sql/item_sum.cc:
a fix (bug #7405: group_concat with distinct and rollup => ignores distinct in some rows).
Code changed in order to work with rollup extension.
mysql-test/r/delete.result:
Test case for bug in multiple-table-delete where some rows was not deleted
mysql-test/t/delete.test:
Test case for bug in multiple-table-delete where some rows was not deleted
sql/item_subselect.cc:
Code cleanup
sql/opt_range.cc:
Code cleanup
sql/sql_delete.cc:
Fixed bug in multiple-table-delete where some rows was not deleted
This happend when the first table-to-delete-from was not the the table that was scanned.
Fixed this by only doing 'delete-on-the-fly' for the first table.
Fixed also some wrong error handling in multi-table-delete
into mysql.com:/opt/local/work/mysql-5.0-7306-new
sql/item_subselect.cc:
Auto merged
sql/mysql_priv.h:
Auto merged
sql/opt_range.cc:
Auto merged
sql/sql_base.cc:
Auto merged
sql/sql_insert.cc:
Auto merged
sql/sql_lex.cc:
Auto merged
sql/sql_lex.h:
Auto merged
sql/sql_olap.cc:
Auto merged
sql/sql_parse.cc:
Auto merged
sql/sql_select.cc:
Auto merged
"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.
mysql-test/r/group_min_max.result:
Test results fixed for EXPLAINs when using GROUP_MIN_MAX access plan.
sql/item_subselect.cc:
Move setting of the internal LIMIT used for IN/ALL/ANY/EXISTS
subqueries to one place: Item_exists_subselect::fix_length_and_dec().
This implies that unit->select_limit_cnt is not set until the item is
fixed. This is OK, as now LIMIT values are not used until JOIN::optimize.
sql/mysql_priv.h:
setup_tables no longer needs a special flag for the case when
it's called from JOIN::reinit() (we don't need to call setup_tables
between two executions of a correlated subquery).
sql/opt_range.cc:
Fix a glitch in GROUP_MIN_MAX access plan: we should use table metadata,
not field data, to evaluate max_used_key_length, which is then
used for explain.
sql/sp.cc:
- setup_tables signature changed.
sql/sql_base.cc:
- setup_tables no longer needs a special mode for subqueries.
Unused checks were removed.
sql/sql_delete.cc:
- setup_tables signature changed
sql/sql_help.cc:
- setup_tables signature changed
sql/sql_insert.cc:
- setup_tables signature changed
sql/sql_lex.cc:
Consolidate setting of internal LIMIT for IN/ALL/ANY/EXISTS subqeries
in one place, and hence remove it from st_select_lex::test_limit().
sql/sql_lex.h:
Cleanup signature of st_select_lex_unit::init_prepare_fake_select_lex().
sql/sql_load.cc:
- setup_tables signature changed
sql/sql_olap.cc:
- setup_tables signature changed
sql/sql_parse.cc:
- st_select_lex_unit::set_limit() signature changed
sql/sql_select.cc:
Move setting of JOIN::select_limit from JOIN::prepare
to JOIN::optimize. At prepare, limit is unknown yet.
Remove excessive cleanups from JOIN::reinit which were overwriting
join->join_tab[i]->table->used_keys. This fixes the bug which was triggered
by the change in item_subselect.cc.
sql/sql_union.cc:
Class st_select_lex_unit was changed to avoid calls to
st_select_lex_unit::set_limit from places where it may be unknown.
Now unit->select_limit_cnt is set at ::exec().
st_select_lex_unit::init_prepare_fake_select_lex():
- move out set_limit functionality
- remove a few lines of dead code.
st_select_lex_unit::prepare():
- now we don't call set_limit at the time of prepare, so the value
of unit->select_limit_cnt may be unknown here. Use sl->select_limit
instead.
st_select_lex_unit::exec():
- cleanup
- call set_limit explicitly as it has been moved out of
init_prepare_fake_select_lex.
sql/sql_update.cc:
- setup_tables signature changed
databases" and basic handling of errors which happen in triggers.
(The bug itself was fixed by several previous patches).
Fixed bug in multi-delete which were exposed by these tests.
mysql-test/r/trigger.result:
Added test for bug #5894 "Triggers with altered tables cause corrupt
databases" and basic handling of errors which happen in triggers.
mysql-test/t/trigger.test:
Added test for bug #5894 "Triggers with altered tables cause corrupt
databases" and basic handling of errors which happen in triggers.
sql/sql_delete.cc:
multi_delete::send_eof():
Fixed bug which were exposed by testing of basic handling of errors
in triggers. Error which happened in multi_delete::do_deletes()
was wiped out by thd->clear_error() and thus never sent to client
(and this stalled client connection).
Added test cases for bug #7894.
sql_select.cc:
Fixed bug #7894: GROUP BY queries with ROLLUP returned
wrong results for expressions containing group by columns.
The fix ensured correct results by replacement of all
occurrences of group by fields in non-aggregate expressions
for corresponding ref objects and preventing creation of
fields in temporary tables for expression containing group
by fields.
sql/sql_select.cc:
Fixed bug #7894: GROUP BY queries with ROLLUP returned
wrong results for expressions containing group by columns.
The fix ensured correct results by replacement of all
occurrences of group by fields in non-aggregate expressions
for corresponding ref objects and preventing creation of
fields in temporary tables for expression containing group
by fields.
mysql-test/t/olap.test:
Added test cases for bug #7894.
mysql-test/r/olap.result:
Added test cases for bug #7894.
into mysql.com:/home/pem/work/mysql-5.0
sql/item_cmpfunc.cc:
Auto merged
sql/item_func.cc:
Auto merged
sql/item_subselect.cc:
Auto merged
sql/mysql_priv.h:
Auto merged
sql/mysqld.cc:
Auto merged
sql/sp_head.cc:
Auto merged
sql/sql_parse.cc:
Auto merged
sql/sql_yacc.yy:
Auto merged
server"
Bug appears only on Windows platform. Freeing memory in
TMP_TABLE_PARAM::cleanup() allocated by new Copy_fields[0] in
setup_copy_fields() results in memory destruction. In test IF used instead
of CONCAT because IF have more stable crash.
sql/sql_select.cc:
Fix bug #9593 The combination of COUNT, DISTINCT and CONCAT seems to lock the server
mysql-test/t/count_distinct.test:
test for bug #9593 The combination of COUNT, DISTINCT and CONCAT seems to lock the server
mysql-test/r/count_distinct.result:
Test for bug #9593 The combination of COUNT, DISTINCT and CONCAT seems to lock the server
to handle large time values as documented. (Bug #10590)
mysql-test/r/func_time.result:
Add new results
mysql-test/t/func_time.test:
Add new regression test
sql/item_timefunc.cc:
Fix handling of '%h', '%I', and '%l' format specifiers
for TIME_FORMAT() to handle large time values correctly.
mysql-test/r/federated.result:
added explain results to join test results per Timour's request.
mysql-test/t/federated.test:
Added explain to test per Timour's request
sql/ha_federated.cc:
better comment to explain why we need table->status set to 0
by simply disabling FLUSH for stored functions. (I can't really work.)
mysql-test/r/sp-error.result:
New test case for BUG#8409.
mysql-test/t/sp-error.test:
New test case for BUG#8409.
sql/sql_yacc.yy:
Disable FLUSH for stored functions.
Check in Jan's fix to bug #10746 and also add a note to ::start_stmt() that stored procs in 5.0 call it
sql/ha_innodb.cc:
Check in Jan's fix to bug #10746 and also add a note to ::start_stmt() that stored procs in 5.0 call it
set net->no_send_error to 0 before execution of each element of
multiquery statement to provide the sending of error to client
tests/mysql_client_test.c:
Fix for bug #9992: mysql_next_result hangs on error
test case
results.
Actually a problem when converting decimal to int for user variables.
mysql-test/r/sp.result:
New test case for BUG#9559.
mysql-test/t/sp.test:
New test case for BUG#9559.
sql/item_func.cc:
Don't set the unsigned flag when converting decimal user var to int.
We can't have Item_trigger_field as aggregated object inside of
sp_instr_set_trigger_field class since in this case its destructor
will be called twice. So instead let us create this Item separately
and store pointer to it in instruction object.
sql/sp_head.cc:
sp_instr_set_trigger_field:
We can't have Item_trigger_field as aggregated object since in this
case its destructor will be called twice, so let us store pointer
to this Item (Another way to avoid this is to exclude this Item
from free_list but this can't be done in elegant way in 5.0 and
will also cause additional problems with Item::cleanup()).
sql/sp_head.h:
sp_instr_set_trigger_field:
We can't have Item_trigger_field as aggregated object since in this
case its destructor will be called twice, so let us store pointer
to this Item (Another way to avoid this is to exclude this Item
from free_list but this can't be done in elegant way in 5.0 and
will also cause additional problems with Item::cleanup()).
sql/sql_yacc.yy:
We can't have Item_trigger_field as aggregated object inside of
sp_instr_set_trigger_field class since in this case its destructor
will be called twice. So instead let us create this Item separately
and store pointer to it in instruction object.
- Remove the expected warrnings when "show binary logs" are called on zero size binary log files.
mysql-test/r/rpl_log.result:
Remove the warnings that is no longer displayed when "show binary logs" is called on zero size log files.
Checked against 4.0, no warnings there.
mysql-test/r/rpl_rotate_logs.result:
Remove the warnings that is no longer displayed when "show binary logs" is called on zero size log files.
Checked against 4.0, no warnings there.
Only add --defaults-extra-file=$DATADIR/my.cnf if $DATADIR/my.cnf is readable.
scripts/mysqld_safe.sh:
Only add --defaults-extra-file=$DATADIR/my.cnf if $DATADIR/my.cnf is readable.
Without this patch, running mysqld_safe gives:
Could not open required defaults file: /opt/home/tim/m/50/m/data/my.cnf
Fatal error in defaults handling. Program aborted
Could not open required defaults file: /opt/home/tim/m/50/m/data/my.cnf
Fatal error in defaults handling. Program aborted
These errors are printed by my_print_defaults; mysqld also encounters the
same problem and quits.
mysql-test/r/union.result:
Convert back the latin1 characters that where accidently changed during last merge
mysql-test/t/union.test:
Convert back the latin1 characters that where accidently changed during last merge
BitKeeper/deleted/.del-ha_isam.cc~4dce65904db2675e:
Auto merged
BitKeeper/deleted/.del-ha_isammrg.cc~dc682e4755d77a2e:
Auto merged
client/sql_string.cc:
Auto merged
client/sql_string.h:
Auto merged
include/my_global.h:
Auto merged
include/my_sys.h:
Auto merged
mysql-test/mysql-test-run.sh:
Auto merged
mysys/my_open.c:
Auto merged
mysys/raid.cc:
Auto merged
ndb/src/kernel/blocks/dbtux/Dbtux.hpp:
Auto merged
sql/field.cc:
Auto merged
sql/ha_berkeley.cc:
Auto merged
sql/ha_blackhole.cc:
Auto merged
sql/ha_heap.cc:
Auto merged
sql/ha_innodb.cc:
Auto merged
sql/ha_myisam.cc:
Auto merged
sql/ha_myisammrg.cc:
Auto merged
sql/ha_ndbcluster.cc:
Auto merged
sql/handler.cc:
Auto merged
sql/item.cc:
Auto merged
sql/item_cmpfunc.cc:
Auto merged
sql/item_func.cc:
Auto merged
sql/item_geofunc.cc:
Auto merged
sql/item_strfunc.cc:
Auto merged
sql/item_subselect.cc:
Auto merged
sql/item_sum.cc:
Auto merged
sql/item_timefunc.cc:
Auto merged
sql/item_uniq.cc:
Auto merged
sql/item_uniq.h:
Auto merged
sql/log_event.cc:
Auto merged
sql/log_event.h:
Auto merged
sql/procedure.cc:
Auto merged
sql/protocol.cc:
Auto merged
sql/protocol_cursor.cc:
Auto merged
sql/set_var.cc:
Auto merged
sql/sql_analyse.cc:
Auto merged
sql/sql_analyse.h:
Auto merged
sql/sql_lex.cc:
Auto merged
sql/sql_map.cc:
Auto merged
sql/sql_olap.cc:
Auto merged
sql/sql_string.cc:
Auto merged
sql/sql_udf.cc:
Auto merged
sql/tztime.cc:
Auto merged
sql/opt_range.cc:
Manual merge
sql/sql_parse.cc:
Use select_lex pointer instead of lex->select_lex
sql/sql_repl.cc:
Function moved to log.cc, fix made there instead
sql/sql_class.cc:
Auto merged
sql/sql_select.cc:
Auto merged
- Add O_SHARE when opening file.
sql/log.cc:
Add O_SHARE when opening file. This is from mysql-4.1 but since function has been moved from sql_repl.cc, it was missed during merge.
into mysql.com:/home/psergey/mysql-5.0-conf-fixes-2
sql/mysql_priv.h:
Auto merged
sql/mysqld.cc:
Auto merged
sql/sql_class.cc:
Auto merged
sql/sql_class.h:
Auto merged
sql/sql_select.cc:
Auto merged
sql/sql_show.cc:
Auto merged
sql/structs.h:
Auto merged
This only showed up on two known platforms, both ia64 (one HP-UX and one Linux wich icc).
For some reason, they only get half the stack size they are supposed to have, which
makes deep SP recursion overrun the stack before check_stack_overrun() is triggered.
Also made som minor fixes in the check_stack_overrun() definition, supporting variable
margins.
No test case added, but the reason for the bug report was a failed existing test case
on these machines, which now works.
sql/item_cmpfunc.cc:
Adopted call to new check_stack_overrun() definition.
sql/item_func.cc:
Adopted calls to new check_stack_overrun() definition.
sql/item_subselect.cc:
Adopted call to new check_stack_overrun() definition.
sql/mysql_priv.h:
Changed definition of check_stack_overrun(), thread_stack_min variable no longer needed.
sql/mysqld.cc:
thread_stack_min variable no longer needed (with variable margin arg. to check_thread_overrun()).
On (at least some) ia64 machines, it seems we only get half the requested stack in reality,
so deep SP recursion crashes before the thread overrun check is triggered.
We work around this by requesting twice the stack requested to get expected size.
sql/sp_head.cc:
Adopted call to new check_stack_overrun() definition, and use a wider margin
execution.
sql/sql_parse.cc:
Modified check_stack_overrun() definition. Removed unused arg that didn't sever any
purpose, and added a 'margin' argument to support different margins for different
callrers.
sql/table.cc:
Adopted call to new check_stack_overrun() definition.
