This was done to make it clear that a update_row() should not change the
row.
This was not done for handler::write_row() as this function still needs
to update auto_increment values in the row. This should at some point
be moved to handler::ha_write_row() after which write_row can also have
const arguments.
Fixed the bug by failing the statement with an error message that explains
that an auto-increment column may not be used in an expression for a
check constraint.
Added a test case in check_constraint.test.
Updated existing tests and results.
When master_use_gtid=no, the IO thread loads the slave GTID state from
the master during connect. This races with the SQL thread when
gtid_ignore_duplicates=1. If an event is in the relay log from before
the new connect and has not been applied yet, moving the slave
position causes the SQL thread to think that event should be skipped
due to gtid_ignore_duplicates=1.
This patch simply disables gtid_ignore_duplicates when not using GTID,
which seems to be what one would expect.
Problem was that we got an error in sequence_insert while opening the
newly created sequence table in an prepared statement as the table id
didn't match.
Fixed by temporarly removing the reprepare observer during
sequence_insert as there can never be a table missmatch in this case.
In case of error on opening VIEW (absent table for example) it is still possible to print its definition but some variable is not set (table_list->derived->derived) so it is better do not try to test it when there is safer alternative (table_list itself).
Working features:
CREATE OR REPLACE [TEMPORARY] SEQUENCE [IF NOT EXISTS] name
[ INCREMENT [ BY | = ] increment ]
[ MINVALUE [=] minvalue | NO MINVALUE ]
[ MAXVALUE [=] maxvalue | NO MAXVALUE ]
[ START [ WITH | = ] start ] [ CACHE [=] cache ] [ [ NO ] CYCLE ]
ENGINE=xxx COMMENT=".."
SELECT NEXT VALUE FOR sequence_name;
SELECT NEXTVAL(sequence_name);
SELECT PREVIOUS VALUE FOR sequence_name;
SELECT LASTVAL(sequence_name);
SHOW CREATE SEQUENCE sequence_name;
SHOW CREATE TABLE sequence_name;
CREATE TABLE sequence-structure ... SEQUENCE=1
ALTER TABLE sequence RENAME TO sequence2;
RENAME TABLE sequence TO sequence2;
DROP [TEMPORARY] SEQUENCE [IF EXISTS] sequence_names
Missing features
- SETVAL(value,sequence_name), to be used with replication.
- Check replication, including checking that sequence tables are marked
not transactional.
- Check that a commit happens for NEXT VALUE that changes table data (may
already work)
- ALTER SEQUENCE. ANSI SQL version of setval.
- Share identical sequence entries to not add things twice to table list.
- testing insert/delete/update/truncate/load data
- Run and fix Alibaba sequence tests (part of mysql-test/suite/sql_sequence)
- Write documentation for NEXT VALUE / PREVIOUS_VALUE
- NEXTVAL in DEFAULT
- Ensure that NEXTVAL in DEFAULT uses database from base table
- Two NEXTVAL for same row should give same answer.
- Oracle syntax sequence_table.nextval, without any FOR or FROM.
- Sequence tables are treated as 'not read constant tables' by SELECT; Would
be better if we would have a separate list for sequence tables so that
select doesn't know about them, except if refereed to with FROM.
Other things done:
- Improved output for safemalloc backtrack
- frm_type_enum changed to Table_type
- Removed lex->is_view and replaced with lex->table_type. This allows
use to more easy check if item is view, sequence or table.
- Added table flag HA_CAN_TABLES_WITHOUT_ROLLBACK, needed for handlers
that want's to support sequences
- Added handler calls:
- engine_name(), to simplify getting engine name for partition and sequences
- update_first_row(), to be able to do efficient sequence implementations.
- Made binlog_log_row() global to be able to call it from ha_sequence.cc
- Added handler variable: row_already_logged, to be able to flag that the
changed row is already logging to replication log.
- Added CF_DB_CHANGE and CF_SCHEMA_CHANGE flags to simplify
deny_updates_if_read_only_option()
- Added sp_add_cfetch() to avoid new conflicts in sql_yacc.yy
- Moved code for add_table_options() out from sql_show.cc::show_create_table()
- Added String::append_longlong() and used it in sql_show.cc to simplify code.
- Added extra option to dd_frm_type() and ha_table_exists to indicate if
the table is a sequence. Needed by DROP SQUENCE to not drop a table.
This happens because the master writes a table_map event to the binary log, but no row event.
The slave has a check that there should always be a row event if there was a table_map event, which
causes a crash.
Fixed by remembering in the cache what kind of events are logged
and ignore cached statements which is just a table map event.
Annotate_rows_log_event again. When a new annotate event comes,
the server applies it first (which backs up thd->query_string),
then frees the old annotate event, if any. Normally there isn't.
But with sub-statements (e.g. triggers) new annotate event comes
before the first one is freed, so the second event backs up
thd->query_string that was set by the first annotate event. Then
the first event is freed, together with its query string. And then
the second event restores thd->query_string to this freed memory.
Fix: free old annotate event before applying the new one.
automatic shortening of a too-long non-unique key should
be not a warning, but a note. It's a normal optimization,
doesn't affect correctness, and should never be converted to
an error, no matter how strict sql_mode is.
When a CTE referring to another CTE from the same with clause
was used twice then the server could not find the second CTE and
reported a bogus error message.
This happened because for any unit that was created as a clone of
a CTE specification the pointer to the WITH clause that owned this CTE
was not set.
An attempt to mark reference as dependent lead to transfering this property to
original view field and through it to other references of this field which
can't be dependent.
Parse context frames (sp_pcontext) can have holes in variable run-time offsets,
the missing offsets reside on the children contexts in such cases.
Example:
CREATE PROCEDURE p1() AS
x0 INT:=100; -- context 0, position 0, run-time 0
CURSOR cur(
p0 INT, -- context 1, position 0, run-time 1
p1 INT -- context 1, position 1, run-time 2
) IS SELECT p0, p1;
x1 INT:=101; -- context 0, position 1, run-time 3
BEGIN
...
END;
Fixing a few methods to take this into account:
- sp_pcontext::find_variable()
- sp_pcontext::retrieve_field_definitions()
- LEX::sp_variable_declarations_init()
- LEX::sp_variable_declarations_finalize()
- LEX::sp_variable_declarations_rowtype_finalize()
- LEX::sp_variable_declarations_with_ref_finalize()
Adding a convenience method:
sp_pcontext::get_last_context_variable(uint offset_from_the_end);
to access variables from the end, rather than from the beginning.
This helps to loop through the context variable array (m_vars)
on the fragment that does not have any holes.
Additionally, renaming sp_pcontext::find_context_variable() to
sp_pcontext::get_context_variable(). This method simply returns
the variable by its index. So let's rename to avoid assumptions
that some heavy lookup is going on inside.
Fixed that the Column_definition::pack_flag member corresponding to
ROW-type SP variables and their fields was not properly initialized.
This lead to sporadic test failures. Valgrind complained about jumps
depending on uninitialized value in VALGRIND builds.
This patch makes sure that sp_head::fill_spvar_definition() is always
called for ROW variables and their fields.
Additionally, fixed that a function with a scalar parameter
erroneously acceptes ROWs with one fields. Now an error is returned.
Implementing cursor%ROWTYPE variables, according to the task description.
This patch includes a refactoring in how sp_instr_cpush and sp_instr_copen
work. This is needed to implement MDEV-10598 later easier, to allow variable
declarations go after cursor declarations (which is currently not allowed).
Before this patch, sp_instr_cpush worked as a Query_arena associated with
the cursor. sp_instr_copen::execute() switched to the sp_instr_cpush's
Query_arena when executing the cursor SELECT statement.
Now the Query_arena associated with the cursor is stored inside an instance
of a new class sp_lex_cursor (a LEX descendand) that contains the cursor SELECT
statement.
This simplifies the implementation, because:
- It's easier to follow the code when everything related to execution
of the cursor SELECT statement is stored inside the same sp_lex_cursor
object (rather than distributed between LEX and sp_instr_cpush).
- It's easier to link an sp_instr_cursor_copy_struct to
sp_lex_cursor rather than to sp_instr_cpush.
- Also, it allows to perform sp_instr_cursor_copy_struct::exec_core()
without having a pointer to sp_instr_cpush, using a pointer to sp_lex_cursor
instead. This will be important for MDEV-10598, because sp_instr_cpush will
happen *after* sp_instr_cursor_copy_struct.
After MDEV-10598 is done, this declaration:
DECLARE
CURSOR cur IS SELECT * FROM t1;
rec cur%ROWTYPE;
BEGIN
OPEN cur;
FETCH cur INTO rec;
CLOSE cur;
END;
will generate about this code:
+-----+--------------------------+
| Pos | Instruction |
+-----+--------------------------+
| 0 | cursor_copy_struct rec@0 | Points to sp_cursor_lex through m_lex_keeper
| 1 | set rec@0 NULL |
| 2 | cpush cur@0 | Points to sp_cursor_lex through m_lex_keeper
| 3 | copen cur@0 | Points to sp_cursor_lex through m_cursor
| 4 | cfetch cur@0 rec@0 |
| 5 | cclose cur@0 |
| 6 | cpop 1 |
+-----+--------------------------+
Notice, "cursor_copy_struct" and "set" will go before "cpush".
Instructions at positions 0, 2, 3 point to the same sp_cursor_lex instance.
The bug was introduced in the patch for "MDEV-10597 Cursors with parameters".
The LEX created in assignment_source_expr was not put into
thd->lex->sphead->m_lex (the stack of LEX'es), so syntax error in "expr"
caused a wrong memory cleanup in sp_head::~sp_head().
The fix changes the code to use sp_head::push_lex() followed by
sp_head::restore_lex(), like it happens in all other similar cases.
Allowing qualified procedure names to be used without the CALL keyword:
BEGIN
test.p1(10);
test.p2;
END;
Note:
- COMMIT and ROLLBACK cannot be used in a direct assignment anymore:
COMMIT:= 10;
ROLLBACK:= 10;
But as they are reserved keywords in Oracle anyway, this is not a problem.
- SHUTDOWN now also cannot be used in direct a direct assignment:
SHUTDOWN:=10;
If this causes migration problems in the future, the grammar should
be modified.
Note:
Variables with names COMMIT, ROLLBACK and SHUTDOWN can still be assigned
with the SET statement, e.g. SET COMMIT=10;
Part 2:
Moving the part of Sql_condition that contain condition items
(such as m_class_origin, m_cursor_name, etc) into a separate
class Sql_condition_items. This allows to remove duplicate code in
different Sql_condition constructors.
Also, introducing new Sql_condition constructors and removing the method
Sql_condition::set(). All code sequences that called an Sql_condition
constructor followed by Sql_condition::set() are now replaced to
the new constructor calls. This gives light performance improvement,
as the relevant members are now initialized only one time.
An additional change for "Part 9: EXCEPTION handlers"
This construct:
EXCEPTION WHEN OTHERS THEN ...;
now catches warning-alike conditions, e.g. NO_DATA_FOUND.
The crash happened because of a wrong reset_lex() .. restore_lex() sequence.
The Item in WHERE clause and the corresponding sp_instr_jump_if_not() were
erroneously created using different LEX.
This is a fix for "MDEV-10580 sql_mode=ORACLE: FOR loop statement"
The tokenizer now treats digits followed by two dots (e.g. '1..')
as an integer number '1' followed by DOT_DOT_SYM.
Previously this sequence was treated as a double number '1.' followed by '.'.
Fixed that the ITERATE statement inside a FOR LOOP statement did not
increment the index variable before jumping to the beginning
of the loop, which caused the loop to repeat endlessly.
Adding methods:
- LEX::sp_while_loop_expression()
- LEX::sp_while_loop_finalize()
to reuse code between sql_yacc.yy and sql_yacc_ora.yy.
FOR loop will also reuse these methods.
Part 5: EXIT statement
Adding unconditional EXIT statement:
EXIT [ label ]
Conditional EXIT statements with WHERE clause
will be added in a separate patch.
Moving similar code from sql_yacc.yy and sql_yacc_ora.yy to methods:
LEX::maybe_start_compound_statement()
LEX::sp_push_loop_label()
LEX::sp_push_loop_empty_label()
LEX::sp_pop_loop_label()
LEX::sp_pop_loop_empty_label()
The EXIT statement will also reuse this code.
Moving the code from *.yy to methods:
LEX::sp_change_context()
LEX::sp_leave_statement()
LEX::sp_iterate_statement()
to reuse the same code between LEAVE and ITERATE statements.
EXIT statement will also reuse the same code.
When processing an SP body:
CREATE PROCEDURE p1 (parameters)
AS [ declarations ]
BEGIN statements
[ EXCEPTION exceptions ]
END;
the parser generates two "jump" instructions:
- from the end of "declarations" to the beginning of EXCEPTION
- from the end of EXCEPTION to "statements"
These jumps are useless if EXCEPTION does not exist.
This patch makes sure that these two "jump" instructions are
generated only if EXCEPTION really exists.
- Part 9: EXCEPTION handlers
The top-most stored routine blocks now support EXCEPTION clause
in its correct place:
AS [ declarations ]
BEGIN statements
[ EXCEPTION exceptions ]
END
Inner block will be done in a separate commit.
- Part 14: IN OUT instead of INOUT (in SP parameter declarations)
Part 13: RETURN vs RETURNS in function definition:
CREATE FUNCTION f1(a INT) RETURN INT ...
Part 12: No parentheses if no arguments:
CREATE FUNCTION f1 RETURN INT ...
Part 12: No parentheses if no arguments
Now "CREATE PROCEDURE p1 AS" is supported with no parentheses after the name.
Note, "CREATE FUNCTION f1 AS" is not supported yet, due to grammar conflict
with UDFs. Functions will be done in a separate patch.
Part 9: EXCEPTION handlers
- Adding exception handler syntax:
WHEN exception_name THEN statement
- Adding EXCEPTION section intoi the top BEGIN..END SP block.
Note, currently EXCEPTION goes in the beginning of the top BEGIN..END
SP block.
TODO:
- add EXCEPTION section into inner blocks
- move EXCEPTION to the end of the block
1. Adding const qualifiers into a few method parameters.
2. Adding methods:
- sp_label::block_label_declare()
- LEX::sp_block_init()
- LEX::sp_block_finalize()
to share more code between the files sql_yacc.yy and sql_yacc_ora.yy,
as well as between the rules sp_labeled_block, sp_unlabeled_block,
sp_unlabeled_block_not_atomic.
3. sql_yacc.yy, sql_yacc_ora.yy changes:
- Removing sp_block_content
- Reorganizing the grammar so the rules sp_labeled_block,
sp_unlabeled_block, sp_unlabeled_block_not_atomic now
contain both BEGIN_SYM and END keywords. Previously,
BEGIN_SYM and END resided in different rules.
This change makes the grammar easier to read,
as well as simplifies adding Oracle-style DECLARE section (coming soon):
DECLARE
..
BEGIN
..
END;
Good side effects:
- SP block related grammar does not use Lex->name any more.
- The "splabel" member was removed from %union
- Adding a new grammar file sql_yacc_ora.yy, which is currently
almost a full copy of sql_yacc.yy.
Note, it's now assumed that sql_yacc.yy and sql_yacc_ora.yy
use the same set of %token directives and exactly the same
%union directive.
These declarations should eventually be moved into a shared
included file, to make sure that sql_yacc.h and sql_yacc_ora.h
are compatible.
- Removing the "-p MYSQL" flag from cmake/bison.cmake, using
the %name-prefix directive inside sql_yacc.yy and sql_yacc_ora.yy instead
- Adding other CMake related changes to build sql_yacc_ora.o
form sql_yacc_ora.yy
- Adding NUMBER(M,N) as a synonym to DECIMAL(M,N) as the first
Oracle compatibility syntax understood in sql_mode=ORACLE.
- Adding prototypes to functions add_virtual_expression()
and handle_sql2003_note184_exception(), so they can be used
in both sql_yacc.yy and sql_yacc_ora.yy.
- Adding a new test suite compat/oracle, with the first test "type_number".
Use this:
./mtr compat/oracle.type_number # to run a single test
./mtr --suite=compat/oracle # to run the entire new suite
- Adding compat/oracle into the list of default suites,
so BuildBot can run it automatically on pushes.
Also fixed a wrong result for a test case for mdev-7691
(the alternative one).
The test cases for all these bug have materialized semi-joins used
inside dependent sub-queries.
The patch actually reverts the change inroduced by Monty in 2003.
It looks like this change is not valid anymore after the implementation
of semi-joins.
Adjusted output from EXPLAIN for many other test cases.
The patch actually fixes the old defect of the optimizer that
could not extract keys for range access from IN predicates
with row arguments.
This problem was resolved in the mysql-5.7 code. The patch
supersedes what was done there:
- it can build range access when not all components of
the first row argument are refer to the columns of the table
for which the range access is constructed.
- it can use equality predicates to build range access
to the table that is not referred to in this argument.
To export symbols from the mysqld.exe, use lib.exe with /DEF, rather than
pre-link step when building mysqld.exe.
This helps to avoid relinking all plugins, if mysqld.exe was recompiled
but the list of its exports has not changed.
Also removed unnecessary DEPENDS in some ADD_CUSTOM_COMMAND (gen_lex_token,
gen_lex_hash etc). They confuse VS generator which tends to
recreate headers and do unnecessary recompilations.
Added CHECK constraints to I_S.TABLE_CONSTRAINTS.
Fixed a bug regarding virtual column definitions whose name is the field name.
Added test case: check_constraint_show
The function Item::split_sum_func2() incorrectly processed the function
items with window functions that were not window functions themselfes
and were used as arguments of other functions.
The function st_select_lex_unit::exec_recursive() incorrectly determined
that a CTE mutually recursive with some others was stabilized in the case
when the non-recursive part of the CTE returned an empty set. As a result
the server fell into an infinite loop when executing a query using
this CTE.
Mutually recursive CTE could cause a crash of the server in the case
when they were not Standard compliant. The crash happened in
mysql_derived_prepare(), because the destructor the derived_result
object created for a CTE that was mutually recursive with some others
was called twice. Yet this destructor should not be called for resursive
references.
The method With_element::check_unrestricted_recursive() icorrectly performed
the check that no recursive reference is not encountered in inner parts of
outer joins. As a result the server reported errors for valid specifications
with outer joins.
Terminate idle transactions safely in server layer by setting up socket timeout
parameter. Percona provides another patch to resolve similar problem, but it
calls server layer's callback in InnoDB plugin to close THD, which crashes in
some testcases. See https://bugs.launchpad.net/percona-server/+bug/901060 for
more detailed information.
1. export parameter trx_idle_timeout to handle all kinds of transactions, the priority is highest
2. export parameter trx_readonly_idle_timeout to handle read-only transactions
3. export parameter trx_changes_idle_timeout to handle read-write transactions
Extended syntax so that it is now possible to set lock_wait_timeout for the
following statements:
SELECT ... FOR UPDATE [WAIT n|NOWAIT]
SELECT ... LOCK IN SHARED MODE [WAIT n|NOWAIT]
LOCK TABLE ... [WAIT n|NOWAIT]
CREATE ... INDEX ON tbl_name (index_col_name, ...) [WAIT n|NOWAIT] ...
ALTER TABLE tbl_name [WAIT n|NOWAIT] ...
OPTIMIZE TABLE tbl_name [WAIT n|NOWAIT]
DROP INDEX ... [WAIT n|NOWAIT]
TRUNCATE TABLE tbl_name [WAIT n|NOWAIT]
RENAME TABLE tbl_name [WAIT n|NOWAIT] ...
DROP TABLE tbl_name [WAIT n|NOWAIT] ...
Valid range of lock_wait_timeout and innodb_lock_wait_timeout was extended so
that 0 is acceptable value (means no wait).
This is amended AliSQL patch. We prefer Oracle syntax for [WAIT n|NOWAIT]
instead of original [WAIT [n]|NO_WAIT].
When "simulate_failed_connection_1" DBUG keyword is set, current thread's
DBUG stack points to global variable in dbug (init_settings).
This global variable could be overriden while current THD is still active,
producing crash in worst scenario.
Added DBUG_SET()so that the current thread own dbug
context that cannot concurrently modified by anyone else.
- Before this patch during startup all slave threads was started without
any check that they had started properly.
- If one did a START SLAVE, STOP SLAVE or CHANGE MASTER as first command to the server
there was a chance that server could access structures that where not
properly initialized which could lead to crashes in
Log_event::read_log_event
- Fixed by waiting for slave threads to start up properly also during
server startup, like we do with START SLAVE.
The following is an updated commit message for the following commit
that was pushed before I had a chance to update the commit message:
c5e25c8b40
Fixed dead locks when doing stop slave while slave was starting.
- Added a separate lock for protecting start/stop/reset of a specific slave.
This solves some possible dead locks when one calls stop slave while
the slave is starting as the old run_locks was over used for other things.
- Set hash->records to 0 before calling free of all hash elements.
This was set to stop concurrent threads to loop over hash elements and
access members that was already freed.
This was a problem especially in start_all_slaves/stop_all_slaves
as the mutex protecting the hash was temporarily released while a slave
was started/stopped.
- Because of change to hash->records during hash_reset(),
any_slave_sql_running() will return 1 during shutdown as one can't
loop over master_info_index->master_info_hash while hash_reset() of it
is in progress.
This also fixes a potential old bug in any_slave_sql_running() where
during shutdown and ~Master_info_index(), my_hash_free() we could
potentially try to access elements that was already freed.
* fix 32bit redefine of char_to_byte_length_safe
Error can be seen here: https://hastebin.com/azomatocoq.scala
sql/item.h: defined uint32 char_to_byte_length_safe(size_t, uint32)
sql/sql_type.h: defined uint32 char_to_byte_length_safe(uint32, uint32)
As @knielsen observed, C++ overloading allows this on 64bit
because size_t and uint32 differ.
Since item.h does a #include "sql_type.h", there is no point in
redifining this function in item.h
* avoid -Wconversion warnings, fix spacing; thanks @svoj
From IRC:
<svoj> EricHerman: nice, thanks! I believe we also should use function body as in item.h. It was fixed by dr-m recently.
<EricHerman> svoj, so you suggest the "return tmp > UINT_MAX32 ? UINT_MAX32 : static_cast<uint32>(tmp);"
<EricHerman> I can use that, sure.
<svoj> EricHerman: yes, please! And fix indentation on these lines (should be 2 spaces instead of 3).
<svoj> EricHerman: the patch where this hunk comes from: git show 89d80c1b0 sql/item.h
In get_mm_tree we have to change Field_geom::geom_type to
GEOMETRY as we have to let storing all types of the spatial features
in the field. So now we restore the original geom_type as it's
done.
failed with SELECT SQ, TEXT field
The functon find_all_keys does call Item_subselect::walk, which calls walk() for the subquery
The issue is that when a field is represented by Item_outer_ref(Item_direct_ref(Item_copy_string( ...))).
Item_copy_string does have a pointer to an Item_field in Item_copy::item but does not implement Item::walk method, so we are not
able to set the bitmap for that field. This is the reason why the assert fails.
Fixed by adding the walk method to Item_copy class.
character (").
The my_wildcmp function doesn't expect the string parameter to
have escapements, only the template. So the string
should be unescaped if necessary.
