support for "%.*s" format
sql/item.cc:
Removed one redundant Item_param::set_value function
sql/item.h:
Removed one redundant Item_param::set_value function
sql/mysqld.cc:
Reformmated the code
sql/share/czech/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/dutch/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/english/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/estonian/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/french/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/german/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/greek/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/hungarian/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/italian/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/japanese/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/korean/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/norwegian-ny/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/norwegian/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/polish/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/portuguese/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/romanian/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/russian/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/slovak/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/spanish/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/swedish/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/share/ukrainian/errmsg.txt:
Changed ER_UNKNOWN_STMT_HANDLER format string
sql/sql_class.h:
SQL Prepared statements now can't be used by binary protocol commands
sql/sql_lex.h:
Added support for PREPARE stmt1 FROM @var syntax.
sql/sql_parse.cc:
Added support for PREPARE stmt1 FROM @var syntax.
sql/sql_prepare.cc:
Code cleanup
sql/sql_yacc.yy:
Added support for PREPARE stmt1 FROM @var syntax.
mysql-test/r/ps.result:
Added check if multiple SQL statements inside a PS are disabled
mysql-test/t/ps.test:
Added check if multiple SQL statements inside a PS are disabled
