Events: crash with procedure which alters events with function
Post-review CS
This fix also changes the handling of KILL command combined with
subquery. It changes the error message given back to "not supported",
from parse error. The error for CREATE|ALTER EVENT has also been changed
to generate "not supported yet" instead of parse error.
In case of a SP call, the error is "not supported yet". This change
cleans the parser from code which should not belong to there. Still
LEX::expr_allows_subselect is existant because it simplifies the handling
of SQLCOM_HA_READ which forbids subselects.
Use lazy initialization for Query_tables_list::sroutines hash.
This step should significantly decrease amount of memory consumed
by stored routines as we no longer will allocate chunk of memory
required for this HASH for each statement in routine.
Evaluate "NULL IN (SELECT ...)" in a special way: Disable pushed-down
conditions and their "consequences":
= Do full table scans instead of unique_[index_subquery] lookups.
= Change appropriate "ref_or_null" accesses to full table scans in
subquery's joins.
Also cache value of NULL IN (SELECT ...) if the SELECT is not correlated
wrt any upper select.
select OK.
The SQL parser was using Item::name to transfer user defined function attributes
to the user defined function (udf). It was not distinguishing between user defined
function call arguments and stored procedure call arguments. Setting Item::name
was causing Item_ref::print() method to print the argument as quoted identifiers
and caused views that reference aggregate functions as udf call arguments (and
rely on Item::print() for the text of the view to store) to throw an undefined
identifier error.
Overloaded Item_ref::print to print aggregate functions as such when printing
the references to aggregate functions taken out of context by split_sum_func2()
Fixed the parser to properly detect using AS clause in stored procedure arguments
as an error.
Fixed printing the arguments of udf call to print properly the udf attribute.
account predicates that become sargable after reading const tables.
In some cases this resulted in choosing non-optimal execution plans.
Now info of such potentially saragable predicates is saved in
an array and after reading const tables we check whether this
predicates has become saragable.
should fail to create
The problem was that this type of errors was checked during view
creation, which doesn't happen when CREATE VIEW is a statement of
a created stored routine.
The solution is to perform the checks at parse time. The idea of the
fix is that the parser checks if a construction just parsed is allowed
in current circumstances by testing certain flags, and this flags are
reset for VIEWs.
The side effect of this change is that if the user already have
such bogus routines, it will now get a error when trying to do
SHOW CREATE PROCEDURE proc;
(and some other) and when trying to execute such routine he will get
ERROR 1457 (HY000): Failed to load routine test.p5. The table mysql.proc is missing, corrupt, or contains bad data (internal code -6)
However there should be very few such users (if any), and they may
(and should) drop these bogus routines.
Presence of a subquery in the ON expression of a join
should not block merging the view that contains this join.
Before this patch the such views were converted into
into temporary table views.
containing a select statement that uses an aggregating IN subquery.
Added a parameter to the function fix_prepare_information
to restore correctly the having clause for the second execution.
Saved andor structure of the having conditions at the proper moment
before any calls of split_sum_func2 that could modify the having structure
adding new Item_ref objects. (These additions, are produced not with
the statement mem_root, but rather with the execution mem_root.)
make st_select_lex::setup_ref_array() take into account that
Item_sum-descendant objects located within descendant SELECTs
may be added into ref_pointer_array.
handle them.
Problem:
CREATE|ALTER EVENT, HANDLER READ, KILL, Partitioning uses `expr` from the
parser. This rule comes with all the rings and bells including subqueries.
However, these commands are not subquery safe. For this reason there are two
fuse checks in the parser. They were checking by command id. CREATE EVENT
should forbid subquery is the fix for
bug#16394 Events: Crash if schedule contains SELECT
The fix has been incorporated as part of the patch for WL#3337 (Event scheduler
new architecture).
Solution:
A new flag was added to LEX command_forbids_subselect. The fuse checks were
changed. The commands are responsible to set the value to true whenever
they can't handle subselects.
Zero-length variables caused failures when using the length to look
up the name in a hash. Instead, signal that no zero-length name can
ever be found and that to encounter one is a syntax error.
- if there are two character set definitions in the column declaration,
we replace the first one with the second one as we store both in the LEX->charset
slot. Add a separate slot to the LEX structure to store underscore charset.
- convert default values to the column charset of STRING, VARSTRING fields
if necessary as well.
can be not replicable.
Now CREATE statements for writing in the binlog are created as follows:
- the beginning of the statement is re-created;
- the rest of the statement is copied from the original query.
The problem appears when there is a version-specific comment (produced by
mysqldump), started in the re-created part of the statement and closed in the
copied part -- there is closing comment-parenthesis, but there is no opening
one.
The proper fix could be to re-create original statement, but we can not
implement it in 5.0. So, for 5.0 the fix is just to cut closing
comment-parenthesis. This technique is also used for SHOW CREATE PROCEDURE
statement (so we are able to reuse existing code).
1) Fix for BUG#19630 "stored function inserting into two auto_increment breaks
statement-based binlog":
a stored function inserting into two such tables may fail to replicate
(inserting wrong data in the slave's copy of the second table) if the slave's
second table had an internal auto_increment counter different from master's.
Because the auto_increment value autogenerated by master for the 2nd table
does not go into binlog, only the first does, so the slave lacks information.
To fix this, if running in mixed binlogging mode, if the stored function or
trigger plans to update two different tables both having auto_increment
columns, we switch to row-based for the whole function.
We don't have a simple solution for statement-based binlogging mode, there
the bug remains and will be documented as a known problem.
Re-enabling rpl_switch_stm_row_mixed.
2) Fix for BUG#20630 "Mixed binlogging mode does not work with stored
functions, triggers, views", which was a documented limitation (in mixed
mode, we didn't detect that a stored function's execution needed row-based
binlogging (due to some UUID() call for example); same for
triggers, same for views (a view created from a SELECT UUID(), and doing
INSERT INTO sometable SELECT theview; would not replicate row-based).
This is implemented by, after parsing a routine's body, remembering in sp_head
that this routine needs row-based binlogging. Then when this routine is used,
the caller is marked to require row-based binlogging too.
Same for views: when we parse a view and detect that its SELECT needs
row-based binary logging, we mark the calling LEX as such.
3) Fix for BUG#20499 "mixed mode with temporary table breaks binlog":
a temporary table containing e.g. UUID has its changes not binlogged,
so any query updating a permanent table with data from the temporary table
will run wrongly on slave. Solution: in mixed mode we don't switch back
from row-based to statement-based when there exists temporary tables.
4) Attempt to test mysqlbinlog on a binlog generated by mysqlbinlog;
impossible due to BUG#11312 and BUG#20329, but test is in place for when
they are fixed.
dropping/creating tables".
The bug could lead to a crash when multi-delete statements were
prepared and used with temporary tables.
The bug was caused by lack of clean-up of multi-delete tables before
re-execution of a prepared statement. In a statement like
DELETE t1 FROM t1, t2 WHERE ... the first table list (t1) is
moved to lex->auxilliary_table_list and excluded from lex->query_tables
or select_lex->tables. Thus it was unaccessible to reinit_stmt_before_use
and not cleaned up before re-execution of a prepared statement.
This cut No 7 should finish the part of fixing the parsing of the events :
- Event_timed is no more used during parsing. Less problems because it has
a mutex. Event_parse_data class is used during parsing. It is suited only
for this purpose. It's pretty lightweight
- Late checking of data from parsing is being performed. This should solve
the problems of nested events in SP or other events (for the situation
of no nested bodies). Before if an ALTER EVENT was in a SP, then when the
SP was compiled, and not executed, the actual init_xxx methods of Event_timed
were called, which is wrong.
- It could be a side effect of using a specialized class, but test events_stress is
now 25% quicker.
Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved
to Event_queue.
The problem was that we restored SQL_CACHE, SQL_NO_CACHE flags in SELECT
statement from internal structures based on value set later at runtime, not
the original value set by the user.
The solution is to remember that original value.
SHOW STATUS are not anymore put in slow query log because of no index usage.
Implemntation done by removing orig_sql_command and moving logic of SHOW STATUS to mysql_excute_command()
This simplifies code and allows us to remove some if statements all over the code.
Upgraded uc_update_queries[] to sql_command_flags and added more bitmaps to better categorize commands.
This allowed some overall simplifaction when testing sql_command.
Fixes bugs:
Bug#10210: running SHOW STATUS increments counters it shouldn't
Bug#19764: SHOW commands end up in the slow log as table scans
Old FRM files didn't allow byte 255 as a column name part.
Since now we store column names in UTF8, this restriction
is not required anymore: 255 is not a valid byte in UTF8.
Fix: removing checking against 255.
Bug#18282 "INFORMATION_SCHEMA.TABLES provides inconsistent info about invalid views"
This bug caused crashes or resulted in wrong data being returned
when one tried to obtain information from I_S tables about views
using stored functions.
It was caused by the fact that we were using LEX representing
statement which were doing select from I_S tables as active LEX
when contents of I_S table were built. So state of this LEX both
affected and was affected by open_tables() calls which happened
during this process. This resulted in wrong behavior and in
violations of some of invariants which caused crashes.
This fix tries to solve this problem by properly saving/resetting
and restoring part of LEX which affects and is affected by the
process of opening tables and views in get_all_tables() routine.
To simplify things we separated this part of LEX in a new class
and made LEX its descendant.
A query with a group by and having clauses could return a wrong
result set if the having condition contained a constant conjunct
evaluated to FALSE.
It happened because the pushdown condition for table with
grouping columns lost its constant conjuncts.
Pushdown conditions are always built by the function make_cond_for_table
that ignores constant conjuncts. This is apparently not correct when
constant false conjuncts are present.
Move plugin declarations after system functions have been checked
(Fixes problem with ndb_config failing becasue SHM is not declared)
Fixed some memory leaks
- detect the need for row-based binlogging not at execution stage but earlier at parsing stage; needed for example for CREATE TABLE SELECT UUID().
- more tests of this mixed mode.
A query with a group by and having clauses could return a wrong
result set if the having condition contained a constant conjunct
evaluated to FALSE.
It happened because the pushdown condition for table with
grouping columns lost its constant conjuncts.
Pushdown conditions are always built by the function make_cond_for_table
that ignores constant conjuncts. This is apparently not correct when
constant false conjuncts are present.
The table opening process now works the following way:
- Create common TABLE_SHARE object
- Read the .frm file and unpack it into the TABLE_SHARE object
- Create a TABLE object based on the information in the TABLE_SHARE
object and open a handler to the table object
Other noteworthy changes:
- In TABLE_SHARE the most common strings are now LEX_STRING's
- Better error message when table is not found
- Variable table_cache is now renamed 'table_open_cache'
- New variable 'table_definition_cache' that is the number of table defintions that will be cached
- strxnmov() calls are now fixed to avoid overflows
- strxnmov() will now always add one end \0 to result
- engine objects are now created with a TABLE_SHARE object instead of a TABLE object.
- After creating a field object one must call field->init(table) before using it
- For a busy system this change will give you:
- Less memory usage for table object
- Faster opening of tables (if it's has been in use or is in table definition cache)
- Allow you to cache many table definitions objects
- Faster drop of table
Bad examples of usage of a string with its length fixed.
The incorrect length in the trigger file configuration descriptor
fixed (BUG#14090).
A hook for unknown keys added to the parser to support old .TRG files.
ESCAPE has length of 1 if specified and sql_mode is NO_BACKSLASH_ESCAPES
or has length of 0 or 1 in every other situation.
(approved patch applied on a up-to-date tree re-commit)
When fixing Item_func_plus in ORDER BY clause field c is searched in all
opened tables, but because c is an alias it wasn't found there.
This patch adds a flag to select_lex which allows Item_field::fix_fields()
to look up in select's item_list to find aliased fields.
thd->allow_sum_func was left 'true' after previous statement thus allowing
sum funcs to be present in conditions.
thd->allow_sum_func should be set to 0 for each query and each prepared
statement reinitialization. This is done in lex_start() and
reset_stmt_for_execute().
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.
The idea of the patch is to separate statement processing logic,
such as parsing, validation of the parsed tree, execution and cleanup,
from global query processing logic, such as logging, resetting
priorities of a thread, resetting stored procedure cache, resetting
thread count of errors and warnings.
This makes PREPARE and EXECUTE behave similarly to the rest of SQL
statements and allows their use in stored procedures.
This patch contains a change in behaviour:
until recently for each SQL prepared statement command, 2 queries
were written to the general log, e.g.
[Query] prepare stmt from @stmt_text;
[Prepare] select * from t1 <-- contents of @stmt_text
The chagne was necessary to prevent [Prepare] commands from being written
to the general log when executing a stored procedure with Dynamic SQL.
We should consider whether the old behavior is preferrable and probably
restore it.
This patch refixes Bug#7115, Bug#10975 (partially), Bug#10605 (various bugs
in Dynamic SQL reported before it was disabled).
- 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.
"Interleaved SPs execution is now binlogged properly, "SELECT spfunc()" is binlogged too.
The known remaining issue is binlogging/replication of "a routine is deleted while it is executed" scenario.
"Process NATURAL and USING joins according to SQL:2003".
* Some of the main problems fixed by the patch:
- in "select *" queries the * expanded correctly according to
ANSI for arbitrary natural/using joins
- natural/using joins are correctly transformed into JOIN ... ON
for any number/nesting of the joins.
- column references are correctly resolved against natural joins
of any nesting and combined with arbitrary other joins.
* This patch also contains a fix for name resolution of items
inside the ON condition of JOIN ... ON - in this case items must
be resolved only against the JOIN operands. To support such
'local' name resolution, the patch introduces a stack of
name resolution contexts used at parse time.
NOTICE:
- This patch is not complete in the sense that
- there are 2 test cases that still do not pass -
one in join.test, one in select.test. Both are marked
with a comment "TODO: WL#2486".
- it does not include a new test specific for the task
its body, but lets each statement to get/release its own locks. This allows a broader set
of statements to be executed inside PROCEDUREs (but breaks replication)
This patch should fix BUG#8072, BUG#8766, BUG#9563, BUG#11126
of stored routines definitions even if we already have some tables open and
locked. To avoid deadlocks in this case we have to put certain restrictions
on locking of mysql.proc table.
This allows to use stored routines safely under LOCK TABLES without explicitly
mentioning mysql.proc in the list of locked tables. It also fixes bug #11554
"Server crashes on statement indirectly using non-cached function".
crash if referencing a table" and several other related bugs.
Fix for bug #11834 "Re-execution of prepared statement with dropped function
crashes server." which was spotted during work on previous bugs.
Also couple of nice cleanups:
- Replaced two separate hashes for stored routines used by statement with one.
- Now instead of doing one pass through all routines used in statement for
caching them and then doing another pass for adding their tables to table
list, we do only one pass during which do both things.
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.
"the server side preparedStatement error for LIMIT placeholder",
which moves all uses of LIMIT clause from PREPARE to OPTIMIZE
and later steps.
After-review fixes.
and some SP-related cleanups.
- We don't have separate stage for calculation of list of tables
to be prelocked and doing implicit LOCK/UNLOCK any more.
Instead we calculate this list at open_tables() and do implicit
LOCK in lock_tables() (and UNLOCK in close_thread_tables()).
Also now we support cases when same table (with same alias) is
used several times in the same query in SP.
- Cleaned up execution of SP. Moved all common code which handles
LEX and does preparations before statement execution or complex
expression evaluation to auxilary sp_lex_keeper class. Now
all statements in SP (and corresponding instructions) that
evaluate expression which can contain subquery have their
own LEX.
Collect all tables and SPs refered by a statement, and open all tables
with an implicit LOCK TABLES. Do find things refered by triggers and views,
we open them first (and then repeat this until nothing new is found), before
doing the actual lock tables.
does not work well together". Now using simplier and more correct
implementation of st_lex::unlink_first_table()/link_first_table_back()
(It also nicely handles case when global table list is created because
of implictly used time zone tables). (2nd attempt)
Fix for bug #7705 "CONVERT_TZ() crashes with subquery/WHERE on index
column". Implemented new approach for caching objects for constant
time zone arguments. Now instead of determining whenever these arguments
are constants and performing time zone lookup at fix_fields() stage, we
do it on first get_date() invocation.
Cleanup of global @@time_zone variable handling.
IN and CONVERT_TZ()" (with after review changes).
Now we add implicitly used time zone tables to global table list right
at the parsing stage instead of doing it later in mysql_execute_command()
or in check_prepared_statement().
No special test-case needed since this bug also manifests itself as
timezone2.test failure if one runs it with --ps-protocol option.
out of order". (final version)
Now instead of binding Item_trigger_field to TABLE objects during
trigger definition parsing at table open, we perform pass through
special list of all such objects in trigger. This allows easily check
all references to fields in old/new version of row in trigger during
execution of CREATE TRIGGER statement (this is more courtesy for users
since we can't check everything anyway).
We also report that such reference is bad by returning error from
Item_trigger_field::fix_fields() method (instead of setup_field())
This means that if trigger is broken we will bark during trigger
execution instead of trigger definition parsing at table open.
(i.e. now we allow to open tables with broken triggers).
crashes server." The fix makes Item_func_rand prepared-statements
aware plus it fixes the case when RAND is used in prepared
statements and replication is on (as well as several similar issues).
Until now we did not reset THD before every execution of a prepared
statement, so if some execution had set thd->time_zone_used
or thd->rand_used they would not be reset until next mysql_parse.
Some of post-review fixes done.
Added a check to recover from IGNORE_SPACE in this situation:
<ident-character(s)><space><dot><ident-character(s)>
The ignored space led to the false identification of the dot
as an ident separator (like "db.table").
Fixed (together with Guilhem) bugs in mysqlbinlog regarding --offset
Prefix addresses with 0x for easier comparisons of debug logs
Fixed problem where MySQL choosed index-read even if there would be a much better range on the same index
This fix changed some 'index' queries to 'range' queries in the test suite
Don't create 'dummy' WHERE clause for trivial WHERE clauses where we can remove the WHERE clause.
This fix removed of a lot of 'Using where' notes in the test suite.
Give NOTE instead of WARNING if table/function doesn't exists when using DROP IF EXISTS
Give NOTE instead of WARNING for safe field-type conversions
Mostly needed for Monty for him getting notion what needed for triggers
from new .FRM format.
Things to be done:
- Right placement of trigger's invocations
- Right handling of errors in triggers (including transaction rollback)
- Support for priviliges
- Right handling of DROP/RENAME table (hope that it will be handled automatically
with merging of .TRG into .FRM file)
- Saving/restoring some information critical for trigger creation and replication
with their definitions (e.g. sql_mode, creator, ...)
- Replication
Already has some known bugs so probably not for general review.
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.
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.
