This is an additional fix.
Item::val_xxx methods are supposed to use original data source and
Item::val_xxx_result methods to use the item's result field. But for the
Item_func_set_user_var class val_xxx_result methods were mapped to val_xxx
methods. This leads, in particular, to producing bad sort keys and thus
wrong order of the result set of queries with group by/order by clauses.
The set of val_xxx_result methods is added to the Item_func_set_user_var
class. It's the same as the val_xxx set of method but uses the result_field
to return a value.
mysql-test/t/user_var.test:
Corrected test case for hte bug#28494.
mysql-test/r/user_var.result:
Corrected test case for hte bug#28494.
sql/item_func.cc:
Bug#28494: Grouping by Item_func_set_user_var produces incorrect result.
The set of val_xxx_result methods is added to the Item_func_set_user_var
class. It's the same as the val_xxx set of method but uses the result_field
to return a value.
sql/item_func.h:
Bug#28494: Grouping by Item_func_set_user_var produces incorrect result.
The set of val_xxx_result methods is added to the Item_func_set_user_var
class.
using a derived table over a grouping subselect.
This crash happens only when materialization of the derived tables
requires creation of auxiliary temporary table, for example when
a grouping operation is carried out with usage of a temporary table.
The crash happened because EXPLAIN EXTENDED when printing the query
expression made an attempt to use the objects created in the mem_root
of the temporary table which has been already freed by the moment
when printing is called.
This bug appeared after the method Item_field::print() had been
introduced.
mysql-test/r/subselect.result:
Added a test case for bug #28728.
mysql-test/t/subselect.test:
Added a test case for bug #28728.
sql/sql_select.cc:
Fixed bug #28728: a crash when executing EXPLAIN EXTENDED for a query
using a derived table over a grouping subselect.
The crash happened because EXPLAIN EXTENDED when printing the query
expression made an attempt to use the objects created in the mem_root
of the temporary table which has been already freed by the moment
when printing is accomplished.
The fix in JOIN::exec() ensures using existing objects when printing
subselects for a derived tables by EXPLAIN EXTENDED.
into bk-internal.mysql.com:/data0/bk/mysql-5.0-opt
mysql-test/r/sp.result:
Auto merged
mysql-test/t/sp.test:
Auto merged
mysql-test/t/strict.test:
Auto merged
mysql-test/t/subselect3.test:
Auto merged
sql/mysqld.cc:
Auto merged
sql/sp_head.cc:
Auto merged
sql/sql_base.cc:
Auto merged
sql/sql_class.h:
Auto merged
sql/sql_table.cc:
Auto merged
sql/sql_update.cc:
Auto merged
sql/sql_view.cc:
Manual merge
Problem: we may create a deadlock committing changes in the mysql_alter_table() when
LOCK_open is set. Moreover, "in some variants of the ALTER TABLE commit
happens earlier, outside of LOCK_open, in other later - inside. It's no good, a storage
engine code that is called in between could expect a consistency - either there is a
transaction or there is not".
Fix: move the commit to happen earlier and outside of the LOCK_open.
mysql-test/r/innodb_mysql.result:
Fix for bug #28652: MySQL (with-debug=full) asserts when alter table operations
- test result.
mysql-test/t/innodb_mysql.test:
Fix for bug #28652: MySQL (with-debug=full) asserts when alter table operations
- test case.
sql/sql_table.cc:
Fix for bug #28652: MySQL (with-debug=full) asserts when alter table operations
- commit moved to happen earlier in the mysql_alter_table(),
now we commit changes at the same time as in case when a temporary
table is used.
into dsl-hkibras1-ff5dc300-70.dhcp.inet.fi:/home/elkin/MySQL/TEAM/FIXES/5.0/bug27716-multi_upd_no_binlog
mysql-test/r/innodb.result:
Auto merged
mysql-test/t/innodb.test:
Auto merged
sql/sql_update.cc:
Auto merged
corruption errors: 126,134,145
When one thread attempts to lock two (or more) tables and another
thread executes statement that aborts these locks (e.g. REPAIR
TABLE) we may get a table object with wrong lock type in a table
cache.
For example if SELECT FROM t1,t2 was aborted, subsequent INSERT
INTO t1 may be executed under read lock.
As a result we may get various table corruptions and even a server
crash.
This is fixed by resetting lock type in case lock was aborted by
another thread.
I failed to create reasonable test case for this bug.
sql/lock.cc:
If thr_multi_lock was aborted by another thread, it unlocks tables
that were locked before one that was aborted. Lock type for tables
that were after a table that was aborted preserved. Thus we need
to reset lock data in case thr_multi_lock was aborted.
Implementation of mysql_multi_update did not call multi_update::send_error method in some cases
(see the test reported on bug page and test cases in changeset).
Fixed with deploying the method, ::send_error() is refined to get binlogging code which works whenever
there is modified non-transactional table.
thd->no_trans_update.stmt flag is set in to TRUE to ease testing though being the beginning of relative
bug#27417 fix (addresses a part of those issues).
Eliminating two minor issues (small bugs) in multi_update methods.
This patch for multi-update also addresses a part of the issues reported in bug#13270,bug#23333.
mysql-test/r/innodb.result:
result changed
mysql-test/r/multi_update.result:
results changed
mysql-test/t/innodb.test:
regression test for the bug combining transactional and non-transaction tables
mysql-test/t/multi_update.test:
regression tests added - erred query must be binlogged
sql/sql_update.cc:
making a call to multi_update::send_error when mysql_select return an error;
deploying binlogging inside of ::send_error;
refining multi_update::send_eof() to mark binlogging work done when its call to ::do_updates() errs
and the query is binlogged with the error. ::send_error() will be called after all but
do not do anything;
thd->no_trans_update.stmt is corrected to be set to TRUE according to the existed code pattern.
Although this part relates to another bugs (bug#27417 etc) it is needed here for testing;
Eliminating a separate issue in multi_update::do_updates where
either one of the class members trans_safe, transactional_tables was
set after the per-table loop whereas both should be calculated during
looping.
Eliminating a separate issue in ::send_eof() where thd->no_trans_update.all won't be set TRUE when
amoung others there'd be at least one modified transactional table.
Binlogging is done to satisfy to general character bug#13270,bug#23333 so that those won't have
fixes specific for multi-update.
into gleb.loc:/home/uchum/work/bk/mysql-5.0-opt
mysql-test/r/view.result:
Merge with bug #27827 fix.
mysql-test/t/view.test:
Merge with bug #27827 fix.
The end_update() function uses the Item::save_org_in_field() function to
save original values of items into the group buffer. But for the
Item_func_set_user_var this method was mapped to the save_in_field method.
The latter function wrongly decides to use the result_field. This leads to
saving incorrect value in the grouping buffer and wrong result of the whole
query.
The can_use_result_field argument of the bool type is added to the
Item_func_set_user_var::save_in_field() function. If it is set to FALSE
then the item's result field won't be used. Otherwise it will be detected
whether the result field will be used (old behaviour).
Two wrapping functions for the function above are added to the
Item_func_set_user_var class:
the save_in_field(Field *field, bool no_conversions) - it calls the above
function with the can_use_result_field set to TRUE.
the save_org_in_field(Field *field) - same, but the can_use_result_field
is set to FALSE.
mysql-test/t/user_var.test:
Added a test case for the bug#28494: Grouping by Item_func_set_user_var
produces incorrect result.
mysql-test/r/user_var.result:
Added a test case for the bug#28494: Grouping by Item_func_set_user_var
produces incorrect result.
sql/item_func.cc:
Bug#28494: Grouping by Item_func_set_user_var produces incorrect result.
The can_use_result_field argument of the bool type is added to the
Item_func_set_user_var::save_in_field() function. If it is set to FALSE
then the item's result field won't be used. Otherwise it will be detected
whether the result field will be used (old behaviour).
sql/item_func.h:
Bug#28494: Grouping by Item_func_set_user_var produces incorrect result.
The can_use_result_field argument of the bool type is added to the
Item_func_set_user_var::save_in_field() function.
Two wrapping functions for the function above are added to the
Item_func_set_user_var class:
the save_in_field(Field *field, bool no_conversions) - it calls the above
function with the can_use_result_field set to TRUE.
the save_org_in_field(Field *field) - same, but the can_use_result_field
is set to FALSE.
ON conditions from JOIN expression were ignored at CHECK OPTION
check when updating a multi-table view with CHECK OPTION.
The st_table_list::prep_check_option function has been
modified to to take into account ON conditions at CHECK OPTION check
It was also changed to build the check option condition only once
for any update used in PS/SP.
sql/table.h:
Fixed bug #27827.
The st_table_list::check_option_processed variable has been added
to build the check option condition only once for any update used
in PS/SP.
sql/table.cc:
Fixed bug #27827.
The st_table_list::prep_check_option function has been
modified to to take into account ON conditions at CHECK OPTION check
It was also changed to build the check option condition only once
for any update used in PS/SP.
mysql-test/t/view.test:
Updated test case for bug #27827.
mysql-test/r/view.result:
Updated test case for bug #27827.
When the same VIEW was created at the master side twice,
malformed (truncated after the word 'AS') query string
was forwarded to client side, so error messages on the
master and client was different, and replication was
broken.
The mysql_register_view function call failed
too early: fields of `view' output argument of this
function was not filled yet with correct data required
for query replication.
The mysql_register_view function also copied pointers to
local buffers into a memory allocated by the caller.
sql/sql_view.cc:
Fixed bug #28244.
Checking of existence of VIEW .frm file has been
moved down to the place, when `view' argument is
completely filled with proper data.
view->query.str and view->query.md5 pointers has been
set to the NULL before return from the mysql_register_view
function.
mysql-test/t/rpl_view.test:
Updated test case for bug #28244.
mysql-test/r/rpl_view.result:
Updated test case for bug #28244.
problem #1: udf_example.so does not get built on AIX
solution#1: build it yourself using
cd sql; gcc -g -I ../include/ -I /usr/include/ -lpthread \
-shared -o udf_example.so udf_example.c; mv udf_example.so \
.libs/
problem#2 (the bug): udf_example fails because it does not
recognize the variable LD_LIBRARY_PATH when doing dl_open(),
it looks at LIBPATH
solution#2: add the library path to LIBPATH
problem#3: udf_example returns the wrong result length since
it relies on strmov to return a pointer to the end of the
string that it copies. On AIX builds, where m_string.h is not
included (m_string defines a macro expanding strmov to stpcpy),
there is a macro expanding strmov to strcpy, which returns a
pointer to the first character.
solution#3: define strmov as stpcpy.
problem#4: #2 applies on hp-ux as well, but this platform
looks at SHLIB_PATH
solution#4: added the library path to SHLIB_PATH
mysql-test/mysql-test-run.pl:
bug#27741: Added library paths to LIBPATH ( shared library
path environment variable used on AIX) and
SHLIB_PATH (ditto on hp-ux)
sql/udf_example.c:
bug#27741: define strmov as stpcpy rather than strcpy and
include string.h. The definition for strmov in the standalone
case was not maintained properly
Problem:
HASH indexes on VARCHAR columns with binary collations did not ignore trailing spaces from strings before comparisons. This could result in duplicate records being successfully inserted into a MEMORY table with unique key constraints.
As a direct consequence of the above, internal MEMORY tables used for GROUP BY calculation in testcases for bug #27643 contained duplicate rows which resulted in duplicate key errors when converting those temporary tables to MyISAM. Additionally, that error was incorrectly converted to the 'table is full' error.
Solution:
- ignore trailing spaces in VARCHAR fields with binary collations when calculating hashes.
- return a proper error from create_myisam_from_heap() when conversion fails.
mysql-test/r/ctype_ucs2_def.result:
Added a testcase for bug #27643.
mysql-test/r/heap_hash.result:
Added a testcase for bug #27643.
mysql-test/t/ctype_ucs2_def.test:
Added a testcase for bug #27643.
mysql-test/t/heap_hash.test:
Added a testcase for bug #27643.
sql/sql_select.cc:
Return an appropriate error instead of 'table is full' when conversion from MEMORY to MyISAM fails.
strings/ctype-bin.c:
Added my_hash_sort_8bit_bin() which ignores trailing spaces when calculating hashes, and is now used for VARCHAR columns instead of my_hash_sort_bin().
strings/ctype-mb.c:
Ignore trailing spaces when calculating a string hash in my_hash_sort_mb_bin().
strings/ctype-ucs2.c:
Ignore trailing spaces when calculating a string hash in my_hash_sort_ucs2_bin().
mysqld crashed when a long-running explain query was killed from
another connection.
When the current thread caught a kill signal executing the function
best_extension_by_limited_search it just silently returned to
the calling function greedy_search without initializing elements of
the join->best_positions array.
However, the greedy_search function ignored thd->killed status
after a calls to the best_extension_by_limited_search function, and
after several calls the greedy_search function used an uninitialized
data from the join->best_positions[idx] to search position in the
join->best_ref array.
That search failed, and greedy_search tried to call swap_variables
function with NULL argument - that caused a crash.
sql/sql_select.cc:
Fixed bug #28598.
choose_plan(), greedy_search(), best_extension_by_limited_search()
and find_best() functions have been changed to return TRUE in case
of fatal error.
mysql-test/t/kill.test:
Updated test case for bug #28598.
mysql-test/r/kill.result:
Updated test case for bug #28598.
ENUM fields internally store their values as integers and may use integer
values as indexes to their values. Invalid values are mapped to zero value.
When storing an empty string the ENUM field fails to find an appropriate value
and tries to convert the provided string to integer. The conversion also
fails and error is returned even if the thd->count_cuted_fields is set to
CHECK_FIELD_IGNORE. This makes the range optimizer wrongly decide that an
impossible range is present.
Now the Field_enum::store() returns error while storing an empty string only
if the thd->count_cuted_fields isn't set to CHECK_FIELD_IGNORE.
sql/field.cc:
Bug#28729: Field_enum wrongly reported an error while storing an empty string.
Now the Field_enum::store() returns error while storing an empty string only
if the thd->count_cuted_fields isn't set to CHECK_FIELD_IGNORE.
mysql-test/r/type_enum.result:
Added a test case for the bug#28729: Field_enum wrongly reported an error
while storing an empty string.
mysql-test/t/type_enum.test:
Added a test case for the bug#28729: Field_enum wrongly reported an error
while storing an empty string.
This is for bug #28121.
include/m_string.h:
Got rid of log_01[], because we don't really need it.
sql/item_cmpfunc.cc:
Got rid of log_01[], because we don't really need it.
strings/strtod.c:
Got rid of log_01[], because we don't really need it.
mysql-test/r/auto_increment.result:
result fix
mysql-test/r/create.result:
result fix
mysql-test/r/insert.result:
result fix
mysql-test/r/insert_select.result:
result fix
mysql-test/r/insert_update.result:
result fix
mysql-test/r/key.result:
result fix
mysql-test/r/null.result:
result fix
mysql-test/r/null_key.result:
result fix
mysql-test/r/ps_2myisam.result:
result fix
mysql-test/r/ps_3innodb.result:
result fix
mysql-test/r/ps_4heap.result:
result fix
mysql-test/r/ps_5merge.result:
result fix
mysql-test/r/ps_6bdb.result:
result fix
mysql-test/r/ps_7ndb.result:
result fix
mysql-test/r/strict.result:
result fix
mysql-test/r/view.result:
result fix
mysql-test/r/warnings.result:
result fix
mysql-test/t/strict.test:
test fix
The result of the CHECK OPTION condition evaluation over an
updated record and records of merged tables was arbitrary and
dependant on the order of records in the merged tables during
the execution of SELECT statement.
The CHECK OPTION expression was evaluated over expired record
buffers (with arbitrary data in the fields).
Rowids of tables used in the CHECK OPTION expression were
added to temporary table rows. The multi_update::do_updates()
method was modified to restore necessary record buffers
before evaluation of the CHECK OPTION condition.
sql/sql_class.h:
Fixed bug #29716.
The multi_update::unupdatable_check_opt_tables variable
has been added.
sql/sql_update.cc:
Fixed bug #29716.
Rowids of tables used in the CHECK OPTION expression were
added to temporary table rows. The multi_update::do_updates()
method was modified to restore necessary record buffers
before evaluation of the CHECK OPTION condition.
mysql-test/t/view.test:
Updated test case for bug #28716.
mysql-test/r/view.result:
Updated test case for bug #28716.
Integer values with 10 digits may or may not fit into an int column
(e.g. 2147483647 vs 6147483647).
Thus when creating a temp table column for such an int we must
use bigint instead.
Fixed to use bigint.
Also subsituted a "magic number" with a named constant.
mysql-test/r/analyse.result:
Bug #28492: Adjusted the results after having fixed the bug
mysql-test/r/metadata.result:
Bug #28492: test case
mysql-test/r/olap.result:
Bug #28492: Adjusted the results after having fixed the bug
mysql-test/r/sp.result:
Bug #28492: Adjusted the results after having fixed the bug
mysql-test/r/view.result:
Bug #28492: Adjusted the results after having fixed the bug
mysql-test/t/metadata.test:
Bug #28492: test case
sql/field.h:
Bug #28492: Replaced a magic number with a constant
sql/sql_select.cc:
Bug #28492: Treat integers with 10 and more digits as
bigint.
type assertion.
The bug was introduced by the patch for bug #16377.
The "+ INTERVAL" (Item_date_add_interval) function detects its result type
by the type of its first argument. But in some cases it returns STRING
as the result type. This happens when, for example, the first argument is a
DATE represented as string. All this makes the get_datetime_value()
function misinterpret such result and return wrong DATE/DATETIME value.
To avoid such cases in the fix for #16377 the code that detects correct result
field type on the first execution was added to the
Item_date_add_interval::get_date() function. Due to this the result
field type of the Item_date_add_interval item stored by the send_fields()
function differs from item's result field type at the moment when
the item is actually sent. It causes an assertion failure.
Now the get_datetime_value() detects that the DATE value is returned by
some item not only by checking the result field type but also by comparing
the returned value with the 100000000L constant - any DATE value should be
less than this value.
Removed result field type adjusting code from the
Item_date_add_interval::get_date() function.
sql/item_cmpfunc.cc:
Bug#28450: The Item_date_add_interval in select list may fail the field
type assertion.
Now the get_datetime_value() detects that the DATE value is returned by
some item not only by checking the result field type but also by comparing
the returned value with the 100000000L constant - any DATE value should be
less than this value.
mysql-test/r/func_date_add.result:
Added a test case for the bug#28450: The Item_date_add_interval in select list may fail the field
type assertion.
sql/item_timefunc.cc:
Bug#28450: The Item_date_add_interval in select list may fail the field
type assertion.
Removed result field type adjusting code from the
Item_date_add_interval::get_date() function.
mysql-test/t/func_date_add.test:
Added a test case for the bug#28450: The Item_date_add_interval in select list may fail the field
type assertion.
Refining the tests since pb revealed the older version's fragality - the error from SF() due to killed
may be different on different env:s.
DBUG_ASSERT instead of assert.
mysql-test/r/binlog_killed.result:
new result file
mysql-test/t/binlog_killed.test:
regression for bug#22725 simplified. tests for bug27563, BUG#27565 made inactive.
sql/sql_insert.cc:
DBUG_ASSERT
