mirror/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-01-18 04:53:01 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
mariadb/sql/event_db_repository.cc

1044 lines
29 KiB
C++
Raw Normal View History

first cut of WL#3337 (New event scheduler locking infrastructure). Infrastructure built. Added the foreseen files and change Makefile.am/CMakeLists.txt accordingly.
2006-06-27 08:48:50 +02:00
/* Copyright (C) 2004-2006 MySQL AB
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
#include "mysql_priv.h"
#include "event_db_repository.h"
#include "event_data_objects.h"
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
#include "sp_head.h"
#include "sp.h"
#include "events.h"
WL#3337 (Events new architecture) Cut number 6. Move code from sql_show.cc to event_db_repository.cc that more belongs to the latter.
2006-06-28 14:22:14 +02:00
#include "sql_show.h"
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
#define EVEX_DB_FIELD_LEN 64
#define EVEX_NAME_FIELD_LEN 64
time_t mysql_event_last_create_time= 0L;
static
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
TABLE_FIELD_W_TYPE event_table_fields[ET_FIELD_COUNT] =
{
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
{(char *) STRING_WITH_LEN("db")},
{(char *) STRING_WITH_LEN("char(64)")},
{(char *) STRING_WITH_LEN("utf8")}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
},
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
{(char *) STRING_WITH_LEN("name")},
{(char *) STRING_WITH_LEN("char(64)")},
{(char *) STRING_WITH_LEN("utf8")}
},
{
{(char *) STRING_WITH_LEN("body")},
{(char *) STRING_WITH_LEN("longblob")},
{NULL, 0}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
},
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
{(char *) STRING_WITH_LEN("definer")},
{(char *) STRING_WITH_LEN("char(77)")},
{(char *) STRING_WITH_LEN("utf8")}
},
{
{(char *) STRING_WITH_LEN("execute_at")},
{(char *) STRING_WITH_LEN("datetime")},
{NULL, 0}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
},
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
{(char *) STRING_WITH_LEN("interval_value")},
{(char *) STRING_WITH_LEN("int(11)")},
{NULL, 0}
},
{
{(char *) STRING_WITH_LEN("interval_field")},
{(char *) STRING_WITH_LEN("enum('YEAR','QUARTER','MONTH','DAY',"
"'HOUR','MINUTE','WEEK','SECOND','MICROSECOND','YEAR_MONTH','DAY_HOUR',"
"'DAY_MINUTE','DAY_SECOND','HOUR_MINUTE','HOUR_SECOND','MINUTE_SECOND',"
"'DAY_MICROSECOND','HOUR_MICROSECOND','MINUTE_MICROSECOND',"
"'SECOND_MICROSECOND')")},
{NULL, 0}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
},
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
{(char *) STRING_WITH_LEN("created")},
{(char *) STRING_WITH_LEN("timestamp")},
{NULL, 0}
},
{
{(char *) STRING_WITH_LEN("modified")},
{(char *) STRING_WITH_LEN("timestamp")},
{NULL, 0}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
},
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
{(char *) STRING_WITH_LEN("last_executed")},
{(char *) STRING_WITH_LEN("datetime")},
{NULL, 0}
},
{
{(char *) STRING_WITH_LEN("starts")},
{(char *) STRING_WITH_LEN("datetime")},
{NULL, 0}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
},
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
{(char *) STRING_WITH_LEN("ends")},
{(char *) STRING_WITH_LEN("datetime")},
{NULL, 0}
},
{
{(char *) STRING_WITH_LEN("status")},
{(char *) STRING_WITH_LEN("enum('ENABLED','DISABLED')")},
{NULL, 0}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
},
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
{(char *) STRING_WITH_LEN("on_completion")},
{(char *) STRING_WITH_LEN("enum('DROP','PRESERVE')")},
{NULL, 0}
},
{
{(char *) STRING_WITH_LEN("sql_mode")},
{(char *) STRING_WITH_LEN("set('REAL_AS_FLOAT','PIPES_AS_CONCAT','ANSI_QUOTES',"
"'IGNORE_SPACE','NOT_USED','ONLY_FULL_GROUP_BY','NO_UNSIGNED_SUBTRACTION',"
"'NO_DIR_IN_CREATE','POSTGRESQL','ORACLE','MSSQL','DB2','MAXDB',"
"'NO_KEY_OPTIONS','NO_TABLE_OPTIONS','NO_FIELD_OPTIONS','MYSQL323','MYSQL40',"
"'ANSI','NO_AUTO_VALUE_ON_ZERO','NO_BACKSLASH_ESCAPES','STRICT_TRANS_TABLES',"
"'STRICT_ALL_TABLES','NO_ZERO_IN_DATE','NO_ZERO_DATE','INVALID_DATES',"
"'ERROR_FOR_DIVISION_BY_ZERO','TRADITIONAL','NO_AUTO_CREATE_USER',"
"'HIGH_NOT_PRECEDENCE')")},
{NULL, 0}
},
{
{(char *) STRING_WITH_LEN("comment")},
{(char *) STRING_WITH_LEN("char(64)")},
{(char *) STRING_WITH_LEN("utf8")}
}
};
/*
Puts some data common to CREATE and ALTER EVENT into a row.
SYNOPSIS
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
mysql_event_fill_row()
WL #3337 (Event scheduler new architecture) Cut Nr. 8. All tests pass. Separated Event_scheduler into Event_queue and Event_scheduler. Added new Event_scheduler_ng which is the new scheduler and is used system-wide. Will be moved to the event_scheduler.cc in the future. Using Event_timed in Event_queue as well as cloned during execution. Next step is to have Event_worker_data which will be used during execution and will take ::compile()/::execute() out of Event_timed.
2006-07-04 18:44:35 +02:00
thd THD
table The row to fill out
et Event's data
is_update CREATE EVENT or ALTER EVENT
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
RETURN VALUE
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
0 OK
EVEX_GENERAL_ERROR Bad data
EVEX_GET_FIELD_FAILED Field count does not match. table corrupted?
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
DESCRIPTION
Used both when an event is created and when it is altered.
*/
static int
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
mysql_event_fill_row(THD *thd, TABLE *table, Event_parse_data *et,
my_bool is_update)
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
{
CHARSET_INFO *scs= system_charset_info;
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
enum enum_events_table_field f_num;
Field **fields= table->field;
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
DBUG_ENTER("mysql_event_fill_row");
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
DBUG_PRINT("info", ("dbname=[%s]", et->dbname.str));
DBUG_PRINT("info", ("name =[%s]", et->name.str));
DBUG_PRINT("info", ("body =[%s]", et->body.str));
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if (fields[f_num= ET_FIELD_DEFINER]->
store(et->definer.str, et->definer.length, scs))
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
goto err_truncate;
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if (fields[f_num= ET_FIELD_DB]->store(et->dbname.str, et->dbname.length, scs))
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
goto err_truncate;
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if (fields[f_num= ET_FIELD_NAME]->store(et->name.str, et->name.length, scs))
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
goto err_truncate;
/* both ON_COMPLETION and STATUS are NOT NULL thus not calling set_notnull()*/
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
fields[ET_FIELD_ON_COMPLETION]->store((longlong)et->on_completion, TRUE);
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
fields[ET_FIELD_STATUS]->store((longlong)et->status, TRUE);
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
/*
Change the SQL_MODE only if body was present in an ALTER EVENT and of course
always during CREATE EVENT.
*/
if (et->body.str)
{
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
fields[ET_FIELD_SQL_MODE]->store((longlong)thd->variables.sql_mode, TRUE);
if (fields[f_num= ET_FIELD_BODY]->store(et->body.str, et->body.length, scs))
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
goto err_truncate;
}
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
if (et->expression)
{
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
fields[ET_FIELD_INTERVAL_EXPR]->set_notnull();
fields[ET_FIELD_INTERVAL_EXPR]->store((longlong)et->expression, TRUE);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
fields[ET_FIELD_TRANSIENT_INTERVAL]->set_notnull();
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
/*
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
In the enum (C) intervals start from 0 but in mysql enum valid values
start from 1. Thus +1 offset is needed!
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
*/
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
fields[ET_FIELD_TRANSIENT_INTERVAL]->store((longlong)et->interval+1, TRUE);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
fields[ET_FIELD_EXECUTE_AT]->set_null();
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
if (!et->starts_null)
{
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
fields[ET_FIELD_STARTS]->set_notnull();
fields[ET_FIELD_STARTS]->store_time(&et->starts, MYSQL_TIMESTAMP_DATETIME);
WL#3337 (Event scheduler new architecture) Cleaned up the code a bit. Fixed few leaks. This code still does not load events on server startup from disk. The problem is that there is a need for a THD instance, which does not exist during server boot. This will be solved soon. Still Event_timed is used both for the memory queue and for exectution. This will be changed according to WL#3337 probably in the next commit.
2006-07-05 17:12:50 +02:00
}
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
if (!et->ends_null)
{
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
fields[ET_FIELD_ENDS]->set_notnull();
fields[ET_FIELD_ENDS]->store_time(&et->ends, MYSQL_TIMESTAMP_DATETIME);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
}
}
else if (et->execute_at.year)
{
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
fields[ET_FIELD_INTERVAL_EXPR]->set_null();
fields[ET_FIELD_TRANSIENT_INTERVAL]->set_null();
fields[ET_FIELD_STARTS]->set_null();
fields[ET_FIELD_ENDS]->set_null();
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
fields[ET_FIELD_EXECUTE_AT]->set_notnull();
fields[ET_FIELD_EXECUTE_AT]->
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
store_time(&et->execute_at, MYSQL_TIMESTAMP_DATETIME);
}
else
{
DBUG_ASSERT(is_update);
/*
it is normal to be here when the action is update
this is an error if the action is create. something is borked
*/
}
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
((Field_timestamp *)fields[ET_FIELD_MODIFIED])->set_time();
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
if (et->comment.str)
{
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if (fields[f_num= ET_FIELD_COMMENT]->
store(et->comment.str, et->comment.length, scs))
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
goto err_truncate;
}
DBUG_RETURN(0);
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
err_truncate:
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
my_error(ER_EVENT_DATA_TOO_LONG, MYF(0), fields[f_num]->field_name);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
DBUG_RETURN(EVEX_GENERAL_ERROR);
}
WL#3337 (Events new architecture) Cut number 6. Move code from sql_show.cc to event_db_repository.cc that more belongs to the latter.
2006-06-28 14:22:14 +02:00
/*
Performs an index scan of event_table (mysql.event) and fills schema_table.
Synopsis
Event_db_repository::index_read_for_db_for_i_s()
thd Thread
schema_table The I_S.EVENTS table
event_table The event table to use for loading (mysql.event)
Returns
0 OK
1 Error
*/
int
Event_db_repository::index_read_for_db_for_i_s(THD *thd, TABLE *schema_table,
TABLE *event_table, char *db)
{
int ret=0;
CHARSET_INFO *scs= system_charset_info;
KEY *key_info;
uint key_len;
byte *key_buf= NULL;
LINT_INIT(key_buf);
DBUG_ENTER("Event_db_repository::index_read_for_db_for_i_s");
DBUG_PRINT("info", ("Using prefix scanning on PK"));
event_table->file->ha_index_init(0, 1);
event_table->field[ET_FIELD_DB]->store(db, strlen(db), scs);
key_info= event_table->key_info;
key_len= key_info->key_part[0].store_length;
if (!(key_buf= (byte *)alloc_root(thd->mem_root, key_len)))
{
ret= 1;
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
/* Don't send error, it would be done by sql_alloc_error_handler() */
WL#3337 (Events new architecture) Cut number 6. Move code from sql_show.cc to event_db_repository.cc that more belongs to the latter.
2006-06-28 14:22:14 +02:00
}
else
{
key_copy(key_buf, event_table->record[0], key_info, key_len);
if (!(ret= event_table->file->index_read(event_table->record[0], key_buf,
key_len, HA_READ_PREFIX)))
{
DBUG_PRINT("info",("Found rows. Let's retrieve them. ret=%d", ret));
do
{
ret= copy_event_to_schema_table(thd, schema_table, event_table);
if (ret == 0)
ret= event_table->file->index_next_same(event_table->record[0],
key_buf, key_len);
} while (ret == 0);
}
DBUG_PRINT("info", ("Scan finished. ret=%d", ret));
}
event_table->file->ha_index_end();
/* ret is guaranteed to be != 0 */
if (ret == HA_ERR_END_OF_FILE || ret == HA_ERR_KEY_NOT_FOUND)
DBUG_RETURN(0);
DBUG_RETURN(1);
}
/*
Performs a table scan of event_table (mysql.event) and fills schema_table.
Synopsis
Events_db_repository::table_scan_all_for_i_s()
thd Thread
schema_table The I_S.EVENTS in memory table
event_table The event table to use for loading.
Returns
0 OK
1 Error
*/
int
Event_db_repository::table_scan_all_for_i_s(THD *thd, TABLE *schema_table,
TABLE *event_table)
{
int ret;
READ_RECORD read_record_info;
DBUG_ENTER("Event_db_repository::table_scan_all_for_i_s");
init_read_record(&read_record_info, thd, event_table, NULL, 1, 0);
/*
rr_sequential, in read_record(), returns 137==HA_ERR_END_OF_FILE,
but rr_handle_error returns -1 for that reason. Thus, read_record()
returns -1 eventually.
*/
do
{
ret= read_record_info.read_record(&read_record_info);
if (ret == 0)
ret= copy_event_to_schema_table(thd, schema_table, event_table);
WL#3337 (Event scheduler new architecture) Cleaned up the code a bit. Fixed few leaks. This code still does not load events on server startup from disk. The problem is that there is a need for a THD instance, which does not exist during server boot. This will be solved soon. Still Event_timed is used both for the memory queue and for exectution. This will be changed according to WL#3337 probably in the next commit.
2006-07-05 17:12:50 +02:00
} while (ret == 0);
WL#3337 (Events new architecture) Cut number 6. Move code from sql_show.cc to event_db_repository.cc that more belongs to the latter.
2006-06-28 14:22:14 +02:00
DBUG_PRINT("info", ("Scan finished. ret=%d", ret));
end_read_record(&read_record_info);
/* ret is guaranteed to be != 0 */
DBUG_RETURN(ret == -1? 0:1);
}
/*
Fills I_S.EVENTS with data loaded from mysql.event. Also used by
SHOW EVENTS
Synopsis
Event_db_repository::fill_schema_events()
thd Thread
tables The schema table
db If not NULL then get events only from this schema
Returns
0 OK
1 Error
*/
int
Event_db_repository::fill_schema_events(THD *thd, TABLE_LIST *tables, char *db)
{
TABLE *schema_table= tables->table;
TABLE *event_table= NULL;
Open_tables_state backup;
int ret= 0;
DBUG_ENTER("Event_db_repository::fill_schema_events");
DBUG_PRINT("info",("db=%s", db? db:"(null)"));
thd->reset_n_backup_open_tables_state(&backup);
if (open_event_table(thd, TL_READ, &event_table))
{
sql_print_error("Table mysql.event is damaged.");
thd->restore_backup_open_tables_state(&backup);
DBUG_RETURN(1);
}
/*
1. SELECT I_S => use table scan. I_S.EVENTS does not guarantee order
thus we won't order it. OTOH, SHOW EVENTS will be
ordered.
2. SHOW EVENTS => PRIMARY KEY with prefix scanning on (db)
Reasoning: Events are per schema, therefore a scan over an index
will save use from doing a table scan and comparing
every single row's `db` with the schema which we show.
*/
if (db)
ret= index_read_for_db_for_i_s(thd, schema_table, event_table, db);
else
ret= table_scan_all_for_i_s(thd, schema_table, event_table);
close_thread_tables(thd);
thd->restore_backup_open_tables_state(&backup);
DBUG_PRINT("info", ("Return code=%d", ret));
DBUG_RETURN(ret);
}
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
/*
Open mysql.event table for read
SYNOPSIS
Events::open_event_table()
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
thd [in] Thread context
lock_type [in] How to lock the table
table [out] We will store the open table here
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
RETURN VALUE
1 Cannot lock table
2 The table is corrupted - different number of fields
0 OK
*/
int
Event_db_repository::open_event_table(THD *thd, enum thr_lock_type lock_type,
TABLE **table)
{
TABLE_LIST tables;
DBUG_ENTER("Event_db_repository::open_event_table");
bzero((char*) &tables, sizeof(tables));
tables.db= (char*) "mysql";
tables.table_name= tables.alias= (char*) "event";
tables.lock_type= lock_type;
if (simple_open_n_lock_tables(thd, &tables))
DBUG_RETURN(1);
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
if (table_check_intact(tables.table, ET_FIELD_COUNT,
event_table_fields,
&mysql_event_last_create_time,
ER_CANNOT_LOAD_FROM_TABLE))
{
close_thread_tables(thd);
DBUG_RETURN(2);
}
*table= tables.table;
tables.table->use_all_columns();
DBUG_RETURN(0);
}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
/*
Checks parameters which we got from the parsing phase.
SYNOPSIS
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
check_parse_params()
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
thd THD
et event's data
RETURNS
0 OK
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
EVEX_BAD_PARAMS Error (reported)
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
*/
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
static int
check_parse_params(THD *thd, Event_parse_data *parse_data)
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
{
const char *pos= NULL;
Item *bad_item;
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
int res;
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
DBUG_ENTER("check_parse_params");
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if (parse_data->check_parse_data(thd))
DBUG_RETURN(EVEX_BAD_PARAMS);
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
if (!parse_data->dbname.str ||
(thd->lex->sql_command == SQLCOM_ALTER_EVENT && thd->lex->spname &&
!thd->lex->spname->m_db.str))
{
my_message(ER_NO_DB_ERROR, ER(ER_NO_DB_ERROR), MYF(0));
DBUG_RETURN(EVEX_BAD_PARAMS);
}
if (check_access(thd, EVENT_ACL, parse_data->dbname.str, 0, 0, 0,
is_schema_db(parse_data->dbname.str)) ||
(thd->lex->sql_command == SQLCOM_ALTER_EVENT && thd->lex->spname &&
(check_access(thd, EVENT_ACL, thd->lex->spname->m_db.str, 0, 0, 0,
is_schema_db(thd->lex->spname->m_db.str)))))
DBUG_RETURN(EVEX_BAD_PARAMS);
DBUG_RETURN(0);
}
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
/*
Creates an event in mysql.event
SYNOPSIS
Event_db_repository::create_event()
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
thd [in] THD
et [in] Object containing info about the event
create_if_not [in] Whether to generate anwarning in case event exists
rows_affected [out] How many rows were affected
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
RETURN VALUE
0 - OK
EVEX_GENERAL_ERROR - Failure
DESCRIPTION
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
Creates an event. Relies on mysql_event_fill_row which is shared with
::update_event. The name of the event is inside "et".
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
*/
int
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
Event_db_repository::create_event(THD *thd, Event_parse_data *parse_data,
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
my_bool create_if_not, uint *rows_affected)
{
int ret= 0;
CHARSET_INFO *scs= system_charset_info;
TABLE *table;
Remove trailing whitespace. After merge fixes.
2006-07-10 23:54:46 +02:00
char old_db_buf[NAME_LEN+1];
LEX_STRING old_db= { old_db_buf, sizeof(old_db_buf) };
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
bool dbchanged= FALSE;
Remove trailing whitespace. After merge fixes.
2006-07-10 23:54:46 +02:00
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
DBUG_ENTER("Event_db_repository::create_event");
*rows_affected= 0;
DBUG_PRINT("info", ("open mysql.event for update"));
if (open_event_table(thd, TL_WRITE, &table))
{
my_error(ER_EVENT_OPEN_TABLE_FAILED, MYF(0));
goto err;
}
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if (check_parse_params(thd, parse_data))
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
goto err;
DBUG_PRINT("info", ("name: %.*s", parse_data->name.length,
parse_data->name.str));
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
DBUG_PRINT("info", ("check existance of an event with the same name"));
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if (!find_event_by_name(thd, parse_data->dbname, parse_data->name, table))
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
if (create_if_not)
{
push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_NOTE,
ER_EVENT_ALREADY_EXISTS, ER(ER_EVENT_ALREADY_EXISTS),
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
parse_data->name.str);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
goto ok;
}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
my_error(ER_EVENT_ALREADY_EXISTS, MYF(0), parse_data->name.str);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
goto err;
}
DBUG_PRINT("info", ("non-existant, go forward"));
Remove trailing whitespace. After merge fixes.
2006-07-10 23:54:46 +02:00
if ((ret= sp_use_new_db(thd, parse_data->dbname, &old_db, 0, &dbchanged)))
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
my_error(ER_BAD_DB_ERROR, MYF(0));
goto err;
}
restore_record(table, s->default_values); // Get default values for fields
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
if (system_charset_info->cset->
numchars(system_charset_info, parse_data->dbname.str,
parse_data->dbname.str +
parse_data->dbname.length) > EVEX_DB_FIELD_LEN)
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
my_error(ER_TOO_LONG_IDENT, MYF(0), parse_data->dbname.str);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
goto err;
}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
if (system_charset_info->cset->
numchars(system_charset_info, parse_data->name.str,
parse_data->name.str +
parse_data->name.length) > EVEX_DB_FIELD_LEN)
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
my_error(ER_TOO_LONG_IDENT, MYF(0), parse_data->name.str);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
goto err;
}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
if (parse_data->body.length > table->field[ET_FIELD_BODY]->field_length)
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
my_error(ER_TOO_LONG_BODY, MYF(0), parse_data->name.str);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
goto err;
}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
if (!(parse_data->expression) && !(parse_data->execute_at.year))
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
DBUG_PRINT("error", ("neither expression nor execute_at are set!"));
my_error(ER_EVENT_NEITHER_M_EXPR_NOR_M_AT, MYF(0));
goto err;
}
((Field_timestamp *)table->field[ET_FIELD_CREATED])->set_time();
/*
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
mysql_event_fill_row() calls my_error() in case of error so no need to
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
handle it here
*/
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if ((ret= mysql_event_fill_row(thd, table, parse_data, FALSE)))
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
goto err;
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
/* Close active transaction only if We are going to modify disk */
if (end_active_trans(thd))
goto err;
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
if (table->file->ha_write_row(table->record[0]))
{
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
my_error(ER_EVENT_STORE_FAILED, MYF(0), parse_data->name.str, ret);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
goto err;
}
*rows_affected= 1;
ok:
if (dbchanged)
Remove trailing whitespace. After merge fixes.
2006-07-10 23:54:46 +02:00
(void) mysql_change_db(thd, old_db.str, 1);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
if (table)
close_thread_tables(thd);
DBUG_RETURN(EVEX_OK);
err:
if (dbchanged)
Remove trailing whitespace. After merge fixes.
2006-07-10 23:54:46 +02:00
(void) mysql_change_db(thd, old_db.str, 1);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
if (table)
close_thread_tables(thd);
DBUG_RETURN(EVEX_GENERAL_ERROR);
}
/*
Used to execute ALTER EVENT. Pendant to Events::update_event().
SYNOPSIS
Event_db_repository::update_event()
thd THD
sp_name the name of the event to alter
et event's data
RETURN VALUE
0 OK
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
EVEX_GENERAL_ERROR Error occured and reported
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
NOTES
sp_name is passed since this is the name of the event to
alter in case of RENAME TO.
*/
int
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
Event_db_repository::update_event(THD *thd, Event_parse_data *parse_data,
sp_name *new_name)
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
CHARSET_INFO *scs= system_charset_info;
Remove trailing whitespace. After merge fixes.
2006-07-10 23:54:46 +02:00
TABLE *table= NULL;
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
int ret= EVEX_OPEN_TABLE_FAILED;
DBUG_ENTER("Event_db_repository::update_event");
if (open_event_table(thd, TL_WRITE, &table))
{
my_error(ER_EVENT_OPEN_TABLE_FAILED, MYF(0));
goto err;
}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if (check_parse_params(thd, parse_data))
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
goto err;
DBUG_PRINT("info", ("dbname: %s", parse_data->dbname.str));
DBUG_PRINT("info", ("name: %s", parse_data->name.str));
DBUG_PRINT("info", ("user: %s", parse_data->definer.str));
if (new_name)
DBUG_PRINT("info", ("rename to: %s", new_name->m_name.str));
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
/* first look whether we overwrite */
if (new_name)
{
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
if (!sortcmp_lex_string(parse_data->name, new_name->m_name, scs) &&
!sortcmp_lex_string(parse_data->dbname, new_name->m_db, scs))
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
my_error(ER_EVENT_SAME_NAME, MYF(0), parse_data->name.str);
goto err;
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if (!find_event_by_name(thd, new_name->m_db, new_name->m_name, table))
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
my_error(ER_EVENT_ALREADY_EXISTS, MYF(0), new_name->m_name.str);
goto err;
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
}
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
}
/*
...and then if there is such an event. Don't exchange the blocks
because you will get error 120 from table handler because new_name will
overwrite the key and SE will tell us that it cannot find the already found
row (copied into record[1] later
*/
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
if (EVEX_KEY_NOT_FOUND == find_event_by_name(thd, parse_data->dbname,
parse_data->name, table))
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
my_error(ER_EVENT_DOES_NOT_EXIST, MYF(0), parse_data->name.str);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
goto err;
}
store_record(table,record[1]);
/* Don't update create on row update. */
table->timestamp_field_type= TIMESTAMP_NO_AUTO_SET;
/*
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
mysql_event_fill_row() calls my_error() in case of error so no need to
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
handle it here
*/
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if ((ret= mysql_event_fill_row(thd, table, parse_data, TRUE)))
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
goto err;
if (new_name)
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
{
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
table->field[ET_FIELD_DB]->
store(new_name->m_db.str, new_name->m_db.length, scs);
table->field[ET_FIELD_NAME]->
store(new_name->m_name.str, new_name->m_name.length, scs);
}
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
/* Close active transaction only if We are going to modify disk */
if (end_active_trans(thd))
goto err;
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
if ((ret= table->file->ha_update_row(table->record[1], table->record[0])))
{
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
my_error(ER_EVENT_STORE_FAILED, MYF(0), parse_data->name.str, ret);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
goto err;
}
/* close mysql.event or we crash later when loading the event from disk */
close_thread_tables(thd);
DBUG_RETURN(0);
err:
if (table)
close_thread_tables(thd);
DBUG_RETURN(EVEX_GENERAL_ERROR);
}
/*
Drops an event
SYNOPSIS
Event_db_repository::drop_event()
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
thd [in] THD
db [in] Database name
name [in] Event's name
drop_if_exists [in] If set and the event not existing => warning
onto the stack
rows_affected [out] Affected number of rows is returned heres
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
RETURN VALUE
0 OK
!0 Error (my_error() called)
*/
int
Event_db_repository::drop_event(THD *thd, LEX_STRING db, LEX_STRING name,
bool drop_if_exists, uint *rows_affected)
{
Remove trailing whitespace. After merge fixes.
2006-07-10 23:54:46 +02:00
TABLE *table= NULL;
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
Open_tables_state backup;
int ret;
DBUG_ENTER("Event_db_repository::drop_event");
DBUG_PRINT("enter", ("db=%s name=%s", db.str, name.str));
ret= EVEX_OPEN_TABLE_FAILED;
thd->reset_n_backup_open_tables_state(&backup);
if (open_event_table(thd, TL_WRITE, &table))
{
my_error(ER_EVENT_OPEN_TABLE_FAILED, MYF(0));
goto done;
}
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
switch ((ret= find_event_by_name(thd, db, name, table))) {
case 0:
/* Close active transaction only if we are actually going to modify disk */
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
if ((ret= end_active_trans(thd)))
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
break;
WL#3337 (Events new architecture) This cut No 7 should finish the part of fixing the parsing of the events : - Event_timed is no more used during parsing. Less problems because it has a mutex. Event_parse_data class is used during parsing. It is suited only for this purpose. It's pretty lightweight - Late checking of data from parsing is being performed. This should solve the problems of nested events in SP or other events (for the situation of no nested bodies). Before if an ALTER EVENT was in a SP, then when the SP was compiled, and not executed, the actual init_xxx methods of Event_timed were called, which is wrong. - It could be a side effect of using a specialized class, but test events_stress is now 25% quicker. Cut No8 will start splitting Event_scheduler into 2 parts, the QUEUE will be moved to Event_queue.
2006-06-29 00:42:25 +02:00
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
if ((ret= table->file->ha_delete_row(table->record[0])))
my_error(ER_EVENT_CANNOT_DELETE, MYF(0));
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
break;
case EVEX_KEY_NOT_FOUND:
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
if (drop_if_exists)
{
push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_NOTE,
ER_SP_DOES_NOT_EXIST, ER(ER_SP_DOES_NOT_EXIST),
"Event", name.str);
ret= 0;
} else
my_error(ER_EVENT_DOES_NOT_EXIST, MYF(0), name.str);
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
break;
default:
;
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
}
done:
Remove trailing whitespace. After merge fixes.
2006-07-10 23:54:46 +02:00
if (table)
close_thread_tables(thd);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
thd->restore_backup_open_tables_state(&backup);
DBUG_RETURN(ret);
}
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
/*
Positions the internal pointer of `table` to the place where (db, name)
is stored.
SYNOPSIS
Event_db_repository::find_event_by_name()
thd Thread
db Schema
name Event name
table Opened mysql.event
RETURN VALUE
0 OK
EVEX_KEY_NOT_FOUND No such event
*/
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
int
Event_db_repository::find_event_by_name(THD *thd, LEX_STRING db,
LEX_STRING name, TABLE *table)
{
byte key[MAX_KEY_LENGTH];
DBUG_ENTER("Event_db_repository::find_event_by_name");
DBUG_PRINT("enter", ("name: %.*s", name.length, name.str));
/*
Create key to find row. We have to use field->store() to be able to
handle VARCHAR and CHAR fields.
Assumption here is that the two first fields in the table are
'db' and 'name' and the first key is the primary key over the
same fields.
*/
if (db.length > table->field[ET_FIELD_DB]->field_length ||
name.length > table->field[ET_FIELD_NAME]->field_length)
DBUG_RETURN(EVEX_KEY_NOT_FOUND);
table->field[ET_FIELD_DB]->store(db.str, db.length, &my_charset_bin);
table->field[ET_FIELD_NAME]->store(name.str, name.length, &my_charset_bin);
WL#3337 (Event scheduler new architecture) Cleaned up the code a bit. Fixed few leaks. This code still does not load events on server startup from disk. The problem is that there is a need for a THD instance, which does not exist during server boot. This will be solved soon. Still Event_timed is used both for the memory queue and for exectution. This will be changed according to WL#3337 probably in the next commit.
2006-07-05 17:12:50 +02:00
key_copy(key, table->record[0], table->key_info, table->key_info->key_length);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
if (table->file->index_read_idx(table->record[0], 0, key,
table->key_info->key_length,
HA_READ_KEY_EXACT))
{
DBUG_PRINT("info", ("Row not found"));
DBUG_RETURN(EVEX_KEY_NOT_FOUND);
}
DBUG_PRINT("info", ("Row found!"));
DBUG_RETURN(0);
}
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
/*
Drops all events in the selected database, from mysql.event.
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
SYNOPSIS
Event_db_repository::drop_schema_events()
thd Thread
schema The database to clean from events
RETURN VALUE
0 OK
!0 Error (Reported)
*/
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
int
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
Event_db_repository::drop_schema_events(THD *thd, LEX_STRING schema)
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
return drop_events_by_field(thd, ET_FIELD_DB, schema);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
}
/*
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
Drops all events by field which has specific value of the field
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
SYNOPSIS
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
Event_db_repository::drop_events_by_field()
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
thd Thread
table mysql.event TABLE
field Which field of the row to use for matching
field_value The value that should match
RETURN VALUE
0 OK
!0 Error from ha_delete_row
*/
int
Event_db_repository::drop_events_by_field(THD *thd,
enum enum_events_table_field field,
LEX_STRING field_value)
{
int ret= 0;
TABLE *table;
Open_tables_state backup;
READ_RECORD read_record_info;
WL #3337 (Event scheduler new architecture) Cut Nr. 8. All tests pass. Separated Event_scheduler into Event_queue and Event_scheduler. Added new Event_scheduler_ng which is the new scheduler and is used system-wide. Will be moved to the event_scheduler.cc in the future. Using Event_timed in Event_queue as well as cloned during execution. Next step is to have Event_worker_data which will be used during execution and will take ::compile()/::execute() out of Event_timed.
2006-07-04 18:44:35 +02:00
DBUG_ENTER("Event_db_repository::drop_events_by_field");
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
DBUG_PRINT("enter", ("field=%d field_value=%s", field, field_value.str));
if (open_event_table(thd, TL_WRITE, &table))
{
my_error(ER_EVENT_OPEN_TABLE_FAILED, MYF(0));
DBUG_RETURN(1);
}
/* only enabled events are in memory, so we go now and delete the rest */
init_read_record(&read_record_info, thd, table, NULL, 1, 0);
while (!ret && !(read_record_info.read_record(&read_record_info)) )
{
char *et_field= get_field(thd->mem_root, table->field[field]);
LEX_STRING et_field_lex= {et_field, strlen(et_field)};
DBUG_PRINT("info", ("Current event %s name=%s", et_field,
get_field(thd->mem_root, table->field[ET_FIELD_NAME])));
if (!sortcmp_lex_string(et_field_lex, field_value, system_charset_info))
{
DBUG_PRINT("info", ("Dropping"));
if ((ret= table->file->ha_delete_row(table->record[0])))
my_error(ER_EVENT_DROP_FAILED, MYF(0),
get_field(thd->mem_root, table->field[ET_FIELD_NAME]));
}
}
end_read_record(&read_record_info);
thd->version--; /* Force close to free memory */
DBUG_RETURN(ret);
}
/*
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
Looks for a named event in mysql.event and in case of success returns
an object will data loaded from the table.
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
SYNOPSIS
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
Event_db_repository::find_event()
thd [in] THD
name [in] The name of the event to find
ett [out] Event's data if event is found
tbl [in] TABLE object to use when not NULL
NOTES
1) Use sp_name for look up, return in **ett if found
2) tbl is not closed at exit
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
RETURN VALUE
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
0 ok In this case *ett is set to the event
# error *ett == 0
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
*/
int
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
Event_db_repository::find_event(THD *thd, LEX_STRING dbname, LEX_STRING name,
Event_basic *et)
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
Remove trailing whitespace. After merge fixes.
2006-07-10 23:54:46 +02:00
TABLE *table= NULL;
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
int ret;
DBUG_ENTER("Event_db_repository::find_event");
DBUG_PRINT("enter", ("name: %*s", name.length, name.str));
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if (open_event_table(thd, TL_READ, &table))
{
my_error(ER_EVENT_OPEN_TABLE_FAILED, MYF(0));
ret= EVEX_GENERAL_ERROR;
goto done;
}
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
if ((ret= find_event_by_name(thd, dbname, name, table)))
{
my_error(ER_EVENT_DOES_NOT_EXIST, MYF(0), name.str);
goto done;
}
/*
1)The table should not be closed beforehand. ::load_from_row() only loads
and does not compile
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
2)::load_from_row() is silent on error therefore we emit error msg here
*/
if ((ret= et->load_from_row(table)))
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
{
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
my_error(ER_CANNOT_LOAD_FROM_TABLE, MYF(0), "event");
goto done;
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
}
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
done:
if (table)
close_thread_tables(thd);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
DBUG_RETURN(ret);
WL#3337 (Events new architecture) 5th cut, moved DB related code to Event_db_repository and updated accordingly the remanining code. Moved change/restore_security_context() to class THD Removed events_priv.h Next step is to reorganize create/update_event() and parsing for them. But probably some other refactoring could be done in the meanwhile. The changes so far pass the test suite.
2006-06-28 01:28:03 +02:00
}
WL#3337 (Event scheduler new architecture) Cleaned up the code a bit. Fixed few leaks. This code still does not load events on server startup from disk. The problem is that there is a need for a THD instance, which does not exist during server boot. This will be solved soon. Still Event_timed is used both for the memory queue and for exectution. This will be changed according to WL#3337 probably in the next commit.
2006-07-05 17:12:50 +02:00
/*
Looks for a named event in mysql.event and then loads it from
the table, compiles and inserts it into the cache.
SYNOPSIS
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
Event_db_repository::load_named_event()
thd [in] THD
dbname [in] Event's db name
name [in] Event's name
etn_new [out] The loaded event
WL#3337 (Event scheduler new architecture) Cleaned up the code a bit. Fixed few leaks. This code still does not load events on server startup from disk. The problem is that there is a need for a THD instance, which does not exist during server boot. This will be solved soon. Still Event_timed is used both for the memory queue and for exectution. This will be changed according to WL#3337 probably in the next commit.
2006-07-05 17:12:50 +02:00
RETURN VALUE
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
OP_OK OK
OP_LOAD_ERROR Error during loading from disk
WL#3337 (Event scheduler new architecture) Cleaned up the code a bit. Fixed few leaks. This code still does not load events on server startup from disk. The problem is that there is a need for a THD instance, which does not exist during server boot. This will be solved soon. Still Event_timed is used both for the memory queue and for exectution. This will be changed according to WL#3337 probably in the next commit.
2006-07-05 17:12:50 +02:00
*/
int
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
Event_db_repository::load_named_event(THD *thd, LEX_STRING dbname,
LEX_STRING name, Event_basic *etn)
WL#3337 (Event scheduler new architecture) Cleaned up the code a bit. Fixed few leaks. This code still does not load events on server startup from disk. The problem is that there is a need for a THD instance, which does not exist during server boot. This will be solved soon. Still Event_timed is used both for the memory queue and for exectution. This will be changed according to WL#3337 probably in the next commit.
2006-07-05 17:12:50 +02:00
{
int ret= 0;
Open_tables_state backup;
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
DBUG_ENTER("Event_db_repository::load_named_event");
DBUG_PRINT("enter",("thd=0x%lx name:%*s",thd, name.length, name.str));
WL#3337 (Event scheduler new architecture) Cleaned up the code a bit. Fixed few leaks. This code still does not load events on server startup from disk. The problem is that there is a need for a THD instance, which does not exist during server boot. This will be solved soon. Still Event_timed is used both for the memory queue and for exectution. This will be changed according to WL#3337 probably in the next commit.
2006-07-05 17:12:50 +02:00
thd->reset_n_backup_open_tables_state(&backup);
WL#3337 (Event scheduler new architecture) This patch introduces specialized Event data objects Event_basic as parent. Event_queue_element used for queue storage Event_timed used for SHOW EVENTS/ I_S.EVENTS / SHOW CREATE EVENT Event_job_data using during execution. Methods were moved out of Event_timed to other classes. This patch also introduces Events::LOCK_event_metadata. This patch gives new implementation of Events::dump_internal_status(). Now both the Event_scheduler and Event_queue return information during their ::dump_internal_status(). Shortened a bit the runtime for executing events test cases.
2006-07-10 13:44:43 +02:00
/* No need to use my_error() here because find_event() has done it */
ret= find_event(thd, dbname, name, etn);
WL#3337 (Event scheduler new architecture) Cleaned up the code a bit. Fixed few leaks. This code still does not load events on server startup from disk. The problem is that there is a need for a THD instance, which does not exist during server boot. This will be solved soon. Still Event_timed is used both for the memory queue and for exectution. This will be changed according to WL#3337 probably in the next commit.
2006-07-05 17:12:50 +02:00
thd->restore_backup_open_tables_state(&backup);
/* In this case no memory was allocated so we don't need to clean */
if (ret)
DBUG_RETURN(OP_LOAD_ERROR);
DBUG_RETURN(OP_OK);
}
Reference in a new issue Copy permalink