rows event):
Adding a after image COLS bitmap to Update_rows_log_event (for telling
what columns that are present in the after image of each row update).
Also fixing case where Rows_log_event length was not correctly computed
(happened when the number of columns in a table was more than 251).
- Improve error handling for "out of memory" problems when master is
sending logs to slave. If memory allocation fails the log should
now report error "memory allocation failed reading log event"
sql/log_event.cc:
Return LOG_READ_MEM from "Log_event::read_log_event" if memory
allocation fails.
- Improve error handling for "out of memory" problems when master is
sending logs to slave. If memory allocation fails the log should
now report error "memory allocation failed reading log event"
when it was printing a Query event, it produced invalid SQL (missing
the BINLOG keyword, so the SQL started with the base64 string, which
is incorrect).
Note: no testcase; I have a .test which shows that the bugfix works,
but it triggers BUG#26361 and so gives Valgrind warnings. I'm sending
this test to the fixer of BUG#26361 for her/him to push when she/he
fixes BUG#26361.
client/mysqlbinlog.cc:
writing the header (a line started with "#", i.e. a comment) and the
body (the real operation) of an event to the same IO_CACHE
(result_cache) confused the logic of Log_event::print_base64()
(which is that if the cache is not empty then the BINLOG keyword
should not be printed); it caused the BINLOG keyword to miss hence
a syntactically wrong output of "mysqlbinlog --base64-output"
for Query events.
So we just use the two IO_CACHE already available in "print_event_info".
sql/log_event.cc:
using the new small inline function.
Note that the replication code should one day be fixed to trap all
errors (like disk write errors).
sql/log_event.h:
small inline function to group two operations: copying an IO_CACHE
to a FILE, and reinitializing this IO_CACHE for being filled again.
sql/records.cc:
fix after merge
when it was printing a Query event, it produced invalid SQL (missing
the BINLOG keyword, so the SQL started with the base64 string, which
is incorrect).
Note: no testcase; I have a .test which shows that the bugfix works,
but it triggers BUG#26361 and so gives Valgrind warnings. I'm sending
this test to the fixer of BUG#26361 for her/him to push when she/he
fixes BUG#26361.
into romeo.(none):/home/bk/b23171-mysql-5.1-new-rpl
mysql-test/r/rpl_row_tabledefs_2myisam.result:
Auto merged
mysql-test/r/rpl_row_tabledefs_3innodb.result:
Auto merged
sql/rpl_rli.cc:
Auto merged
sql/rpl_rli.h:
Auto merged
sql/rpl_utility.cc:
Auto merged
sql/rpl_utility.h:
Auto merged
sql/slave.h:
Auto merged
sql/log.cc:
Manual merge
sql/log_event.cc:
Manual merge
sql/log_event.h:
Manual merge
sql/slave.cc:
Manual merge
sql/sql_binlog.cc:
Manual merge
Tail fixes after re-applying patches to older version of clone.
sql/log_event.cc:
Name change of shall_skip() -> do_shall_skip()
Introducing public shall_skip().
Name change of skip reason enumeration constants.
Removing extreneous argument to slave_print_msg() causing compiler warning.
sql/log_event.h:
Adding enumeration for skip reason.
Factoring out exec_event() into exec_relay_log_event().
Making public interface to event execution primitives.
Adding documentation.
Making some (internal) functions const-correct.
sql/rpl_rli.cc:
replicate_same_server_id is now a member variable of RLI.
sql/rpl_rli.h:
replicate_same_server_id is now a member variable of RLI.
sql/slave.cc:
Using RLI-specific member variable replicate_same_server_id instead of
global instance.
Moving comments about skipping logic to exec_relay_log_event().
Moving event execution logic to exec_relay_log_event().
sql/sql_binlog.cc:
Using apply_event() directly and adding comment with explenation.
into romeo.(none):/home/bk/b22583-mysql-5.1-new-rpl
mysql-test/extra/rpl_tests/rpl_log.test:
Auto merged
mysql-test/r/rpl_ndb_log.result:
Auto merged
mysql-test/r/rpl_row_log.result:
Auto merged
mysql-test/r/rpl_row_log_innodb.result:
Auto merged
mysql-test/r/rpl_stm_log.result:
Auto merged
mysql-test/r/rpl_switch_stm_row_mixed.result:
Auto merged
mysql-test/t/mysqlbinlog.test:
Auto merged
mysql-test/t/rpl_switch_stm_row_mixed.test:
Auto merged
sql/log_event.cc:
Auto merged
sql/sql_class.cc:
Auto merged
mysql-test/r/rpl_row_basic_11bugs.result:
Manual merge.
mysql-test/t/rpl_row_basic_11bugs.test:
Manual merge.
field does not work
Fix to prevent MyISAM from reading data from NULL BLOB.
Fix to make record comparison independent of values of unused bits in
record.
Updating binlog positions in tests.
mysql-test/extra/rpl_tests/rpl_multi_query.test:
Binlog position change
mysql-test/extra/rpl_tests/rpl_stm_charset.test:
Binlog position change
mysql-test/include/show_binlog_events.inc:
Binlog position change
mysql-test/r/binlog_stm_binlog.result:
Result change
mysql-test/r/binlog_stm_ctype_ucs.result:
Result change
mysql-test/r/binlog_stm_insert_select.result:
Result change
mysql-test/r/binlog_stm_mix_innodb_myisam.result:
Result change
mysql-test/r/ctype_cp932_binlog_stm.result:
Result change
mysql-test/r/ndb_binlog_multi.result:
Result change
mysql-test/r/rpl_known_bugs_detection.result:
Result change
mysql-test/r/rpl_loaddata.result:
Result change
mysql-test/r/rpl_loaddata_s.result:
Result change
mysql-test/r/rpl_ndb_charset.result:
Result change
mysql-test/r/rpl_ndb_extraCol.result:
Result change
mysql-test/r/rpl_ndb_log.result:
Result change
mysql-test/r/rpl_ndb_multi.result:
Result change
mysql-test/r/rpl_rbr_to_sbr.result:
Result change
mysql-test/r/rpl_rotate_logs.result:
Result change
mysql-test/r/rpl_sp.result:
Result change
mysql-test/r/rpl_stm_charset.result:
Result change
mysql-test/r/rpl_stm_flsh_tbls.result:
Result change
mysql-test/r/rpl_stm_log.result:
Result change
mysql-test/r/rpl_stm_max_relay_size.result:
Result change
mysql-test/r/rpl_stm_multi_query.result:
Result change
mysql-test/r/rpl_stm_reset_slave.result:
Result change
mysql-test/r/rpl_stm_until.result:
Result change
mysql-test/r/rpl_truncate_7ndb.result:
Result change
mysql-test/r/user_var-binlog.result:
Result change
mysql-test/t/binlog_stm_mix_innodb_myisam.test:
Binlog position change
mysql-test/t/ctype_cp932_binlog_stm.test:
Binlog position change
mysql-test/t/mysqlbinlog.test:
Binlog position change
mysql-test/t/mysqlbinlog2.test:
Binlog position change
mysql-test/t/rpl_sp.test:
Binlog position change
mysql-test/t/rpl_stm_flsh_tbls.test:
Binlog position change
sql/log_event.cc:
Emptying the record entirely since it appears MyISAM reads blob column
data even when they are NULL.
Adding code to set unused bits of the records before doing a comparison,
and restoring the original values after. Setting the unused bits is necessary
since NDB does not set them correctly, and resetting them afterwards is needed
because MyISAM compares the record with the one located when updating
or deleting it.
mysql-test/r/rpl_row_basic_11bugs-master.opt:
New BitKeeper file ``mysql-test/r/rpl_row_basic_11bugs-master.opt''
mysql-test/r/rpl_row_basic_11bugs-slave.opt:
New BitKeeper file ``mysql-test/r/rpl_row_basic_11bugs-slave.opt''
field does not work
Fix to prevent MyISAM from reading data from NULL BLOB.
Fix to make record comparison independent of values of unused bits in
record.
Updating binlog positions in tests.
into bodhi.local:/opt/local/work/mysql-5.1-runtime
include/my_global.h:
Auto merged
mysql-test/mysql-test-run.pl:
Auto merged
mysql-test/r/events_scheduling.result:
Auto merged
BitKeeper/deleted/.del-init_db.sql~a77d572c39d5a1f8:
Auto merged
BitKeeper/deleted/.del-mysql_create_system_tables.sh:
Auto merged
mysql-test/r/query_cache.result:
Auto merged
mysql-test/r/view.result:
Auto merged
mysql-test/t/events_scheduling.test:
Auto merged
mysql-test/t/grant_cache.test:
Auto merged
mysql-test/t/query_cache.test:
Auto merged
mysql-test/t/view.test:
Auto merged
scripts/mysql_system_tables_fix.sql:
Auto merged
sql/event_db_repository.cc:
Auto merged
sql/event_queue.cc:
Auto merged
sql/field.cc:
Auto merged
sql/handler.cc:
Auto merged
sql/handler.h:
Auto merged
sql/item.cc:
Auto merged
sql/item.h:
Auto merged
sql/item_func.cc:
Auto merged
sql/item_subselect.cc:
Auto merged
sql/log_event.cc:
Auto merged
sql/mysql_priv.h:
Auto merged
sql/mysqld.cc:
Auto merged
sql/set_var.cc:
Auto merged
sql/sp.cc:
Auto merged
sql/sp_head.cc:
Auto merged
sql/sp_head.h:
Auto merged
sql/sql_base.cc:
Auto merged
sql/sql_class.cc:
Auto merged
sql/sql_class.h:
Auto merged
sql/sql_help.cc:
Auto merged
sql/sql_lex.cc:
Auto merged
sql/sql_lex.h:
Auto merged
sql/sql_parse.cc:
Auto merged
sql/sql_prepare.cc:
Auto merged
sql/sql_select.cc:
Auto merged
sql/sql_show.cc:
Auto merged
sql/sql_table.cc:
Auto merged
sql/sql_yacc.yy:
Auto merged
sql/table.cc:
Auto merged
sql/share/errmsg.txt:
Auto merged
sql/tztime.cc:
Auto merged
storage/innobase/handler/ha_innodb.cc:
Auto merged
storage/myisam/ha_myisam.cc:
Auto merged
storage/myisam/ha_myisam.h:
Auto merged
mysql-test/r/skip_grants.result:
Manualmerge.
mysql-test/r/sp.result:
Manualmerge.
mysql-test/t/skip_grants.test:
Manualmerge.
mysql-test/t/sp.test:
Manualmerge.
sql/event_data_objects.cc:
Manualmerge.
into janus.mylan:/usr/home/serg/Abk/mysql-5.1
include/my_global.h:
Auto merged
sql/ha_ndbcluster.cc:
Auto merged
sql/ha_partition.cc:
Auto merged
sql/handler.cc:
Auto merged
sql/item_subselect.cc:
Auto merged
sql/log.cc:
Auto merged
sql/log_event.cc:
Auto merged
sql/mysql_priv.h:
Auto merged
sql/opt_range.cc:
Auto merged
sql/slave.cc:
Auto merged
sql/slave.h:
Auto merged
sql/sp.cc:
Auto merged
sql/sql_acl.cc:
Auto merged
sql/sql_insert.cc:
Auto merged
sql/sql_select.cc:
Auto merged
sql/sql_select.h:
Auto merged
sql/sql_servers.cc:
Auto merged
sql/table.cc:
Auto merged
sql/table.h:
Auto merged
sql/tztime.cc:
Auto merged
storage/heap/ha_heap.cc:
Auto merged
storage/innobase/handler/ha_innodb.cc:
Auto merged
storage/myisam/ha_myisam.cc:
Auto merged
storage/myisammrg/ha_myisammrg.h:
Auto merged
into weblab.(none):/home/marcsql/TREE/mysql-5.1-26503-merge
mysql-test/t/sp-error.test:
Auto merged
sql/handler.cc:
Auto merged
sql/item.cc:
Auto merged
sql/item.h:
Auto merged
sql/item_func.cc:
Auto merged
sql/item_subselect.cc:
Auto merged
sql/log_event.cc:
Auto merged
sql/sp_head.cc:
Auto merged
sql/sp_head.h:
Auto merged
sql/sql_base.cc:
Auto merged
sql/sql_show.cc:
Auto merged
sql/sql_table.cc:
Auto merged
sql/sql_yacc.yy:
Auto merged
into moonlight.home:/home/tomash/src/mysql_ab/mysql-5.1-bug9953
mysql-test/r/sp-error.result:
Auto merged
mysql-test/r/view.result:
Auto merged
sql/item_timefunc.cc:
Auto merged
sql/lock.cc:
Auto merged
sql/log_event.cc:
Auto merged
sql/mysql_priv.h:
Auto merged
sql/set_var.cc:
Auto merged
sql/sp.cc:
Auto merged
sql/sql_base.cc:
Auto merged
sql/sql_help.cc:
Auto merged
sql/sql_lex.cc:
Auto merged
sql/sql_lex.h:
Auto merged
sql/sql_parse.cc:
Auto merged
sql/sql_show.cc:
Auto merged
sql/sql_view.cc:
Auto merged
sql/table.cc:
Auto merged
sql/tztime.cc:
Auto merged
storage/myisam/ha_myisam.cc:
Auto merged
sql/sql_yacc.yy:
SCCS merged
The problem was that some facilities (like CONVERT_TZ() function or
server HELP statement) may require implicit access to some tables in
'mysql' database. This access was done by ordinary means of adding
such tables to the list of tables the query is going to open.
However, if we issued LOCK TABLES before that, we would get "table
was not locked" error trying to open such implicit tables.
The solution is to treat certain tables as MySQL system tables, like
we already do for mysql.proc. Such tables may be opened for reading
at any moment regardless of any locks in effect. The cost of this is
that system table may be locked for writing only together with other
system tables, it is disallowed to lock system tables for writing and
have any other lock on any other table.
After this patch the following tables are treated as MySQL system
tables:
mysql.help_category
mysql.help_keyword
mysql.help_relation
mysql.help_topic
mysql.proc (it already was)
mysql.time_zone
mysql.time_zone_leap_second
mysql.time_zone_name
mysql.time_zone_transition
mysql.time_zone_transition_type
These tables are now opened with open_system_tables_for_read() and
closed with close_system_tables(), or one table may be opened with
open_system_table_for_update() and closed with close_thread_tables()
(the latter is used for mysql.proc table, which is updated as part of
normal MySQL server operation). These functions may be used when
some tables were opened and locked already.
NOTE: online update of time zone tables is not possible during
replication, because there's no time zone cache flush neither on LOCK
TABLES, nor on FLUSH TABLES, so the master may serve stale time zone
data from cache, while on slave updated data will be loaded from the
time zone tables.
mysql-test/r/help.result:
Update result.
mysql-test/r/lock.result:
Update result.
mysql-test/r/sp-error.result:
Update result.
mysql-test/r/timezone2.result:
Add result for bug#9953: CONVERT_TZ requires mysql.time_zone_name
to be locked.
mysql-test/r/view.result:
Update result: use table t3 rather than utilize MySQL system table.
mysql-test/t/help.test:
Test that we can use HELP even under LOCK TABLES.
mysql-test/t/lock.test:
Test LOCK TABLE on system tables.
mysql-test/t/timezone2.test:
Add test case for bug#9953: CONVERT_TZ requires mysql.time_zone_name
to be locked.
mysql-test/t/view.test:
Update test: use table t3 rather that utilize MySQL system table.
sql/handler.h:
Fix comment for 'count' parameter of check_if_locking_is_allowed().
Add 'current' and 'system_count' parameters.
sql/item_create.cc:
We no longer have LEX::add_time_zone_tables_to_query_tables().
sql/item_timefunc.cc:
We no longer have LEX::time_zone_tables_used, so
Item_func_convert_tz::fix_fields() became the same as base
Item_date_func::fix_fields().
my_tz_find() no longer takes table list, but takes THD pointer now.
sql/item_timefunc.h:
Remove dead field and method.
sql/lock.cc:
Pass values for 'current' and 'system_count' to
check_if_locking_is_allowed().
sql/log_event.cc:
We no longer have my_tz_find_with_opening_tz_tables(), its functions is
performed by my_tz_find().
sql/mysql_priv.h:
Add functions to work with MySQL system tables.
sql/set_var.cc:
my_tz_find() no longer takes table list, but takes THD pointer now.
sql/sp.cc:
Remove close_proc_table(). Use close_system_tables() instead.
Use open_system_tables_for_read() and open_system_table_for_update().
sql/sp.h:
Remove close_proc_table() declaration.
sql/sql_base.cc:
Add implementation of open_system_tables_for_read(),
close_system_tables(), open_system_table_for_update().
sql/sql_help.cc:
Operate on MySQL system tables mysql.help_* with
open_system_tables_for_read() and close_system_tables() to allow the
usage of HELP statement under LOCK TABLES.
sql/sql_lex.cc:
Remove LEX::time_zone_tables_used and
LEX::add_time_zone_tables_to_query_tables() which are no longer used.
sql/sql_lex.h:
Remove LEX::time_zone_tables_used and
LEX::add_time_zone_tables_to_query_tables() which are no longer used.
sql/sql_parse.cc:
Remove references to LEX::time_zone_tables_used and
my_tz_check_n_skip_implicit_tables() which are no longer used.
sql/sql_show.cc:
Use close_system_tables() instead of removed close_proc_table().
sql/sql_view.cc:
LEX::time_zone_tables_used is no longer there.
sql/sql_yacc.yy:
LEX::add_time_zone_tables_to_query_tables() is no longer there.
sql/table.cc:
Add more tables that should be treated as MySQL system tables.
sql/share/errmsg.txt:
Change the error message, as now we allow write-locking of several
system tables if not mixed with ordinary tables.
sql/tztime.cc:
Do not add time zone tables to the list of query tables in
tz_init_table_list().
Remove fake_time_zone_tables_list and my_tz_get_tables_list().
In my_tz_init(), open mysql.time_zone_leap_second with
simple_open_n_lock_tables(), but pass time zone name to my_tz_find(),
which will open and close time zone tables as necessary.
In tz_load_from_open_tables() do not call table->use_all_columns(), as
this was already done in open_system_tables_for_read().
my_tz_find() takes THD pointer instead of table list, and calls
open_system_tables_for_read() and close_system_tables() as necessary.
Remove my_tz_find_with_opening_tz_tables().
sql/tztime.h:
Remove declarations of my_tz_get_table_list(),
my_tz_find_with_opening_tz_tables(), fake_time_zone_tables_list,
definition of my_tz_check_n_skip_implicit_tables().
Update prototype for my_tz_find().
storage/csv/ha_tina.cc:
Add new parameters to check_if_locking_is_allowed().
storage/csv/ha_tina.h:
Add new parameters to check_if_locking_is_allowed().
storage/myisam/ha_myisam.cc:
Add new parameters to check_if_locking_is_allowed(). In this
function we count system tables. If there are system tables, but
there are also non-system tables, we report an error.
storage/myisam/ha_myisam.h:
Add new parameters to check_if_locking_is_allowed().
The problem was that some facilities (like CONVERT_TZ() function or
server HELP statement) may require implicit access to some tables in
'mysql' database. This access was done by ordinary means of adding
such tables to the list of tables the query is going to open.
However, if we issued LOCK TABLES before that, we would get "table
was not locked" error trying to open such implicit tables.
The solution is to treat certain tables as MySQL system tables, like
we already do for mysql.proc. Such tables may be opened for reading
at any moment regardless of any locks in effect. The cost of this is
that system table may be locked for writing only together with other
system tables, it is disallowed to lock system tables for writing and
have any other lock on any other table.
After this patch the following tables are treated as MySQL system
tables:
mysql.help_category
mysql.help_keyword
mysql.help_relation
mysql.help_topic
mysql.proc (it already was)
mysql.time_zone
mysql.time_zone_leap_second
mysql.time_zone_name
mysql.time_zone_transition
mysql.time_zone_transition_type
These tables are now opened with open_system_tables_for_read() and
closed with close_system_tables(), or one table may be opened with
open_system_table_for_update() and closed with close_thread_tables()
(the latter is used for mysql.proc table, which is updated as part of
normal MySQL server operation). These functions may be used when
some tables were opened and locked already.
NOTE: online update of time zone tables is not possible during
replication, because there's no time zone cache flush neither on LOCK
TABLES, nor on FLUSH TABLES, so the master may serve stale time zone
data from cache, while on slave updated data will be loaded from the
time zone tables.
include/my_global.h:
Introduce constants to be used instead of magic numbers.
sql/field.cc:
Polishing: use contants instead of magic numbers.
sql/ha_innodb.cc:
Polishing: use contants instead of magic numbers.
sql/handler.cc:
Polishing: use contants instead of magic numbers.
sql/item.cc:
Polishing: use contants instead of magic numbers.
sql/item.h:
Polishing: use contants instead of magic numbers.
sql/item_func.cc:
Polishing: use contants instead of magic numbers.
sql/item_subselect.cc:
Polishing: use contants instead of magic numbers.
sql/log_event.cc:
Polishing: use contants instead of magic numbers.
sql/sql_base.cc:
Polishing: use contants instead of magic numbers.
sql/sql_select.cc:
Polishing: use contants instead of magic numbers.
sql/sql_show.cc:
Polishing: use contants instead of magic numbers.
sql/sql_table.cc:
Polishing: use contants instead of magic numbers.
into romeo.(none):/home/bk/b22583-mysql-5.1-new-rpl
mysql-test/extra/binlog_tests/binlog.test:
Auto merged
mysql-test/extra/binlog_tests/ctype_cp932_binlog.test:
Auto merged
mysql-test/extra/binlog_tests/mix_innodb_myisam_binlog.test:
Auto merged
mysql-test/t/rpl_row_create_table.test:
Auto merged
sql/log_event.h:
Auto merged
mysql-test/r/binlog_row_binlog.result:
Manual merge
sql/log_event.cc:
Manual merge
sql/sql_class.cc:
Manual merge
Post-merge fixes.
mysql-test/extra/binlog_tests/binlog.test:
Binlog position change
mysql-test/extra/binlog_tests/binlog_insert_delayed.test:
Binlog position change
mysql-test/extra/binlog_tests/ctype_cp932_binlog.test:
Binlog position change
mysql-test/extra/binlog_tests/ctype_ucs_binlog.test:
Binlog position change
mysql-test/extra/binlog_tests/mix_innodb_myisam_binlog.test:
Binlog position change
mysql-test/r/binlog_row_binlog.result:
Result change
mysql-test/r/binlog_row_ctype_ucs.result:
Result change
mysql-test/r/binlog_row_insert_select.result:
Result change
mysql-test/r/binlog_row_mix_innodb_myisam.result:
Result change
mysql-test/r/ctype_cp932_binlog_row.result:
Result change
mysql-test/r/flush_block_commit_notembedded.result:
Result change
mysql-test/r/rpl_row_create_table.result:
Result change
mysql-test/r/rpl_row_delayed_ins.result:
Result change
mysql-test/t/binlog_row_mix_innodb_myisam.test:
Binlog position change
mysql-test/t/rpl_row_create_table.test:
Removing blinding of end_log_pos in SHOW BINLOG EVENTS output.
Binlog position change.
sql/log_event.cc:
Post-merge fixes.
into bodhi.local:/opt/local/work/mysql-5.1-runtime
mysql-test/r/trigger.result:
Auto merged
mysql-test/r/view.result:
Auto merged
mysql-test/t/trigger.test:
Auto merged
sql/log_event.cc:
Auto merged
sql/mysqld.cc:
Auto merged
sql/sp_head.cc:
Auto merged
sql/sp_head.h:
Auto merged
sql/sql_base.cc:
Auto merged
sql/sql_lex.cc:
Auto merged
sql/sql_lex.h:
Auto merged
sql/sql_trigger.cc:
Auto merged
sql/sql_update.cc:
Auto merged
sql/table.cc:
Auto merged
sql/table.h:
Auto merged
sql/sql_class.cc:
Manual merge.
sql/sql_class.h:
Manual merge.
sql/sql_parse.cc:
Manual merge.
sql/sql_prepare.cc:
Manual merge.
sql/sql_yacc.yy:
Manual merge.
fixes).
The legend: on a replication slave, in case a trigger creation
was filtered out because of application of replicate-do-table/
replicate-ignore-table rule, the parsed definition of a trigger was not
cleaned up properly. LEX::sphead member was left around and leaked
memory. Until the actual implementation of support of
replicate-ignore-table rules for triggers by the patch for Bug 24478 it
was never the case that "case SQLCOM_CREATE_TRIGGER"
was not executed once a trigger was parsed,
so the deletion of lex->sphead there worked and the memory did not leak.
The fix:
The real cause of the bug is that there is no 1 or 2 places where
we can clean up the main LEX after parse. And the reason we
can not have just one or two places where we clean up the LEX is
asymmetric behaviour of MYSQLparse in case of success or error.
One of the root causes of this behaviour is the code in Item::Item()
constructor. There, a newly created item adds itself to THD::free_list
- a single-linked list of Items used in a statement. Yuck. This code
is unaware that we may have more than one statement active at a time,
and always assumes that the free_list of the current statement is
located in THD::free_list. One day we need to be able to explicitly
allocate an item in a given Query_arena.
Thus, when parsing a definition of a stored procedure, like
CREATE PROCEDURE p1() BEGIN SELECT a FROM t1; SELECT b FROM t1; END;
we actually need to reset THD::mem_root, THD::free_list and THD::lex
to parse the nested procedure statement (SELECT *).
The actual reset and restore is implemented in semantic actions
attached to sp_proc_stmt grammar rule.
The problem is that in case of a parsing error inside a nested statement
Bison generated parser would abort immediately, without executing the
restore part of the semantic action. This would leave THD in an
in-the-middle-of-parsing state.
This is why we couldn't have had a single place where we clean up the LEX
after MYSQLparse - in case of an error we needed to do a clean up
immediately, in case of success a clean up could have been delayed.
This left the door open for a memory leak.
One of the following possibilities were considered when working on a fix:
- patch the replication logic to do the clean up. Rejected
as breaks module borders, replication code should not need to know the
gory details of clean up procedure after CREATE TRIGGER.
- wrap MYSQLparse with a function that would do a clean up.
Rejected as ideally we should fix the problem when it happens, not
adjust for it outside of the problematic code.
- make sure MYSQLparse cleans up after itself by invoking the clean up
functionality in the appropriate places before return. Implemented in
this patch.
- use %destructor rule for sp_proc_stmt to restore THD - cleaner
than the prevoius approach, but rejected
because needs a careful analysis of the side effects, and this patch is
for 5.0, and long term we need to use the next alternative anyway
- make sure that sp_proc_stmt doesn't juggle with THD - this is a
large work that will affect many modules.
Cleanup: move main_lex and main_mem_root from Statement to its
only two descendants Prepared_statement and THD. This ensures that
when a Statement instance was created for purposes of statement backup,
we do not involve LEX constructor/destructor, which is fairly expensive.
In order to track that the transformation produces equivalent
functionality please check the respective constructors and destructors
of Statement, Prepared_statement and THD - these members were
used only there.
This cleanup is unrelated to the patch.
sql/log_event.cc:
THD::main_lex is private and should not be used.
sql/mysqld.cc:
Move MYSQLerror to sql_yacc.yy as it depends on LEX headers now.
sql/sql_class.cc:
Cleanup: move main_lex and main_mem_root to THD and Prepared_statement
sql/sql_class.h:
Cleanup: move main_lex and main_mem_root to THD and Prepared_statement
sql/sql_lex.cc:
Implement st_lex::restore_lex()
sql/sql_lex.h:
Declare st_lex::restore_lex().
sql/sql_parse.cc:
Consolidate the calls to unit.cleanup() and deletion of lex->sphead
in mysql_parse (COM_QUERY handler)
sql/sql_prepare.cc:
No need to delete lex->sphead to restore memory roots now in case of a
parse error - this is done automatically inside MYSQLparse
sql/sql_trigger.cc:
This code could lead to double deletion apparently, as in case
of an error lex.sphead was never reset.
sql/sql_yacc.yy:
Trap all returns from the parser to ensure that MySQL-specific cleanup
is invoked: we need to restore the global state of THD and LEX in
case of a parsing error. In case of a parsing success this happens as
part of normal grammar reduction process.
fixes).
The legend: on a replication slave, in case a trigger creation
was filtered out because of application of replicate-do-table/
replicate-ignore-table rule, the parsed definition of a trigger was not
cleaned up properly. LEX::sphead member was left around and leaked
memory. Until the actual implementation of support of
replicate-ignore-table rules for triggers by the patch for Bug 24478 it
was never the case that "case SQLCOM_CREATE_TRIGGER"
was not executed once a trigger was parsed,
so the deletion of lex->sphead there worked and the memory did not leak.
The fix:
The real cause of the bug is that there is no 1 or 2 places where
we can clean up the main LEX after parse. And the reason we
can not have just one or two places where we clean up the LEX is
asymmetric behaviour of MYSQLparse in case of success or error.
One of the root causes of this behaviour is the code in Item::Item()
constructor. There, a newly created item adds itself to THD::free_list
- a single-linked list of Items used in a statement. Yuck. This code
is unaware that we may have more than one statement active at a time,
and always assumes that the free_list of the current statement is
located in THD::free_list. One day we need to be able to explicitly
allocate an item in a given Query_arena.
Thus, when parsing a definition of a stored procedure, like
CREATE PROCEDURE p1() BEGIN SELECT a FROM t1; SELECT b FROM t1; END;
we actually need to reset THD::mem_root, THD::free_list and THD::lex
to parse the nested procedure statement (SELECT *).
The actual reset and restore is implemented in semantic actions
attached to sp_proc_stmt grammar rule.
The problem is that in case of a parsing error inside a nested statement
Bison generated parser would abort immediately, without executing the
restore part of the semantic action. This would leave THD in an
in-the-middle-of-parsing state.
This is why we couldn't have had a single place where we clean up the LEX
after MYSQLparse - in case of an error we needed to do a clean up
immediately, in case of success a clean up could have been delayed.
This left the door open for a memory leak.
One of the following possibilities were considered when working on a fix:
- patch the replication logic to do the clean up. Rejected
as breaks module borders, replication code should not need to know the
gory details of clean up procedure after CREATE TRIGGER.
- wrap MYSQLparse with a function that would do a clean up.
Rejected as ideally we should fix the problem when it happens, not
adjust for it outside of the problematic code.
- make sure MYSQLparse cleans up after itself by invoking the clean up
functionality in the appropriate places before return. Implemented in
this patch.
- use %destructor rule for sp_proc_stmt to restore THD - cleaner
than the prevoius approach, but rejected
because needs a careful analysis of the side effects, and this patch is
for 5.0, and long term we need to use the next alternative anyway
- make sure that sp_proc_stmt doesn't juggle with THD - this is a
large work that will affect many modules.
Cleanup: move main_lex and main_mem_root from Statement to its
only two descendants Prepared_statement and THD. This ensures that
when a Statement instance was created for purposes of statement backup,
we do not involve LEX constructor/destructor, which is fairly expensive.
In order to track that the transformation produces equivalent
functionality please check the respective constructors and destructors
of Statement, Prepared_statement and THD - these members were
used only there.
This cleanup is unrelated to the patch.
into mysql.com:/home/bar/mysql-5.1-rpl
mysql-test/r/mysqlbinlog.result:
Auto merged
mysql-test/t/mysqlbinlog.test:
Auto merged
sql/log_event.cc:
Auto merged
sql/log_event.h:
Auto merged
sql/sql_class.cc:
Auto merged
sql/sql_class.h:
Auto merged
sql/sql_load.cc:
Auto merged
sql/sql_yacc.yy:
Auto merged
into mysql.com:/home/bar/mysql-5.0.b15126
sql/log_event.cc:
Auto merged
sql/log_event.h:
Auto merged
sql/sql_class.cc:
Auto merged
sql/sql_class.h:
Auto merged
sql/sql_load.cc:
Auto merged
sql/sql_yacc.yy:
Auto merged
mysql-test/r/mysqlbinlog.result:
After merge fix
mysql-test/t/mysqlbinlog.test:
After merge fix
This patch fixes problem that LOAD DATA could use different
character sets when loading files on master and on slave sides:
- Adding replication of thd->variables.collation_database
- Adding optional character set clause into LOAD DATA
Note, the second way, with explicit CHARACTER SET clause
should be the recommended way to load data using an alternative
character set.
The old way, using "SET @@character_set_database=xxx" should be
gradually depricated.
mysql-test/r/mysqlbinlog.result:
Adding test case
mysql-test/t/mysqlbinlog.test:
Adding test case
sql/log_event.cc:
Adding logging of thd->variables.collation_database
sql/log_event.h:
Adding declarations
sql/sql_class.cc:
Exchange character set is null by default
sql/sql_class.h:
Adding character set into sql_exchange
sql/sql_load.cc:
- Using exchange character set (if it was specified in LOAD DATA syntax)
- Using thd->variables.collation_database by default
sql/sql_yacc.yy:
Adding optional character set clause into LOAD DATA syntax
mysql-test/r/rpl_loaddata2.result:
New BitKeeper file ``mysql-test/r/rpl_loaddata2.result''
mysql-test/std_data/loaddata6.dat:
New BitKeeper file ``mysql-test/std_data/loaddata6.dat''
mysql-test/t/rpl_loaddata2.test:
New BitKeeper file ``mysql-test/t/rpl_loaddata2.test''
