execution. Failure to do so caused the erroneous statements to send nothing
and hang the client.
mysql-test/r/sp-error.result:
Testcase for BUG#9814. Note that the result demonstrates that currently
mysql-test-run ignores errors in multi-statement if they arrive after first
resultset has been received.
mysql-test/t/sp-error.test:
Testcase for BUG#09814.
"Stored procedures: crash with function calling itself".
Disallow recursive stored routines until we either make Item's and LEX
reentrant safe or will use spearate sp_head instances (and thus separate
LEX objects and Item trees) for each routine invocation.
mysql-test/r/sp-error.result:
Added tests for bug #11394 "Recursion in SP crash server" and
bug #11600 "Stored procedures: crash with function calling itself".
(We simply disallow recursion for stored routines).
mysql-test/r/sp.result:
Disabled test cases containing recursive stored routines until we will
support for them.
mysql-test/t/sp-error.test:
Added tests for bug #11394 "Recursion in SP crash server" and
bug #11600 "Stored procedures: crash with function calling itself".
(We simply disallow recursion for stored routines).
mysql-test/t/sp.test:
Disabled test cases containing recursive stored routines until we will
support for them.
sql/share/errmsg.txt:
Added error message saying that recursive stored routines are disallowed.
sql/sp_head.cc:
sp_head::execute():
Since many Item's and LEX members can't be used in reentrant fashion
we have to disable recursion for stored routines. So let us track
routine invocations using sp_head::m_is_invoked member and raise
error when one attempts to call routine recursively.
sql/sp_head.h:
sp_head:
Added m_is_invoked member for tracking of routine invocations and
preventing recursion.
"Stored procedures: crash with function calling itself".
Disallow recursive stored routines until we either make Item's and LEX
reentrant safe or will use spearate sp_head instances (and thus separate
LEX objects and Item trees) for each routine invocation.
include/my_bitmap.h:
new bitmap operation
mysql-test/r/view.result:
added warnings
Correct inserting data check (absence of default value) for view underlying tables (BUG#6443)
mysql-test/t/view.test:
Correct inserting data check (absence of default value) for view underlying tables (BUG#6443)
mysys/my_bitmap.c:
new bitmap operation
sql/field.h:
index of field in table added
sql/item.cc:
Name resolution context added
table list removed from fix_fields() arguments
sql/item.h:
Name resolution context added
table list removed from fix_fields() arguments
sql/item_cmpfunc.cc:
table list removed from fix_fields() arguments
sql/item_cmpfunc.h:
table list removed from fix_fields() arguments
sql/item_func.cc:
table list removed from fix_fields() arguments
sql/item_func.h:
table list removed from fix_fields() arguments
sql/item_row.cc:
table list removed from fix_fields() arguments
sql/item_row.h:
table list removed from fix_fields() arguments
sql/item_strfunc.cc:
fixed server crash on NULL argument
sql/item_strfunc.h:
table list removed from fix_fields() arguments
sql/item_subselect.cc:
table list removed from fix_fields() arguments
sql/item_subselect.h:
table list removed from fix_fields() arguments
sql/item_sum.cc:
table list removed from fix_fields() arguments
sql/item_sum.h:
table list removed from fix_fields() arguments
sql/item_timefunc.cc:
table list removed from fix_fields() arguments
sql/item_timefunc.h:
table list removed from fix_fields() arguments
sql/item_uniq.h:
table list removed from fix_fields() arguments
sql/log_event.cc:
Name resolution context added
sql/log_event.h:
Name resolution context added
sql/mysql_priv.h:
Name resolution context added
sql/set_var.cc:
table list removed from fix_fields() arguments
sql/share/errmsg.txt:
new error message
sql/sp.cc:
Name resolution context added
sql/sp_head.cc:
table list removed from fix_fields() arguments
sql/sp_head.h:
Name resolution context added
sql/sql_base.cc:
table list removed from fix_fields() arguments
Name resolution context added
sql/sql_class.cc:
renamed variable
sql/sql_delete.cc:
Name resolution context added
sql/sql_derived.cc:
Name resolution context added
sql/sql_do.cc:
table list removed from fix_fields() arguments
sql/sql_handler.cc:
Name resolution context added
sql/sql_help.cc:
Name resolution context added
sql/sql_insert.cc:
Name resolution context added
table list removed from fix_fields() arguments
sql/sql_lex.cc:
Name resolution context added
sql/sql_lex.h:
removed resolve mode (information stored into name resolution context)
sql/sql_load.cc:
table list removed from fix_fields() arguments
sql/sql_olap.cc:
Name resolution context added
sql/sql_parse.cc:
Name resolution context added
sql/sql_prepare.cc:
table list removed from fix_fields() arguments
sql/sql_select.cc:
table list removed from fix_fields() arguments
sql/sql_show.cc:
Name resolution context added
sql/sql_trigger.cc:
table list removed from fix_fields() arguments
sql/sql_udf.h:
table list removed from fix_fields() arguments
sql/sql_union.cc:
Name resolution context added
sql/sql_update.cc:
Name resolution context added
sql/sql_view.cc:
Name resolution context added
sql/sql_view.h:
table list removed from fix_fields() arguments
sql/sql_yacc.yy:
Name resolution context added
sql/table.cc:
Name resolution context added
merged view processing moved
sql/table.h:
merged view processing moved
Make sure to cleanup the items for a cursor query after each open, otherwise
it's done too late, after the run-time mem_root is freed.
mysql-test/r/sp.result:
New test case for BUG#11529.
mysql-test/t/sp.test:
New test case for BUG#11529.
sql/sp_head.cc:
Add a back pointer from a sp_cursor to its cpush instruction, and use it to set
the arena and cleanup the items for the cursor's query when opening it.
sql/sp_rcontext.cc:
Store pointer in sp_cursor to its cpush instruction.
sql/sp_rcontext.h:
Store pointer in sp_cursor to its cpush instruction.
We need every instruction to have its own arena, because we want to
track instruction's state (INITIALIZED_FOR_SP -> EXECUTED). Because of
`if' statements and other conditional instructions used in stored
procedures, not every instruction of a stored procedure gets executed
during the first (or even subsequent) execution of the procedure.
So it's better if we track the execution state of every instruction
independently.
All instructions of a given procedure now also share sp_head's
mem_root, but keep their own free_list.
This simplifies juggling with free Item lists in sp_head::execute.
- free_items() moved to be a member of Query_arena.
- logic of 'backup_arena' debug member of Query_arena has been
changed to support
multi-backups. Until now, TRUE 'backup_arena' meant that there is
exactly one active backup of the THD arena. Now it means simply that
the arena is used for backup, so that we can't accidentally overwrite an
existing backup. This allows doing multiple backups, e.g. in
sp_head::execute and Cursor::fetch, when THD arena is already backed up
but we want to set yet another arena (usually the 'permanent' arena,
to save permanent transformations/optimizations of a parsed tree).
sql/sp_head.cc:
- use Query_arena support in sp_head::execute() as now sp_instr inherites
from it.
sql/sp_head.h:
- inherite sp_instr from Query_arena
sql/sql_class.cc:
- changed the principle of Query_arena::backup_arena; free_items is now
a member of Query_arena.
sql/sql_class.h:
- changed the principle of Query_arena::backup_arena; free_items is now
a member of Query_arena.
sql/sql_prepare.cc:
free_items() is now a member of Query_arena.
sql/sql_select.cc:
free_items() now automatically sets free_list to zero.
We need every instruction to have its own arena, because we want to
track instruction's state (INITIALIZED_FOR_SP -> EXECUTED). Because of
`if' statements and other conditional instructions used in stored
procedures, not every instruction of a stored procedure gets executed
during the first (or even subsequent) execution of the procedure.
So it's better if we track the execution state of every instruction
independently.
All instructions of a given procedure now also share sp_head's
mem_root, but keep their own free_list.
This simplifies juggling with free Item lists in sp_head::execute.
- free_items() moved to be a member of Query_arena.
- logic of 'backup_arena' debug member of Query_arena has been
changed to support
multi-backups. Until now, TRUE 'backup_arena' meant that there is
exactly one active backup of the THD arena. Now it means simply that
the arena is used for backup, so that we can't accidentally overwrite an
existing backup. This allows doing multiple backups, e.g. in
sp_head::execute and Cursor::fetch, when THD arena is already backed up
but we want to set yet another arena (usually the 'permanent' arena,
to save permanent transformations/optimizations of a parsed tree).
main_mem_root is moved out of class Query_arena.
sql/sp_head.cc:
Adjust to the changed Query_arena constructor.
main_mem_root is moved out of class Query_arena.
sql/sp_head.h:
main_mem_root is moved out of class Query_arena: add it to class sp_head.
sql/sql_class.cc:
main_mem_root is moved out of class Query_arena: remove
constructors no longer relevant, remove dead code.
sql/sql_class.h:
main_mem_root is moved out of class Query_arena.
sql/sql_prepare.cc:
It's better to not use main_mem_root anywhere: logically, it's not
a public member (can't fix sp_head::make_field and Item_subselect::exec
to make it protected)
sql/sql_select.cc:
New Cursor constructor, which avoids unneeded memory allocation
when initializign main_mem_root.
sql/sql_select.h:
main_mem_root is moved out of class Query_arena.
mysql-test/r/sp.result:
test commented until bug#11394 fix
test for bug#10136
mysql-test/t/sp.test:
test commented until bug#11394 fix
bug10136
sql/sp_head.cc:
fixed items cleunup for SP
query transformations to the PREPARE stage (prepared statements).
sql/item.h:
Remove an unused friend declaration.
sql/mysql_priv.h:
Change signature of insert_fields()
sql/sp_head.cc:
Make is_stmt_prepare_or_first_sp_execute really work: reset SP state
to EXECUTED after execution.
sql/sql_base.cc:
allocate_view_names flag of insert_fields is removed.
The purpose of this variable was to amend the case when a view
is replaced with a base table between subsequent executions of a prepared
statement: in that case the new table theoretically can be used instead
of the view. If allocate_view_names was set,
all the references to the view expressions were replaced with Item_field's
which in turn could have been resolved by their names.
But this approach doesn't work for other reasons, so let's not try
to help what must be solved by TDC.
sql/sql_class.h:
Add is_first_sp_execute() helper method.
sql/sql_handler.cc:
insert_fields signature changed.
sql/sql_lex.h:
Add a comment for variable 'first_execution'.
into mysql.com:/home/pem/work/mysql-5.0
mysql-test/r/sp.result:
Auto merged
mysql-test/t/sp.test:
Auto merged
sql/item_func.cc:
Auto merged
sql/sp_head.cc:
Auto merged
sql/sp_rcontext.cc:
Auto merged
Fixed valgrind complaints. This fixes the memory leak problems for
procedured, and partially for functions. There's still a leak involving
results from functions that turned out to be too involved, so it will be
fixed separately.
mysql-test/r/sp.result:
Fixed some minor mistake (spotted while debugging).
mysql-test/t/sp.test:
Fixed some minor mistake (spotted while debugging).
sql/item_func.cc:
Moved Item_func_sp::cleanup() from item_func.h to ease debugging,
and made a debug output come out right.
sql/item_func.h:
Moved Item_func_sp::cleanup() to item_func.cc to ease debugging.
sql/sp_head.cc:
Fixed valgrind problems with the previous memory leak fix (unit cleanup and
putting result field in a differen mem_root), and removed prealloc flag from
init_alloc_root() calls.
sql/sp_rcontext.cc:
New mem_root pointer used for return fields from functions.
sql/sp_rcontext.h:
New mem_root pointer used for return fields from functions.
Fixed valgrind complaints. This fixes the memory leak problems for
procedured, and partially for functions. There's still a leak involving
results from functions that turned out to be too involved, so it will be
fixed separately.
CURSOR_TYPE_READ_ONLY". The bug was that we (me) don't perform proper
cleanups of the prepared statement when done fetching from a cursor.
Another patch.
sql/mysql_priv.h:
Rename reset_stmt_for_execute to init_stmt_before_use (to correspond to
cleanup_stmt_and_thd_after_use).
sql/sp_head.cc:
Rename.
sql/sql_prepare.cc:
Move common cleanup code to a cleanup function, call it when we close
a cursor.
sql/sql_select.cc:
Cleanup.
sql/sql_select.h:
No need for init_thd, this code has been inlined in Cursor::open.
tests/mysql_client_test.c:
Add a test case for Bug#10729 "mysql_stmt_attr_set CURSOR_TYPE_READ_ONLY"
(problem reusing a prepared statemnt if there was a cursor)
CURSOR_TYPE_READ_ONLY". The bug was that we (me) don't perform proper
cleanups of the prepared statement when done fetching from a cursor.
Another patch.
sql/item.h:
Auto merged
sql/item_subselect.cc:
Auto merged
sql/sp_head.cc:
Auto merged
sql/sql_class.h:
Auto merged
sql/sql_handler.cc:
Auto merged
sql/sql_parse.cc:
Auto merged
sql/sql_repl.cc:
Auto merged
sql/sql_select.cc:
Auto merged
sql/sql_yacc.yy:
Auto merged
error for LIMIT placeholder".
The patch adds grammar support for LIMIT ?, ? and changes the
type of ST_SELECT_LEX::select_limit,offset_limit from ha_rows to Item*,
so that it can point to Item_param.
mysql-test/include/ps_modify.inc:
Fix existing tests: now LIMIT can contain placeholders.
mysql-test/include/ps_query.inc:
Fix existing tests: now LIMIT can contain placeholders.
mysql-test/r/ps.result:
Add basic test coverage for LIMIT ?, ? and fix test results.
mysql-test/r/ps_2myisam.result:
Fix test results: now LIMIT can contain placeholders.
mysql-test/r/ps_3innodb.result:
Fix test results: now LIMIT can contain placeholders.
mysql-test/r/ps_4heap.result:
Fix test results: now LIMIT can contain placeholders.
mysql-test/r/ps_5merge.result:
Fix test results: now LIMIT can contain placeholders.
mysql-test/r/ps_6bdb.result:
Fix test results: now LIMIT can contain placeholders.
mysql-test/r/ps_7ndb.result:
Fix test results: now LIMIT can contain placeholders.
mysql-test/t/ps.test:
Add basic test coverage for LIMIT ?, ?.
sql/item.h:
Add a short-cut for (ulonglong) val_int() to Item.
Add a constructor to Item_int() that accepts ulonglong.
Simplify Item_uint constructor by using the c-tor above.
sql/item_subselect.cc:
Now select_limit has type Item *.
We can safely create an Item in Item_exists_subselect::fix_length_and_dec():
it will be allocated in runtime memory root and freed in the end of
execution.
sql/sp_head.cc:
Add a special initalization state for stored procedures to
be able to easily distinguish the first execution of a stored procedure
from prepared statement prepare.
sql/sql_class.h:
Introduce new state 'INITIALIZED_FOR_SP' to be able to easily distinguish
the first execution of a stored procedure from prepared statement prepare.
sql/sql_derived.cc:
- use unit->set_limit() to set unit->select_limit_cnt, offset_limit_cnt
evreryplace. Add a warning about use of set_limit in
mysql_derived_filling.
sql/sql_error.cc:
- use unit->set_limit() to set unit->select_limit_cnt, offset_limit_cnt
evreryplace.
- this change is also aware of bug#11095 "show warnings limit 0 returns
all rows instead of zero rows", so the one who merges the bugfix from
4.1 can use local version of sql_error.cc.
sql/sql_handler.cc:
- use unit->set_limit() to initalize
unit->select_limit_cnt,offset_limit_cnt everyplace.
sql/sql_lex.cc:
Now ST_SELECT_LEX::select_limit, offset_limit have type Item *
sql/sql_lex.h:
Now ST_SELECT_LEX::select_limit, offset_limit have type Item *
sql/sql_parse.cc:
- use unit->set_limit() to initalize
unit->select_limit_cnt,offset_limit_cnt everyplace.
- we can create an Item_int to set global limit of a statement:
it will be created in the runtime mem root and freed in the end of
execution.
sql/sql_repl.cc:
Use unit->set_limit to initialize limits.
sql/sql_select.cc:
- select_limit is now Item* so the proper way to check for default value
is to compare it with NULL.
sql/sql_union.cc:
Evaluate offset_limit_cnt using the new type of ST_SELECT_LEX::offset_limit
sql/sql_view.cc:
Now ST_SELECT_LEX::select_limit, offset_limit have type Item *
sql/sql_yacc.yy:
Add grammar support for LIMIT ?, ? clause.
error for LIMIT placeholder".
The patch adds grammar support for LIMIT ?, ? and changes the
type of ST_SELECT_LEX::select_limit,offset_limit from ha_rows to Item*,
so that it can point to Item_param.
Include "mysql_priv.h" before testing USE_PRAGMA_IMPLEMENTATION
sql/ha_federated.cc:
Include "mysql_priv.h" before testing USE_PRAGMA_IMPLEMENTATION
sql/sp_cache.cc:
Include "mysql_priv.h" before testing USE_PRAGMA_IMPLEMENTATION
sql/sp_head.cc:
Include "mysql_priv.h" before testing USE_PRAGMA_IMPLEMENTATION
sql/sp_pcontext.cc:
Include "mysql_priv.h" before testing USE_PRAGMA_IMPLEMENTATION
sql/sp_rcontext.cc:
Include "mysql_priv.h" before testing USE_PRAGMA_IMPLEMENTATION
client/mysql.cc:
Auto merged
libmysql/libmysql.c:
Auto merged
sql/ha_ndbcluster.cc:
Auto merged
sql/item.cc:
Auto merged
sql/item.h:
Auto merged
sql/item_func.cc:
Auto merged
sql/item_subselect.cc:
Auto merged
sql/item_sum.cc:
Auto merged
sql/mysql_priv.h:
Auto merged
sql/opt_range.cc:
Auto merged
sql/sql_base.cc:
Auto merged
sql/sql_bitmap.h:
Auto merged
sql/sql_class.h:
Auto merged
sql/sql_insert.cc:
Auto merged
sql/sql_parse.cc:
Auto merged
sql/sql_select.cc:
Auto merged
Simple optimization for 2 argument usage to function of variable arguments
Fix stack overrun when using 1+1+1+1+1+1+1+....
Update crash-me results for 5.0
Don't call post_open if pre_open() fails (optimization)
sql-bench/limits/mysql-4.1.cfg:
Rename: sql-bench/limits/mysql.cfg -> sql-bench/limits/mysql-4.1.cfg
libmysql/libmysql.c:
More portable define
mysql-test/mysql-test-run.sh:
Write also InnoDB warnings to warnings.log
mysql-test/t/type_newdecimal.test:
Don't get errors if innodb is not defined
mysys/my_alloc.c:
Cleanup comments
mysys/thr_lock.c:
Cleanup comments
sql/item.h:
Remove not needed initializer
sql/item_func.cc:
Simple optimization for 2 argument usage to function of variable arguments
sql/mysql_priv.h:
We use more stackspace with the introduction of int_op() etc.
This change ensures we don't run out of stack when doing 1+1+1+1...
(Tested on x86, 32 bit)
sql/sp_head.cc:
Don't call post_open if pre_open() fails
sql/sp_rcontext.cc:
More comments
Change so that post_open() doesn't have to be called if pre_open() fails
sql/sql_parse.cc:
Fold long lines
sql/sql_select.cc:
Simple reorganization to reduce number of if's
Ensure that table_map is updated for where clause (fixed warning from valgrind)
Simple optimization for 2 argument usage to function of variable arguments
Fix stack overrun when using 1+1+1+1+1+1+1+....
Update crash-me results for 5.0
Don't call post_open if pre_open() fails (optimization)
Free memory after all SP invokation.
sql/sp_head.cc:
Fixed memory leaks in SP invokation. Have to release memory after each call,
or items and contexts are not released until it's too late.
Have to catch errors from SELECT when opening a cursor.
mysql-test/r/sp.result:
New test case for BUG#10961.
mysql-test/t/sp.test:
New test case for BUG#10961.
sql/protocol.h:
Init data in Protocol_cursor constructor, for error cases.
sql/sp_head.cc:
Catch "hidden" errors during SELECT when opening a cursor.
sql/item_cmpfunc.cc:
Put the buff parameter back in check_stack_overrun().
sql/item_func.cc:
Put the buff parameter back in check_stack_overrun().
sql/item_subselect.cc:
Put the buff parameter back in check_stack_overrun().
sql/mysql_priv.h:
Put the buff parameter back in check_stack_overrun().
sql/mysqld.cc:
Style fixes.
sql/sp_head.cc:
Put the buff parameter back in check_stack_overrun().
sql/sql_parse.cc:
Put the buff parameter back in check_stack_overrun(), and added comment explaining the purpose.
sql/table.cc:
Put the buff parameter back in check_stack_overrun().
