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.
mysql-test/r/events.result:
update results
mysql-test/r/events_bugs.result:
update results
mysql-test/r/events_logs_tests.result:
update results
mysql-test/r/events_scheduling.result:
update results
mysql-test/t/events.test:
update test
make --sleep more appropriate . saving some time could mean failure on loaded boxes though :(
add tests for previously uncovered branches.
mysql-test/t/events_bugs.test:
update test
make --sleep more appropriate . saving some time could mean failure on loaded boxes though :(
add tests for previously uncovered branches.
mysql-test/t/events_logs_tests.test:
make the test shorter by time
mysql-test/t/events_scheduling.test:
when selecting always use ORDER BY
mysql-test/t/events_stress.test:
sleep 2.5secs for shorter stress test
sql/event_data_objects.cc:
Event_timed is no more used during execution.
Event_timed is no more used during in the memory queue.
Event_timed is only used for SHOW CREATE EVENT/ I_S.EVENTS/ SHOW EVENTS
Event_basic is the parent of almost all Event data objects.
Event_basic -> Event_queue_element (used for the memory queue) -> Event_timed
Event_basic -> Event_job_data (the object used for execution)
Sql_alloc -> Event_parse_data (used during parsing)
sql/event_data_objects.h:
Event_timed is no more used during execution.
Event_timed is no more used during in the memory queue.
Event_timed is only used for SHOW CREATE EVENT/ I_S.EVENTS/ SHOW EVENTS
Event_basic is the parent of almost all Event data objects.
Event_basic -> Event_queue_element (used for the memory queue) -> Event_timed
Event_basic -> Event_job_data (the object used for execution)
Sql_alloc -> Event_parse_data (used during parsing)
sql/event_db_repository.cc:
Cosmetics.
load_named_event now uses Event_basic, for polymorphism
find_event uses Event_basic, to be polymorphic.
use Field **fields= table->field and then index fields[...]
Add documentation.
Fix documentation.
sql/event_db_repository.h:
Event_db_repository depends only on Event_basic's interface
sql/event_queue.cc:
Cosmetics.
Don't use Event_timed for the queue and giving back object for execution.
Event_queue_element is for the queue, Event_job_data is for execution.
Add Event_queue::dump_internal_status() for SHOW SCHEDULER STATUS command
sql/event_queue.h:
Cosmetics.
Don't use Event_timed for the queue and giving back object for execution.
Event_queue_element is for the queue, Event_job_data is for execution.
Add Event_queue::dump_internal_status() for SHOW SCHEDULER STATUS command
sql/event_scheduler_ng.cc:
Add back Event_scheduler::cond_wait()
Add back Event_scheduler::dump_internal_status()
Using Event_job_data for execution. Make the scheduler thread unkillable (thd->command= COM_DAEMON).
Add a lot of documentation.
sql/event_scheduler_ng.h:
Add back Event_scheduler::cond_wait()
Add back Event_scheduler::dump_internal_status()
Using Event_job_data for execution.
sql/events.cc:
Documentation
Add LOCK_event_metadata
sql/events.h:
Change the signature of Events::drop_event() not to use sp_name but LEX_STRING
sql/share/errmsg.txt:
Fix error message
sql/sql_parse.cc:
Events::drop_event() has new signature
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.
sql/event_data_objects.cc:
Strip unneeded stuff from class Event_timed
Event_timed is still used for the queue and execution.
That will be changed in the next commit.
sql/event_data_objects.h:
Strip unneeded stuff from class Event_timed
Event_timed is still used for the queue and execution.
That will be changed in the next commit.
sql/event_db_repository.cc:
Cosmetics.
Add a new method load_named_event_job, to be made complete in the
next commit. It will load from disk an instance of Event_job_data to
be used during execution.
sql/event_db_repository.h:
find_event does not need MEM_ROOT anymore
because the memory is allocated on Event's own root.
sql/event_queue.cc:
Remove dead code.
Move dumping of the queue to separate method.
Make critical sections in create_event & update_event
as small as possible - load the new event outside of the section
and free the object also outside of it.
sql/event_queue.h:
init -> init_queue -> easier for ctags
deinit -> deinit_queue -> easier for ctags
sql/event_scheduler.cc:
empty this file
sql/event_scheduler.h:
empty this file
sql/event_scheduler_ng.cc:
add back DBUG_RETURN(0) in thread handlers.
We don't stop running events when stopping the scheduler. Therefore
remove this method now. If it is needed later it can be added back.
sql/event_scheduler_ng.h:
Remove stop_all_running_threads()
init -> init_scheduler
deinit -> deinit_scheduler
easier for ctags
sql/events.cc:
Cosmetics
sql/events.h:
Cosmetics
sql/set_var.cc:
Remove references to dead code
sql/sql_parse.cc:
Reorganize a bit.
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.
mysql-test/r/events.result:
update result
mysql-test/r/events_bugs.result:
update result
mysql-test/r/ps_1general.result:
update result
mysql-test/r/skip_name_resolve.result:
update result
mysql-test/r/sp-threads.result:
update result
mysql-test/r/sp_notembedded.result:
update result
mysql-test/r/status.result:
update result
mysql-test/t/events_stress.test:
Make event_stress a bit longer
sql/Makefile.am:
Add new event_scheduler_ng.h/cc . These are only to be in the experimental
clone. Later their content will be moved to event_scheduler.h/cc
sql/event_data_objects.cc:
Allocate strings memory on own memory root, instead
on the schedulers. Thus don't "leak" memory. This should
fix bug#18683 memory leak in event scheduler
sql/event_data_objects.h:
add mem_root
add THD - this is only temporal, will be moved to class Event_job_data
once Event_job_data is responsible for the execution.
sql/event_db_repository.cc:
Remove unused code.
Cosmetic changes
sql/event_queue.cc:
Now use the Event_scheduler_ng (NextGen)
sql/event_queue.h:
Now use the Event_scheduler_ng (NextGen)
sql/event_scheduler.cc:
This file is no more used, but will be soon.
sql/event_scheduler.h:
This file is no more used but will be soon
sql/events.cc:
Now use the Event_scheduler_ng (NextGen)
sql/events.h:
Now use the Event_scheduler_ng (NextGen)
sql/mysqld.cc:
Make it again possible to kill the scheduler thread
sql/set_var.cc:
Now use the Event_scheduler_ng (NextGen)
sql/share/errmsg.txt:
Shorten the message.
sql/sql_show.cc:
Loading is on a own root, then don't use thd->mem_root
This is the first cut of separating Event_scheduler in two
classes which are more specialized.
Inheritance was used to separate methods and member variables.
Still Event_scheduler is a child of Event_queue. This dependency
will be removed soon.
sql/event_data_objects.cc:
add comments
sql/event_db_repository.cc:
coding style
sql/event_db_repository.h:
add a call, will be implemented later
sql/event_queue.cc:
Event_queue, still as super-class of Event_scheduler
sql/event_queue.h:
Event_queue as super-class of Event_scheduler. Trying to
separate the two classes
sql/event_scheduler.cc:
Event_scheduler as child class of Event_queue.
Trying to separate both classes.
sql/event_scheduler.h:
Event_scheduler as child class of Event_queue.
Trying to separate both classes.
sql/events.cc:
Don't allocate on the stack the scheduler but on the heap.
The exact way it is done will be changed, that's ok for now.
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.
mysql-test/r/events.result:
update result
mysql-test/t/events.test:
disabled is actually wrong, should be disable, but because of the early
checking it was never parsed.
sql/event_data_objects.cc:
move add init_xxx methods from Event_timed to Event_parse_data
Event_parse data does not need definer_user and definer_host
in Event_timed::compile() do not use lex.et, well there is no more lex.et :)
sql/event_data_objects.h:
move parsing responsibilities from Event_timed to Event_parse_data
sql/event_db_repository.cc:
No more Event_timed comes from parsing but Event_parse_data
The initialization of Item*-s from parsing is done late, and not
during the actual parsing. This is the right way to go because
if an ALTER EVENT is inside a SP or CREATE EVENT it should not be
executed (initialized) during parsing, as it was done.
sql/event_db_repository.h:
No more Event_timed comes from parsing but Event_parse_data
The initialization of Item*-s from parsing is done late, and not
during the actual parsing. This is the right way to go because
if an ALTER EVENT is inside a SP or CREATE EVENT it should not be
executed (initialized) during parsing, as it was done.
sql/event_scheduler.cc:
No more Event_timed comes from parsing but Event_parse_data
The initialization of Item*-s from parsing is done late, and not
during the actual parsing. This is the right way to go because
if an ALTER EVENT is inside a SP or CREATE EVENT it should not be
executed (initialized) during parsing, as it was done.
sql/event_scheduler.h:
No more Event_timed comes from parsing but Event_parse_data
The initialization of Item*-s from parsing is done late, and not
during the actual parsing. This is the right way to go because
if an ALTER EVENT is inside a SP or CREATE EVENT it should not be
executed (initialized) during parsing, as it was done.
sql/events.cc:
No more Event_timed comes from parsing but Event_parse_data
The initialization of Item*-s from parsing is done late, and not
during the actual parsing. This is the right way to go because
if an ALTER EVENT is inside a SP or CREATE EVENT it should not be
executed (initialized) during parsing, as it was done.
sql/events.h:
No more Event_timed comes from parsing but Event_parse_data
The initialization of Item*-s from parsing is done late, and not
during the actual parsing. This is the right way to go because
if an ALTER EVENT is inside a SP or CREATE EVENT it should not be
executed (initialized) during parsing, as it was done.
sql/sql_lex.cc:
lex->et_compile_phase and lex->et are no more.
Use lex->event_parse_data
sql/sql_lex.h:
lex->et_compile_phase and lex->et are no more.
Use lex->event_parse_data
sql/sql_parse.cc:
lex->et_compile_phase and lex->et are no more.
Use lex->event_parse_data
ACL checks were moved inside the Events subsystem.
Also ending of the transaction is performed only just
before doing disk operation. Therefore only when needed.
sql/sql_yacc.yy:
lex->et and lex->et_parse_phase are no more
Use the specialized for parsing Event_parse_data
sql/CMakeLists.txt:
Auto merged
sql/events.cc:
Auto merged
sql/mysqld.cc:
Auto merged
sql/set_var.cc:
Auto merged
sql/sql_class.cc:
Auto merged
sql/sql_class.h:
Auto merged
sql/sql_lex.h:
Auto merged
sql/sql_parse.cc:
Auto merged
sql/sql_yacc.yy:
Auto merged
sql/sql_show.cc:
SCCS merged
Cut 7 (refactoring)
db_repository is no more embedded in the Events
singleton. Therefore a change to Events_db_repository
won't mean recompile of all files in the server which include events.h
sql/event_data_objects.cc:
db_repository is no more embedded in the Events
singleton. Therefore a change to Events_db_repository
won't mean recompile of all files in the server which include events.h
sql/event_db_repository.cc:
db_repository is no more embedded in the Events
singleton. Therefore a change to Events_db_repository
won't mean recompile of all files in the server which include events.h
sql/events.cc:
db_repository is no more embedded in the Events
singleton. Therefore a change to Events_db_repository
won't mean recompile of all files in the server which include events.h
sql/events.h:
db_repository is no more embedded in the Events
singleton. Therefore a change to Events_db_repository
won't mean recompile of all files in the server which include events.h
sql/mysqld.cc:
db_repository is no more embedded in the Events
singleton. Therefore a change to Events_db_repository
won't mean recompile of all files in the server which include events.h
Cut number 6. Move code from sql_show.cc to event_db_repository.cc
that more belongs to the latter.
sql/event_db_repository.cc:
move code that works with mysql.event from sql_show.cc to
event_db_repository.cc . Route through Event_db_repository's interface
which is proxied by class Events. The code relies on a function from
sql_show.cc which does the actual storage in the schema table. I think
it's better to leave the function there because the structure of
I_S.EVENTS is defined in sql_show.cc
sql/event_db_repository.h:
I_S / SHOW EVENTS handling hooks
sql/event_scheduler.cc:
use the pointer to db_repository which Event_scheduler already has
sql/events.cc:
Put a comment to get_instance
sql/events.h:
callback for I_S (sql_show.cc)
sql/sql_show.cc:
move code that belongs more to Event_db_repository than to here.
Use a callback of class Events. Only 1 function is left here, because
it copies data into the actual rows of I_S.EVENTS and belongs to this file.
sql/sql_show.h:
export this function will be called from event_db_repository.cc
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.
BitKeeper/deleted/.del-events_priv.h~2e8bce2cf35997df:
Delete: sql/events_priv.h
sql/Makefile.am:
events_priv.h is no more
sql/event_data_objects.cc:
reorganize events code
sql/event_data_objects.h:
reorganize events code
sql/event_db_repository.cc:
reorganize events code
sql/event_db_repository.h:
reorganize events code
sql/event_scheduler.cc:
reorganize events code
sql/event_scheduler.h:
reorganize events code
sql/events.cc:
reorganize events code
sql/events.h:
reorganize events code
sql/mysqld.cc:
reorganize events code
sql/set_var.cc:
reorganize events code
sql/sql_class.cc:
add ::change_security_context() and restore_security_context()
sql/sql_class.h:
add ::change_security_context() and restore_security_context()
sql/sql_db.cc:
reorganize Events code
sql/sql_parse.cc:
reorganize Events code
sql/sql_show.cc:
reorganize Events code
Third cut to simplify parsing phase. Now DROP EVENT works.
Overloaded few functions to be able to use either sp_name or pass two LEX_STRINGs
instead of a Event_timed pointer. This is transitional and eventually the old
functions will be removed. For now DROP EVENT also works, does not need anymore
a parsing object (Event_timed) and definer initialization because everyone who
has EVENT_ACL can drop events, and this is checked on execution time in sql_parse.cc
from the security context, as it should be.
sql/event_data_objects.cc:
overload few functions
sql/event_scheduler.cc:
Event_scheduler::drop_event() actually does not need Event_timed object
but just an identifier, hence pass only sp_name.
Overloaded Event_scheduler::find_event() to work with sp_name object. Eventually
the old version will be removed. This is being done as transitional step to
be able to test frequently code.
sql/event_scheduler.h:
Event_scheduler::drop_event() actually does not need Event_timed object
but just an identifier, hence pass only sp_name.
Overloaded Event_scheduler::find_event() to work with sp_name object. Eventually
the old version will be removed. This is being done as transitional step to
be able to test frequently code.
sql/events.cc:
Change db_drop_event() not to use Event_timed, either coming from parsing
or from Event_timed::drop, but use LEX_STRINGs. sp_name is not convinient
because in Event_timed::drop a temporary object has to be created. Hence,
dereference the sp_name in Events::drop_event() and pass the LEX_STRINGs.
sql/events.h:
Change db_drop_event() not to use Event_timed, either coming from parsing
or from Event_timed::drop, but use LEX_STRINGs. sp_name is not convinient
because in Event_timed::drop a temporary object has to be created. Hence,
dereference the sp_name in Events::drop_event() and pass the LEX_STRINGs.
sql/events_priv.h:
Change db_drop_event() not to use Event_timed, either coming from parsing
or from Event_timed::drop, but use LEX_STRINGs. sp_name is not convinient
because in Event_timed::drop a temporary object has to be created. Hence,
dereference the sp_name in Events::drop_event() and pass the LEX_STRINGs.
sql/sql_parse.cc:
SQLCOM_DROP_EVENT does not need lex->event_parse_data object and
is more like SQLCOM_SHOW_CREATE_EVENT. Therefore, move it to the block that
handles the latter.
sql/sql_yacc.yy:
DROP EVENT does not need a parsing object, just a name.
Store it as lex->spname. Pretty similar handling to the one of
SHOW CREATE EVENT.
Second cut of separating parsing phase from execution phase
Separate Event_timed from parsing phase and introducing Event_parse_data.
sql/event_data_objects.cc:
second cut,
copy Event_timed::init_body() to Event_parse_data::init_body()
Init the body during parsing, everything else keep as a pointer to
Item or some other pointer to use for later initialization.
sql/event_data_objects.h:
get the identifier as sp_name*, later will initialize our structures
sql/events.cc:
for easy transition add temporarily parse_data, later Event_timed *et will be removed.
Do slow transition because Event_timed is so tightly integrated that a front-attack
by removing things from this class was unsuccessful. Do things step by step by eliminating
dependencies. Hence, the code could be checked with the current test suite too
(early testing)
sql/events.h:
for easy transition add temporarily parse_data, later Event_timed *et will be removed.
Do slow transition because Event_timed is so tightly integrated that a front-attack
by removing things from this class was unsuccessful. Do things step by step by eliminating
dependencies. Hence, the code could be checked with the current test suite too
(early testing)
BitKeeper/etc/ignore:
Added libmysql/viosocket.o.6WmSJk libmysqld/event_data_objects.cc libmysqld/event_db_repository.cc libmysqld/event_queue.cc server-tools/instance-manager/net_serv.cc to the ignore list
sql/share/errmsg.txt:
remove this message, not used and needed for now
sql/sql_lex.h:
for easy transition add temporarily parse_data, later Event_timed *et will be removed.
Do slow transition because Event_timed is so tightly integrated that a front-attack
by removing things from this class was unsuccessful. Do things step by step by eliminating
dependencies. Hence, the code could be checked with the current test suite too
(early testing)
sql/sql_parse.cc:
for easy transition add temporarily parse_data, later Event_timed *et will be removed.
Do slow transition because Event_timed is so tightly integrated that a front-attack
by removing things from this class was unsuccessful. Do things step by step by eliminating
dependencies. Hence, the code could be checked with the current test suite too
(early testing)
sql/sql_yacc.yy:
for easy transition add temporarily parse_data, later Event_timed *et will be removed.
Do slow transition because Event_timed is so tightly integrated that a front-attack
by removing things from this class was unsuccessful. Do things step by step by eliminating
dependencies. Hence, the code could be checked with the current test suite too
(early testing)
Infrastructure built. Added the foreseen files and change Makefile.am/CMakeLists.txt
accordingly.
libmysqld/Makefile.am:
add new files. this is first cut of WL3337u
sql/CMakeLists.txt:
add more files to build
sql/Makefile.am:
add new files. this is first cut of WL3337
sql/event_scheduler.cc:
event_timed.h -> event_data_objects.h (WL#3337)
sql/events.cc:
event_timed.h -> event_data_objects.h (WL#3337)
sql/share/errmsg.txt:
new error message
sql/event_data_objects.cc:
event_timed.h -> event_data_objects.h (WL#3337)
sql/event_data_objects.h:
event_timed.h -> event_data_objects.h (WL#3337)
sql/sql_parse.cc:
event_timed.h -> event_data_objects.h (WL#3337)
sql/sql_show.cc:
event_timed.h -> event_data_objects.h (WL#3337)
sql/sql_yacc.yy:
event_timed.h -> event_data_objects.h (WL#3337)
into mysql.com:/opt/local/work/mysql-5.1-runtime
mysql-test/r/information_schema.result:
Auto merged
sql/CMakeLists.txt:
Auto merged
sql/mysql_priv.h:
Auto merged
sql/mysqld.cc:
Auto merged
sql/events.cc:
Auto merged
sql/sql_parse.cc:
Auto merged
sql/sql_show.cc:
Auto merged
sql/sql_yacc.yy:
Auto merged
mysql-test/t/events_microsec.test:
SCCS merged
BitKeeper/etc/ignore:
auto-union
libmysqld/Makefile.am:
Auto merged
sql/Makefile.am:
Auto merged
sql/events.cc:
Auto merged
sql/mysql_priv.h:
Auto merged
sql/mysqld.cc:
Auto merged
sql/sql_db.cc:
Auto merged
sql/sql_parse.cc:
Auto merged
sql/sql_show.cc:
Auto merged
sql/sql_yacc.yy:
Auto merged
sql/time.cc:
Auto merged
Unify method naming -> create/update/drop_event
Move class Event_timed to event_timed.h
class Events is in events.h (renamed from event.h)
The implementation is in events.cc (renamed from event.h)
BitKeeper/deleted/.del-event_executor.cc~f4a4645b973838ab:
Delete: sql/event_executor.cc
include/my_time.h:
add a boundary
libmysqld/CMakeLists.txt:
event.cc -> events.cc
libmysqld/Makefile.am:
event.cc -> event.cc
sql/CMakeLists.txt:
event.cc -> events.cc
sql/Makefile.am:
event.cc -> events.cc
event_priv.h -> events_priv.h
sql/event_scheduler.cc:
event.h -> events.h
add_event -> create_event
replace_event -> update_event()
event_timed_compare_q is only used in event_scheduler.cc , so moving it
from event.cc and making it static
sql/event_scheduler.h:
add_event -> create_event
replace_event -> update_event
sql/event_timed.cc:
moved extern interval_type_to_name to mysql_priv.h
sql/mysql_priv.h:
moved my_time_compare to time.cc
sql/mysqld.cc:
event.h -> events.h
sql/sql_db.cc:
event.h -> events.h
sql/sql_parse.cc:
event.h -> events.h
class Event_timed moved to event_timed.h
sql/sql_show.cc:
event.h -> events.h
class Event_timed moved to event_timed.h
sql/sql_yacc.yy:
event.h -> events.h
class Event_timed moved to event_timed.h
sql/events.cc:
add_event -> create_event
replace_event -> update_event
event_timed_compare_q moved to event_scheduler.cc
sql/events.h:
class Event_timed moved to event_timed.h
sql/events_priv.h:
my_time_compare moved to mysql_priv.h
event_timed_compare_q is static in event_scheduler.cc
sql/time.cc:
moved interval_type_to_name from event.cc to here.
BitKeeper/etc/ignore:
Added libmysqld/events.cc to the ignore list
