Remove a bogus debug assertion that triggered the bug.
Add assertions precisely where records must not be delete-marked.
And a comment to clarify when the record is allowed to be delete-marked.
'CREATE TABLE IF NOT EXISTS ... SELECT' behaviour
BUG#55474, BUG#55499, BUG#55598, BUG#55616 and BUG#55777 are fixed
in this patch too.
This is the 5.1 part.
It implements:
- if the table exists, binlog two events: CREATE TABLE IF NOT EXISTS
and INSERT ... SELECT
- Insert nothing and binlog nothing on master if the existing object
is a view. It only generates a warning that table already exists.
mysql-test/r/trigger.result:
Ather this patch, 'CREATE TABLE IF NOT EXISTS ... SELECT' will not
insert anything if the creating table already exists and is a view.
sql/sql_class.h:
Declare virtual function write_to_binlog() for select_insert.
It's used to binlog 'create select'
sql/sql_insert.cc:
Implement write_to_binlog();
Use write_to_binlog() instead of binlog_query() to binlog the statement.
if the table exists, binlog two events: CREATE TABLE IF NOT EXISTS
and INSERT ... SELECT
sql/sql_lex.h:
Declare create_select_start_with_brace and create_select_pos.
They are helpful for binlogging 'create select'
sql/sql_parse.cc:
Do nothing on master if the existing object is a view.
sql/sql_yacc.yy:
Record the relative postion of 'SELECT' in the 'CREATE ...SELECT' statement.
Record whether there is a '(' before the 'SELECT' clause.
* no automatic --loose-skip-innodb added by mtr based on the test name.
instead loose-skip-innodb is now in the default_mysqld.cnf
* have_innodb_plugin.inc is changed to give a verbose "skip" message
(instead of "require: true")
* My::Suite class. It's support in mtr, and everywhere
* support for suite.pm
* when sorting tests, take combinations into account
* support for SUITENAME_COMBINATIONS
* no special treatment for innodb_plugin in mtr_cases.pm
* two special pre-created config groups: ENV and OPT
* allow option names to start from #
* allow magic option to have an argument
* remove dead code
* fix @-substitution to works as expected
* new processes take the value of $opt_verbose automatically, no need to pass it to a constructor
* innodb_plugin suite uses suite.pm and combinations file to test as much as possible
(innodb plugin, xtradb plugin, xtradb static - whatever available)
* besides test-master.opt and test-slave.opt a test.opt file is also
loaded, both for master and slave
* .opt files for all included files are loaded too
* progress report in the xterm titlebar
An user assignment variable expression that's
evaluated in a logical expression context
(Item::val_bool()) can be pre-calculated in a
temporary table for GROUP BY.
However when the expression value is used after the
temp table creation it was re-evaluated instead of
being read from the temp table due to a missing
val_bool_result() method.
Fixed by implementing the method.
The server was not checking for errors generated during
the execution of Item::val_xxx() methods when copying
data to the group, order, or distinct temp table's row.
Fixed by extending the copy_funcs() to return an error
code and by checking for that error code on the places
copy_funcs() is called.
Test case added.
variable assignments
The assert() that is firing is checking if expressions that can't be
null return a NULL when evaluated.
MAKEDATE() function can return NULL if the second argument is
less then or equal to 0. Thus its nullability depends not only on
the nullability of its arguments but also on their values.
Fixed by (overoptimistically) setting MAKEDATE() to be nullable
despite the nullability of its arguments.
Test added.
Had to update one test result to reflect the metadata change.
Print out checked file names in maria_check -s (unless you use a second -s)
Some trivial optimizations
storage/maria/ma_bitmap.c:
Trivial optimizations:
- Combine common code (to be able to remove duplicate mutex_lock call)
- Move setting of thread specific variables outside of mutex
storage/maria/ma_check.c:
Fixed wrong argument to printf
storage/maria/maria_chk.c:
Print out checked table names unless -s -s
storage/maria/maria_read_log.c:
Added option --start-from-checkpoint (to help find bugs in checkpoints)
feature
The test for bug no 50939 was put in range.test which isn't such a good idea
since it requires partitioning. Fixed by moving the test case to
partitioning_range.test.
client/mysqlcheck.c:
Added missing casts
client/mysqldump.c:
Added missing casts
client/mysqlimport.c:
Added missing casts
extra/my_print_defaults.c:
Added missing casts
mysql-test/mysql-test-run.pl:
Added suppression for non-critical warning on windows
storage/maria/maria_pack.c:
Added missing casts
storage/xtradb/buf/buf0lru.c:
Added missing casts
storage/xtradb/fil/fil0fil.c:
Added missing casts
storage/xtradb/handler/i_s.cc:
Added extra argument to call store() function for longlong.
storage/xtradb/srv/srv0srv.c:
Added cast to suppress compiler warning
support-files/compiler_warnings.supp:
Added suppression for some non critical compiler warnings on Windows
unittest/mytap/tap.h:
Fixed prototypes to be same as the actual functions
maria_chk & maria_read_log now reads block size from control file.
mysql-test/suite/maria/r/maria.result:
Updated results after trivial change of maria_chk's output
storage/maria/ma_bitmap.c:
More DBUG_PRINT
storage/maria/ma_blockrec.c:
Fixed bug that we didn't mark page full in bitmap if directory is full
storage/maria/ma_check.c:
Write out if directory is full for errors in bitmap
storage/maria/ma_control_file.c:
Don't give error for wrong block size if block size is 0
storage/maria/maria_chk.c:
Read block size from control file
In case of -dvv, write also out bitmap information (good for debugging)
storage/maria/maria_read_log.c:
Read block size from control file
Fixed that maria_read_log works with different page size than TRANSLOG_PAGE_SIZE
Added straight_join to make results predicatable
mysql-test/suite/pbxt/r/range.result:
Added straight_join to make results predicatable
mysql-test/suite/pbxt/t/range.test:
Added straight_join to make results predicatable
mysys/my_sync.c:
Fixed typo
Fixed LP#613418 (M)aria recovery failure: ma_key_recover.c:981: _ma_apply_redo_index: Assertion `check_page_length == page_length' failed
include/my_sys.h:
Added my_disable_sync
mysql-test/mysql-test-run.pl:
Added --sync-sys=0 to run test suite faster
mysys/my_static.c:
Added my_disable_sync
mysys/my_sync.c:
Added my_disable_sync
sql/mysqld.cc:
Added -sync-sys=0 option for mysqld to skip sync() calls for faster testing
storage/maria/ma_key_recover.c:
More DBUG_ASSERT()
Added logging of KEY_OP_DEBUG to make examening of logs easier
Fixed testing of page length in recovery to ensure we don't overwrite checksum (previous tests was too relaxed)
Fixed bug in recovery logging of split pages which caused failure during recovery:
- Length was not adjusted properly for pages to be split
- Added KEY_OP_MAX_PAGELENGTH to tell recovery that page is now full length
- This fixed LP#613418
storage/maria/ma_key_recover.h:
Changed prototype for ma_log_change() for KEY_OP_DEBUG
storage/maria/ma_loghandler.h:
Added new enums for better debugging of recovery logs
storage/maria/ma_rt_index.c:
Added debugging information to calls to ma_log_change()
storage/maria/ma_write.c:
Added debugging information to calls to ma_log_change() and ma_log_split()
INSERT IGNORE ... SELECT ... UNION SELECT ...
This assert was triggered by INSERT IGNORE ... SELECT. The assert checks that a
statement either sends OK or an error to the client. If the bug was triggered
on release builds, it caused OK to be sent to the client instead of the correct
error message (in this case ER_FIELD_SPECIFIED_TWICE).
The reason the assert was triggered, was that lex->no_error was set to TRUE
during JOIN::optimize() because of IGNORE. This causes all errors to be ignored.
However, not all errors can be ignored. Some, such as ER_FIELD_SPECIFIED_TWICE
will cause the INSERT to fail no matter what. But since lex->no_error was set,
the critical errors were ignored, the INSERT failed and neither OK nor the
error message was sent to the client.
This patch fixes the problem by temporarily turning off lex->no_error in
places where errors cannot be ignored during processing of INSERT ... SELECT.
Test case added to insert.test.
Fixed compiler warnings
Disabled some tests that doesn't work on windows (uses shell tools or strange characters)
client/mysqlshow.c:
Fixed compiler warnings
client/mysqlslap.c:
Fixed compiler warnings
mysql-test/mysql-test-run.pl:
Use an error file instead of /dev/null
(Fixes problem on Windows and the output may be usefull)
mysql-test/suite/maria/r/maria.result:
Test case for bug LP#613408 (not complete)
mysql-test/suite/maria/t/maria.test:
Test case for bug LP#613408 (not complete)
mysql-test/suite/percona/percona_log_slow_slave_statements-and-use_global_long_query_time.test:
This test doesn't work one windows (needs shell tools)
mysql-test/suite/percona/percona_log_slow_slave_statements.test:
This test doesn't work one windows (needs shell tools)
mysql-test/suite/percona/percona_slow_query_log-control_global_slow.test:
This test doesn't work one windows (needs shell tools)
mysql-test/suite/percona/percona_slow_query_log-log_slow_filter.test:
This test doesn't work one windows (needs shell tools)
mysql-test/suite/percona/percona_slow_query_log-log_slow_verbosity.test:
This test doesn't work one windows (needs shell tools)
mysql-test/suite/percona/percona_slow_query_log-long_query_time.test:
This test doesn't work one windows (needs shell tools)
mysql-test/suite/percona/percona_slow_query_log-microseconds_in_slow_query_log.test:
This test doesn't work one windows (needs shell tools)
mysql-test/suite/percona/percona_slow_query_log-min_examined_row_limit.test:
This test doesn't work one windows (needs shell tools)
mysql-test/suite/percona/percona_slow_query_log-use_global_long_query_time.test:
This test doesn't work one windows (needs shell tools)
mysql-test/t/ctype_filesystem.test:
This test doesn't work one windows (problem with character sets)
mysql-test/t/events_time_zone.test:
Use longer times to get predictable tests
mysql-test/t/show_check-master.opt:
set long query time to get more predictable tests
storage/maria/ma_check.c:
Restore info->s->lock_key_trees after repair.
Disable logging to temp tables for all repair cases (safety fix)
storage/maria/ma_state.c:
Ensurethat info->state_start doesn't point to freed memory. (Could happen after running an internal repair to fast create indexes)
storage/maria/trnman.c:
Added longer comment
The CONVERT_TZ function crashes the server when the
timezone argument is an empty SET field value.
1) The CONVERT_TZ may find a timezone string in the
tz_names hash.
2) A string representation of the empty SET is a
String of zero length with the NULL pointer.
3) If the key argument length is zero, hash functions
do comparison using the length of the record being
compared against.
I.e. a zero-length String buffer is an invalid
argument for hash search functions, and if String
points to NULL buffer, hashcmp() fails with SEGV
accessing that memory.
The my_tz_find function has been modified to
treat empty Strings as invalid timezone values
to skip unnecessary hash search.
mysql-test/r/timezone2.result:
Test case for bug #55424.
mysql-test/t/timezone2.test:
Test case for bug #55424.
sql/sql_string.h:
Bug #55424: convert_tz crashes when fed invalid data
Added "const" modifier to String::is_empty().
sql/tztime.cc:
Bug #55424: convert_tz crashes when fed invalid data
The my_tz_find function has been modified to
treat empty Strings as invalid timezone values
to skip unnecessary hash search.
Fixed compiler warnings
client/mysqlslap.c:
Fixed compiler warnings
mysql-test/suite/maria/r/maria.result:
Test case for LP#614265
mysql-test/suite/maria/t/maria.test:
Test case for LP#614265
mysql-test/suite/pbxt/t/skip_name_resolve-master.opt:
Ensure that we get restart before test (as test uses show processlist)
sql/handler.cc:
Added cloned marker if clone was called (for safety checks & debugging)
sql/handler.h:
Added cloned marker if clone was called (for safety checks & debugging)
storage/maria/ha_maria.cc:
In clone call, set file->trn if cloned file had this set. This is needed as maria_create_trn_for_mysql() and thus file->trn is never set for cloned table.
Ensure that file->trn is properly reset after calls to repair/check/zerofill.
Increment locked table count if file->trn is set (as we decrement this in the unlock call)
tests/mysql_client_test.c:
Fixed compiler warnings
Added option: --warning-for-wrong-transaction-id to maria_check
mysql-test/r/innodb-use-sys-malloc.result:
Remove not used files
mysql-test/t/innodb-use-sys-malloc-master.opt:
Remove not used files
storage/maria/ma_recovery.c:
Added support for --lsn-end
storage/maria/ma_recovery.h:
New prototype
storage/maria/ma_write.c:
Added logging of KEY_OP_CHECK for all possible keyblock splits.
(Helps us find error in log file handling)
storage/maria/maria_chk.c:
Added option: --warning-for-wrong-transaction-id
--update-status now resets open count if check succeds
storage/maria/maria_read_log.c:
Added option --end-lsn for only appling part of log (good for debugging)
Changed options to use '-' instead of '_'
mysql-test/suite/funcs_1/r/is_columns_is.result:
Update after change of comment length
mysql-test/suite/funcs_1/r/is_engines.result:
Update after change of comment length
mysql-test/suite/innodb_plugin/t/disabled.def:
Disable a test that gives valgrind warnings for xtradb
mysql-test/suite/innodb_plugin/t/innodb_bug52745.test:
Test requires persian collation
Currently we do a full validation of AHI whenever check tables is
called on any table. This patch fixes this by only doing this full
check in debug versions.
bug#55716
rb://423
approved by: Marko
file .\item_subselect.cc, line 836
IN quantified predicates are never executed directly. They are rather wrapped
inside nodes called IN Optimizers (Item_in_optimizer) which take care of the
execution. However, this is not done during query preparation. Unfortunately
the LIKE predicate pre-evaluates constant right-hand side arguments even
during name resolution. Likely this is meant as an optimization.
Fixed by not pre-evaluating LIKE arguments in view prepare mode.
mysql-test/suite/percona/percona_log_slow_slave_statements.test:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-control_global_slow.result:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-control_global_slow.test:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-log_slow_filter.result:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-log_slow_filter.test:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-log_slow_verbosity.result:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-log_slow_verbosity.test:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-long_query_time.result:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-long_query_time.test:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-microseconds_in_slow_query_log.result:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-microseconds_in_slow_query_log.test:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-min_examined_row_limit.result:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-min_examined_row_limit.test:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-use_global_long_query_time.result:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/percona/percona_slow_query_log-use_global_long_query_time.test:
Replaced 'grep | wc -l' with 'grep -c' as output of wc -l is not portable (for example on Mac)
mysql-test/suite/rpl/r/rpl_stm_until.result:
Updated results
mysql-test/suite/rpl/t/rpl_stm_until.test:
Fixed random failure in test suite: master could be reading data from master-bin.00001 at the same time as test suite was changing it.
Handle overflow when reading value from SELECT MAX(C) FROM T;
Call ha_innobase::info() after initializing the autoinc value
in ha_innobase::open().
Fix for both the builtin and plugin.
rb://402
The bug is due to a double delete of a BLOB, once via:
rollback -> btr_cur_pessimistic_delete()
and the second time via purge.
The bug is in row_upd_clust_rec_by_insert(). There we relinquish ownership
of the non-updated BLOB columns in btr_cur_mark_extern_inherited_fields()
before building the row entry that will be inserted and whose contents will
be logged in the UNDO log. However, we don't set the BLOB column later to
INHERITED so that a possible rollback will not free the original row's
non-updated BLOB entries. This is because the condition that checks for
that is in :
if (node->upd_ext) {}.
node->upd_ext is non-NULL only if a BLOB column was updated and that column
is part of some key ordering (see row_upd_replace()). This results in the
non-update BLOB columns being deleted during a rollback and subsequently by
purge again.
rb://413
