"real" table fails in JOINs".
This is a regression caused by the fix for Bug 18444.
This fix removed the assignment of empty_c_string to table->db performed
in add_table_to_list, as neither me nor anyone else knew what it was
there for. Now we know it and it's covered with tests: the only case
when a table database name can be empty is when the table is a derived
table. The fix puts the assignment back but makes it a bit more explicit.
Additionally, finally drop sp.result.orig which was checked in by mistake.
context.
Routine arguments were evaluated in the security context of the routine
itself, not in the caller's context.
The bug is fixed the following way:
- Item_func_sp::find_and_check_access() has been split into two
functions: Item_func_sp::find_and_check_access() itself only
finds the function and check that the caller have EXECUTE privilege
on it. New function set_routine_security_ctx() changes security
context for SUID routines and checks that definer have EXECUTE
privilege too.
- new function sp_head::execute_trigger() is called from
Table_triggers_list::process_triggers() instead of
sp_head::execute_function(), and is effectively just as the
sp_head::execute_function() is, with all non-trigger related code
removed, and added trigger-specific security context switch.
- call to Item_func_sp::find_and_check_access() stays outside
of sp_head::execute_function(), and there is a code in
sql_parse.cc before the call to sp_head::execute_procedure() that
checks that the caller have EXECUTE privilege, but both
sp_head::execute_function() and sp_head::execute_procedure() call
set_routine_security_ctx() after evaluating their parameters,
and restore the context after the body is executed.
run at startup"
The server returned an error when trying to execute init-file with a
stored procedure that could return multiple result sets to the client.
A stored procedure can return multiple result sets if it contains
PREPARE, SELECT, SHOW and similar statements.
The fix is to set client_capabilites|=CLIENT_MULTI_RESULTS in
sql_parse.cc:handle_bootstrap(). There is no "client" really, so
nothing is ever sent. This makes init-file feature behave consistently:
the prepared statements that can be called directly in the init-file
can be used in a stored procedure too.
Re-committed the patch originally submitted by Per-Erik after review.
NDB table".
SQL-layer was not marking fields which were used in triggers as such. As
result these fields were not always properly retrieved/stored by handler
layer. So one might got wrong values or lost changes in triggers for NDB,
Federated and possibly InnoDB tables.
This fix solves the problem by marking fields used in triggers
appropriately.
Also this patch contains the following cleanup of ha_ndbcluster code:
We no longer rely on reading LEX::sql_command value in handler in order
to determine if we can enable optimization which allows us to handle REPLACE
statement in more efficient way by doing replaces directly in write_row()
method without reporting error to SQL-layer.
Instead we rely on SQL-layer informing us whether this optimization
applicable by calling handler::extra() method with
HA_EXTRA_WRITE_CAN_REPLACE flag.
As result we no longer apply this optimzation in cases when it should not
be used (e.g. if we have on delete triggers on table) and use in some
additional cases when it is applicable (e.g. for LOAD DATA REPLACE).
Finally this patch includes fix for bug#20728 "REPLACE does not work
correctly for NDB table with PK and unique index".
This was yet another problem which was caused by improper field mark-up.
During row replacement fields which weren't explicity used in REPLACE
statement were not marked as fields to be saved (updated) so they have
retained values from old row version. The fix is to mark all table
fields as set for REPLACE statement. Note that in 5.1 we already solve
this problem by notifying handler that it should save values from all
fields only in case when real replacement happens.
Produce a warning if DATA/INDEX DIRECTORY is specified in
ALTER TABLE statement.
Ignoring of these options is documented in the symbolic links
section of the manual.
Bug#19022 "Memory bug when switching db during trigger execution"
Bug#17199 "Problem when view calls function from another database."
Bug#18444 "Fully qualified stored function names don't work correctly in
SELECT statements"
Documentation note: this patch introduces a change in behaviour of prepared
statements.
This patch adds a few new invariants with regard to how THD::db should
be used. These invariants should be preserved in future:
- one should never refer to THD::db by pointer and always make a deep copy
(strmake, strdup)
- one should never compare two databases by pointer, but use strncmp or
my_strncasecmp
- TABLE_LIST object table->db should be always initialized in the parser or
by creator of the object.
For prepared statements it means that if the current database is changed
after a statement is prepared, the database that was current at prepare
remains active. This also means that you can not prepare a statement that
implicitly refers to the current database if the latter is not set.
This is not documented, and therefore needs documentation. This is NOT a
change in behavior for almost all SQL statements except:
- ALTER TABLE t1 RENAME t2
- OPTIMIZE TABLE t1
- ANALYZE TABLE t1
- TRUNCATE TABLE t1 --
until this patch t1 or t2 could be evaluated at the first execution of
prepared statement.
CURRENT_DATABASE() still works OK and is evaluated at every execution
of prepared statement.
Note, that in stored routines this is not an issue as the default
database is the database of the stored procedure and "use" statement
is prohibited in stored routines.
This patch makes obsolete the use of check_db_used (it was never used in the
old code too) and all other places that check for table->db and assign it
from THD::db if it's NULL, except the parser.
How this patch was created: THD::{db,db_length} were replaced with a
LEX_STRING, THD::db. All the places that refer to THD::{db,db_length} were
manually checked and:
- if the place uses thd->db by pointer, it was fixed to make a deep copy
- if a place compared two db pointers, it was fixed to compare them by value
(via strcmp/my_strcasecmp, whatever was approproate)
Then this intermediate patch was used to write a smaller patch that does the
same thing but without a rename.
TODO in 5.1:
- remove check_db_used
- deploy THD::set_db in mysql_change_db
See also comments to individual files.
Addendum fixes after changing the condition variable
for the global read lock.
The stress test suite revealed some deadlocks. Some were
related to the new condition variable (COND_global_read_lock)
and some were general problems with the global read lock.
It is now necessary to signal COND_global_read_lock whenever
COND_refresh is signalled.
We need to wait for the release of a global read lock if one
is set before every operation that requires a write lock.
But we must not wait if we have locked tables by LOCK TABLES.
After setting a global read lock a thread waits until all
write locks are released.
schemas
The function check_one_table_access() called to check access to tables in
SELECT/INSERT/UPDATE was doing additional checks/modifications that don't hold
in the context of setup_tables_and_check_access().
That's why the check_one_table() was split into two : the functionality needed by
setup_tables_and_check_access() into check_single_table_access() and the rest of
the functionality stays in check_one_table_access() that is made to call the new
check_single_table_access() function.
function crashes server".
Attempts to execute prepared multi-delete statement which involved trigger or
stored function caused server crashes (the same happened for such statements
included in stored procedures in cases when one tried to execute them more
than once).
The problem was caused by yet another incorrect usage of check_table_access()
routine (the latter assumes that table list which it gets as argument
corresponds to value LEX::query_tables_own_last). We solve this problem by
juggling with LEX::query_tables_own_last value when we call
check_table_access() for LEX::auxilliary_table_list (better solution is too
intrusive and should be done in 5.1).
there was two problems about charsets in embedded server
1. mysys/charset.c - defined there default_charset_info variable is
modified by both server and client code (particularly when
--default-charset option is handled)
In embedded server we get two codelines modifying one variable.
I created separate default_client_charset_info for client code
2. mysql->charset and mysql->options.charset initialization isn't
properly done for embedded server - necessary calls added
There was an incomplete reset of the name resolution context, that caused
INSERT ... SELECT ... JOIN statements to resolve not by joint row type calculated
for the join.
Removed the redundant re-initialization of the context, because
mysql_insert_select_prepare() now correctly saves/restores the context.
There was a wrong determination of the DB name (witch is
not always the one in TABLE_LIST because derived tables
may be calculated using temp tables that have their db name
set to "").
The fix determines the database name according to the type
of table reference, and calls the function check_access()
with the correct db name so the correct set of grants is found.
There actually was 3 different problems -
hash_user_connections wasn't cleaned
one strdupped database name wasn't freed
and stmt->mem_root wasn't cleaned as it was
replased with mysql->field_alloc for result
For the last one - i made the library using stmt's
fields to store result if it's the case.
In multi-table delete a table for delete can't be used for selecting in
subselects. Appropriate error was raised but wasn't checked which leads to a
crash at the execution phase.
The mysql_execute_command() now checks for errors before executing select
for multi-delete.
The check for view security was lacking several points :
1. Check with the right set of permissions : for each table ref that
participates in a view there were the right credentials to use in it's
security_ctx member, but these weren't used for checking the credentials.
This makes hard enforcing the SQL SECURITY DEFINER|INVOKER property
consistently.
2. Because of the above the security checking for views was just ruled out
in explicit ways in several places.
3. The security was checked only for the columns of the tables that are
brought into the query from a view. So if there is no column reference
outside of the view definition it was not detecting the lack of access to
the tables in the view in SQL SECURITY INVOKER mode.
The fix below tries to fix the above 3 points.
which explicitly or implicitly uses stored function gives 'Table not locked'
error"
Test case for these bugs crashed in --ps-protocol mode. The crash was caused
by incorrect usage of check_grant() routine from create_table_precheck()
routine. The former assumes that either number of tables to be inspected by
it is limited explicitly (i.e. is is not UINT_MAX) or table list used and
thd->lex->query_tables_own_last value correspond to each other.
create_table_precheck() was not fulfilling this condition and crash happened.
The fix simply sets number of tables to be inspected by check_grant() to 1.
after merge.
Concurrent read and update of privilege structures (like simultaneous
run of SHOW GRANTS and ADD USER) could result in server crash.
Ensure that proper locking of ACL structures is done.
No test case is provided because this bug can't be reproduced
deterministically.
There were two distict bugs: parse error was returned for valid
statement and that error wasn't reported to the client.
The fix ensures that EXPLAIN SELECT..INTO is accepted by parser and any
other parse error will be reported to the client.
Bug#17667: An attacker has the opportunity to bypass query logging.
This adds a new, local-only printf format specifier to our *printf functions
that allows us to print known-size buffers that must not be interpreted as
NUL-terminated "strings."
It uses this format-specifier to print to the log, thus fixing this
problem.
The bug caused wrong result sets for union constructs of the form
(SELECT ... ORDER BY order_list1 [LIMIT n]) ORDER BY order_list2.
For such queries order lists were concatenated and limit clause was
completely neglected.
After FLUSH STATUS max_used_connections was reset to 0, and haven't
been updated while cached threads were reused, until the moment a new
thread was created.
The first suggested fix from original bug report was implemented:
a) On flushing the status, set max_used_connections to
threads_connected, not to 0.
b) Check if it is necessary to increment max_used_connections when
taking a thread from the cache as well as when creating new threads
counter".
When TRUNCATE TABLE was called within an stored procedure the
auto_increment counter was not reset to 0 even if straight
TRUNCATE for this table did this.
This fix makes TRUNCATE in stored procedures to be handled exactly
in the same way as straight TRUNCATE. We achieve this by rolling
back the fix for bug 8850, which is no longer needed since stored
procedures don't require prelocked mode anymore (and TRUNCATE is
not allowed in stored functions or triggers).
The idea is to add DEFINER-clause in CREATE PROCEDURE and CREATE FUNCTION
statements. Almost all support of definer in stored routines had been already
done before this patch.
NOTE: this patch changes behaviour of dumping stored routines in mysqldump.
Before this patch, mysqldump did not dump DEFINER-clause for stored routines
and this was documented behaviour. In order to get full information about stored
routines, one should have dumped mysql.proc table. This patch changes this
behaviour, so that DEFINER-clause is dumped.
Since DEFINER-clause is not supported in CREATE PROCEDURE | FUNCTION statements
before this patch, the clause is covered by additional version-specific comments.
if --skip-grant-tables specified.
The problem is that there is a check that prevents creating a definer
with empty host name.
In --skip-grant-tables mode this check prevents the user from creating a
trigger/view without explicitly specifying its definer. This happens, because
in --skip-grant-tables mode CURRENT_USER is ''@''. According to Sanja this
check was implemented intentionally.
However, according to the MySQL manual it is possible to specify empty host
name (as well as empty user name). Moreover, the behaviour for stored routines
is different in this aspect -- we allow them to be created with implicit
definer.
Based on this, we believe it is OK to change the behaviour for views to be
similar with the behaviour for stored routines.
The idea of the fix is to extend support of non-SUID triggers for backward
compatibility. Formerly non-SUID triggers were appeared when "new" server
is being started against "old" database. Now, they are also created when
"new" slave receives updates from "old" master.
- Added empty constructors and virtual destructors to many classes and structs
- Removed some usage of the offsetof() macro to instead use C++ class pointers
column is increasing when table is recreated with PS/SP":
make use of create_field::char_length more consistent in the code.
Reinit create_field::length from create_field::char_length
for every execution of a prepared statement (actually fixes the
bug).
After trying multiple inheritance (to messy and hard make it work) and
sublassing jump_if_not (worked, but ugly), decided to on this solution
instead:
Inserting an abstract sp_instr_opt_meta class as parent for all instructions
with destinations makes it possible to handle a continuation pointer for
sp_instr_set_case_expr too.
Note: No special test case; the fix is captured by the changed behaviour of
bug14643_2, and bug14498_4 (formerly disabled), in sp.test.
Since replication rules execute after `mysql_multi_update_prepare' returns we
delay to `break' in case this functions returns non-zero (some tables are not found)
for to examine if there is an ignore rule for a not-found table. By doing that
it is guaranteed do/ignore replication rules logically preceed opening table routine.
There are two main idea of this fix:
- introduce a common function for server and client to split user value
(<user name>@<host name>) into user name and host name parts;
- dump DEFINER clause in correct format in mysqldump.
- Fixed tests
- Optimized new code
- Fixed some unlikely core dumps
- Better bug fixes for:
- #14397 - OPTIMIZE TABLE with an open HANDLER causes a crash
- #14850 (ERROR 1062 when a quering a view using a Group By on a column that can be null
Problem #1: INSERT...SELECT, Version for 5.0.
Extended the unique table check by a check of lock data.
Merge sub-tables cannot be detected by doing name checks only.
Problem #1: INSERT...SELECT, Version for 4.1.
INSERT ... SELECT with the same table on both sides (hidden
below a MERGE table) does now work by buffering the select result.
The duplicate detection works now after open_and_lock_tables()
on the locks.
I did not find a test case that failed without the change in
sql_update.cc. I made the change anyway as it should in theory
fix a possible MERGE table problem with multi-table update.
according to the standard.
The idea is to use Field-classes to implement stored routines
variables. Also, we should provide facade to Item-hierarchy
by Item_field class (it is necessary, since SRVs take part
in expressions).
The patch fixes the following bugs:
- BUG#8702: Stored Procedures: No Error/Warning shown for inappropriate data
type matching;
- BUG#8768: Functions: For any unsigned data type, -ve values can be passed
and returned;
- BUG#8769: Functions: For Int datatypes, out of range values can be passed
and returned;
- BUG#9078: STORED PROCDURE: Decimal digits are not displayed when we use
DECIMAL datatype;
- BUG#9572: Stored procedures: variable type declarations ignored;
- BUG#12903: upper function does not work inside a function;
- BUG#13705: parameters to stored procedures are not verified;
- BUG#13808: ENUM type stored procedure parameter accepts non-enumerated
data;
- BUG#13909: Varchar Stored Procedure Parameter always BINARY string (ignores
CHARACTER SET);
- BUG#14161: Stored procedure cannot retrieve bigint unsigned;
- BUG#14188: BINARY variables have no 0x00 padding;
- BUG#15148: Stored procedure variables accept non-scalar values;
Problem #1: INSERT...SELECT
INSERT ... SELECT with the same table on both sides (hidden
below a MERGE table) does now work by buffering the select result.
The duplicate detection works now after open_and_lock_tables()
on the locks.
I did not find a test case that failed without the change in
sql_update.cc. I made the change anyway as it should in theory
fix a possible MERGE table problem with multi-table update.
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).
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.
handling of savepoints in stored routines.
Fixed ha_rollback_to_savepoint()/ha_savepoint()/ha_release_savepoint()
functions to properly handle savepoints inside of stored functions and
triggers.
Also now when we invoke stored function or trigger we create new savepoint
level. We destroy it at the end of function/trigger execution and return back
to old savepoint level.
Since long, the compiled code of stored routines has been printed in the trace file
when starting mysqld with the "--debug" flag. (At creation time only, and only in
debug builds of course.) This has been helpful when debugging stored procedure
execution, but it's a bit awkward to use. Also, the printing of some of the
instructions is a bit terse, in particular for sp_instr_stmt where only the command
code was printed.
This improves the printout of several of the instructions, and adds the debugging-
only commands "show procedure code <name>" and "show function code <name>".
(In non-debug builds they are not available.)
we changing current db temporarily and restore it when sp is created. however thd->db
in this case becomes empty string rather than NULL and so all checks of thd->db == NULL
will be false. So if after this we'll issue create procedure sp2()... without specifying
db it will succeed and create sp with db=NULL, which causes mysqldto crash on
show procedure status statement.
This patch fixes the problem.
it is added a check of not being empty value. When modifying SP with Admin
application on win32 it does not pass curent database so sp is stored with
db=null which causes a crash later on show procedure status;
Indeed now that stored procedures CALL is not binlogged, but instead the invoked substatements are,
the restrictions applied by log-bin-trust-routine-creators=0 are superfluous for procedures.
They still need to apply to functions where function calls are written to the binlog (for example as "DO myfunc(3)").
We rename the variable to log-bin-trust-function-creators but allow the old name until some future version (and issue a warning if old name is used).
the READ_ONLY global variable now allows statements which are to update only temporary tables
(note: if a statement, after parse stage, looks like it will update a non-temp table, it will be rejected,
even if at execution it would have turned out that 0 rows would be updated; for example
UPDATE my_non_tem_table SET a=1 WHERE 1 = 0; will be rejected).
- 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
allow select into outfile from I_S tables
it is enough to add FILE_ACL for I_S tables only to 'check_table_access' function
as we use 'any_db' for 'check_access' function in places where FILE_ACL is required
Ensure that ccache is also used for C programs
mysql: Ensure that 'delimiter' works the same way in batch mode as in normal mode
mysqldump: Change to use ;; (instead of //) as a stored procedure/trigger delimiter
Fixed test cases by adding missing DROP's and rename views to be of type 'v#'
Removed MY_UNIX_PATH from fn_format()
Removed current_db_used from TABLE_LIST
Removed usage of 'current_thd' in Item_splocal
Removed some compiler warnings
A bit faster longlong2str code
--with-embedded-privilege-control options". One more (hopefully last) build
failure which was introduced during work on WL#2787 "Add view definer/owner
to the view definition..."
present): the problem originally was that the tables in auxilliary_tables did not have
the correct real_name, which caused problems in the second call to tables_ok().
The fix corrects the real_name problem, and also sets the updating flag properly,
which makes the second call to tables_ok() unnecessary.
Information_schema DB
Bug#9846 Inappropriate error displayed while
dropping table from 'INFORMATION_SCHEMA'
Bug#10734 Grant of privileges other than 'select' and
'create view' should fail on schema
Bug#10708 SP's can use INFORMATION_SCHEMA as ROUTINE_SCHEMA
cumulative fix for bugs above(after review, 2nd version)
added privilege check for information schema db & tables
The problem was that in the first production in rule 'join_table', that
processes simple cross joins, the parser was processing the second join operand
before the first one due to unspecified priorities of JOINs. As a result in the
case of cross joins the parser constructed a tree with incorrect nesting:
the expression "t1 join t2 join t3 on some_cond" was interpreted as
"t1 join (t2 join t3 on some_cond)" instead of
"(t1 join t2) join t3 on some_cond".
Because of this incorrect nesting the method make_join_on_context picked an
incorrect table as the first table of the name resolution context.
The solution assignes correct priorities to the related production.
* Allocate thd->user_var_events elements on appropriate mem_root
* If several SP statements are binlogged as a single statement, collect all user var
accesses they make (grep for StoredRoutinesBinlogging for details)
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.
multi-threaded environment".
To avoid deadlocks between several simultaneously run account management
commands (particularly between FLUSH PRIVILEGES/SET PASSWORD and GRANT
commands) we should always take table and internal locks during their
execution in the same order. In other words we should first open and lock
privilege tables and only then obtain acl_cache::lock/LOCK_grant locks.
sql_parse.cc:
mysqld.cc:
Added --skip-client-character-set-handshake.
When this option is activated, client side character set
(which is sent in handshake) is ignored, and server side
default-character-set value is used for character_set_client
and character_set_results, thus reprodicing 4.0 behaviour.
"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.
We should not assume that "thd" argument of reload_acl_and_cache() is
non-zero. Failure to do so will cause server to crash when one sends
SIGHUP to it.
- Corrected problem with N-way nested natural joins in PS mode.
- Code cleanup
- More asserts to check consistency of name resolution contexts
- Fixed potential memory leak of name resolution contexts
This allows us to use statement replication with functions and triggers
The following things are fixed with this patch:
- NOW() and automatic timestamps takes the value from the main event for functions and triggers (which allows these to replicate with statement level logging)
- No side effects for triggers or functions with auto-increment values(), last_insert_id(), rand() or found_rows()
- Triggers can't return result sets
Fixes bugs:
#12480: NOW() is not constant in a trigger
#12481: Using NOW() in a stored function breaks statement based replication
#12482: Triggers has side effects with auto_increment values
#11587: trigger causes lost connection error
