An error occurred fetching the project authors.
- 23 Nov, 2006 2 commits
-
-
unknown authored
Removing valgrind memory checks from code. sql/log_event.cc: Removing valgrind memory checks from code.
-
unknown authored
Patch to fix valgrind warning that uninitialized bytes were written. The warning was caused by set_default(), which assumes that the field pointer is pointing into record[0], while all other functions honors the move_field_offset(). sql/log_event.cc: Changing unpack_row() to always unpack into table->record[0] since the Field class hierarchy contain too many dependencies on the record to operate on being in table->record[0]. Changing code to use the new unpack_row(), which requires some juggling of records in one case.
-
- 21 Nov, 2006 1 commit
-
-
unknown authored
Removing DBUG_DUMP printouts for valgrind builds since they trigger warnings. Removing valgrind memory checks completely. Removing bzero() of record when opening table that was added earlier. sql/log_event.cc: Removing DBUG_DUMP of records for valgrind builds since they trigger false valgrind warnings. Removing valgrind checks completely, since they can trigger false valgrind warnings. sql/sql_class.cc: Removing DBUG_DUMP of records for valgrind builds since they trigger false valgrind warnings. sql/table.cc: Removing extreneous bzero() that were added previously.
-
- 20 Nov, 2006 1 commit
-
-
unknown authored
Fix to correct behaviour of find_and_fetch_row() for tables that have primary keys stored in storage engines that support the fast method to fetch rows given a primary key. The method uses position() to retrieve the key for a given record and rnd_pos() to position the internal "cursor" at the row. Rnd_pos() returns the found record in table->record[0], so the record has to be moved to table->record[1] for further processing after calling find_and_fetch_row(). sql/log_event.cc: Adding code to one exit branch of find_and_fetch_row() to move output record from table->record[0] to table->record[1]. Adding function to do valgrind memory check. Adding valgrind memory checks to check that records are defined when they should be. Adding Doxygen comment to find_and_fetch_row() with pre- and post-conditions.
-
- 14 Nov, 2006 1 commit
-
-
unknown authored
ver 5.0 and 5.1 refinement. adding to the MAX_SIZE_LOG_EVENT_STATUS estimation status vars and EXECUTE_LOAD_QUERY_EXTRA_HEADER_LEN sql/log_event.cc: assertion on size of status vars sql/log_event.h: MAX_SIZE_LOG_EVENT_STATUS as the max bytes of contribution of status vars into repl message; MAX_SIZE_LOG_EVENT_STATUS is getting more wrt to 4.1 version.
-
- 03 Nov, 2006 2 commits
-
-
unknown authored
sql/log_event.cc: Fixing comment sql/rpl_rli.cc: Removing code that was moved to other file.
-
unknown authored
Bug#23831 deadlock not noticed RBR bug in that when replicated msta (multi-statement-trans-action) deadlocks with a local at write row event or gets timed-out, the event handler did not return the correct error code. Wrong error code stops slave sql thread instead of to proceed with rollback and replay. The correct code is typed in error log and stored for error handling rotine to conduct rollback and replay of the transaction. The handling for the rbr remains the same as for the sbr events. Particularly, timed-out transaction still is rolled back - look at the related bugs. mysql-test/extra/rpl_tests/rpl_deadlock.test: fixing deadlock stimulation, commented on potential race condition and left todo. mysql-test/r/rpl_deadlock_innodb.result: changed sql/log_event.cc: returning the correct errors for write row event type; storing DL, TO error codes to check at grand-caller in step with sbr.
-
- 01 Nov, 2006 2 commits
-
-
unknown authored
Changes to the code are necessary to compile on Windows. Typecasts are needed in several DBUG_DUMP statements. CAB 01 Nov 2006 sql/log_event.cc: The changes are so that the code will compile on Windows. Typecasts are needed on the DBUG_DUMP statements to cast a byte * to a char *. sql/sql_class.cc: The changes are necessary to get the code to compile on Windows. The DBUG_DUMP statements need a typecast.
-
unknown authored
Not replicating the mysql database *at all* any more. All changes to mysql tables are replicated by replicating the statements that do the changes. mysql-test/r/rpl_do_grant.result: Result change mysql-test/t/rpl_do_grant.test: Deleting grants on both master and server since the mysql database is not replicated any more. sql/handler.cc: Removing code that selectively checks what tables in the mysql database to ignore. The entire mysql database is not replicated any more. sql/log_event.cc: Disabling code to reload grants and access when certain tables are touched. sql/rpl_rli.h: Removing unused code. sql/slave.cc: Removing unused code. sql/sql_class.cc: Mysql queries (i.e., queries that affects tables in the mysql database) are now replicated by statement instead of replicating the tables in the mysql database.
-
- 13 Oct, 2006 1 commit
-
-
unknown authored
VisualStudio needs cast from (byte*) to (const char*), 3rd arg to _db_dump_() sql/sql_class.cc: VisualStudio needs cast from (byte*) to (const char*), 3rd arg to _db_dump_() sql/log_event.cc: VisualStudio needs cast from (byte*) to (const char*), 3rd arg to _db_dump_()
-
- 10 Oct, 2006 2 commits
-
-
unknown authored
Removing code to step the group log position and just stepping the event log position. If the group log position were stepped one time too many, it might be that the group starts at a position that is not possible, e.g., at a Rows_log_event, or between an Intvar_log_event and the following associated Query_log_event. sql/log_event.cc: Removing code to step the group log position and just stepping the event log position. If the group log position were stepped one time too many, it might be that the group starts at a position that is not possible, e.g., at a Rows_log_event, or between an Intvar_log_event and the following associated Query_log_event. sql/slave.cc: Removing code to step the group log position and just stepping the event log position when executing a Format_description_log_event. If the group log position were stepped one time too many, it might be that the group starts at a position that is not possible, e.g., at a Rows_log_event, or between an Intvar_log_event and the following associated Query_log_event.
-
unknown authored
sql/log_event.cc: Adding casts to make it compile on Windows.
-
- 06 Oct, 2006 2 commits
-
-
unknown authored
Note: bug#21726 does not directly apply to 4.1, as it doesn't have stored procedures. However, 4.1 had some bugs that were fixed in 5.0 by the patch for bug#21726, and this patch is a backport of those fixes. Namely, in 4.1 it fixes: - LAST_INSERT_ID(expr) didn't return value of expr (4.1 specific). - LAST_INSERT_ID() could return the value generated by current statement if the call happens after the generation, like in CREATE TABLE t1 (i INT AUTO_INCREMENT PRIMARY KEY, j INT); INSERT INTO t1 VALUES (NULL, 0), (NULL, LAST_INSERT_ID()); - Redundant binary log LAST_INSERT_ID_EVENTs could be generated. mysql-test/r/rpl_insert_id.result: Add result for bug#21726: Incorrect result with multiple invocations of LAST_INSERT_ID. mysql-test/t/rpl_insert_id.test: Add test case for bug#21726: Incorrect result with multiple invocations of LAST_INSERT_ID. sql/item_func.cc: Add implementation of Item_func_last_insert_id::fix_fields(), where we set THD::last_insert_id_used when statement calls LAST_INSERT_ID(). In Item_func_last_insert_id::val_int(), return THD::current_insert_id if called like LAST_INSERT_ID(), otherwise return value of argument if called like LAST_INSERT_ID(expr). sql/item_func.h: Add declaration of Item_func_last_insert_id::fix_fields(). sql/log_event.cc: Do not set THD::last_insert_id_used on LAST_INSERT_ID_EVENT. Though we know the statement will call LAST_INSERT_ID(), it wasn't called yet. sql/set_var.cc: In sys_var_last_insert_id::value_ptr(), set THD::last_insert_id_used, and return THD::current_insert_id for @@LAST_INSERT_ID. sql/sql_class.h: Update comments. Remove THD::insert_id(), as it has lost its purpose now. sql/sql_insert.cc: Now it is OK to read THD::last_insert_id directly. sql/sql_load.cc: Now it is OK to read THD::last_insert_id directly. sql/sql_parse.cc: In mysql_execute_command(), remember THD::last_insert_id (first generated value of the previous statement) in THD::current_insert_id, which then will be returned for LAST_INSERT_ID() and @@LAST_INSERT_ID. sql/sql_select.cc: If "IS NULL" is replaced with "= <LAST_INSERT_ID>", use right value, which is THD::current_insert_id, and also set THD::last_insert_id_used to issue binary log LAST_INSERT_ID_EVENT. sql/sql_update.cc: Now it is OK to read THD::last_insert_id directly. tests/mysql_client_test.c: Add test case for bug#21726: Incorrect result with multiple invocations of LAST_INSERT_ID.
-
unknown authored
crash for, e.g., NDB): Before, mysqlbinlog printed table map events as a separate statement, so when executing the event, the opened table was subsequently closed when the statement ended. Instead, the row-based events that make up a statement are now printed as *one* BINLOG statement, which means that the table maps and the following *_rows_log_event events are executed fully before the statement ends. Changing implementation of BINLOG statement to be able to read the emitted format, which now consists of several chunks of BASE64-encoded data. client/mysqlbinlog.cc: Using IO_CACHE to print events instead of directly to file. Factoring out code to write event header and base64 representation into separate function. mysys/mf_iocache2.c: Correcting name in documentation. sql/log_event.cc: Adding class Write_on_release_cache that holds an IO_CACHE and that will write contents of IO_CACHE to a designated file on destruction. Changing signature of event printing functions print_header() and print_base64() to write to IO_CACHE and changing *all* calls in those functions in accordance. This means that all printing functions now print to an IO_CACHE instead of to a file, and that the IO_CACHE is then copied to the file. The print() function have the same signature as before, but since it is using print_header() and print_base64(), the data will now be printed to an IO_CACHE and then copied to the file. Changing row-based replication events to incrementally build one BINLOG statement for all events making up a statement. sql/log_event.h: Changing signature of event printing functions print_header() and print_base64() to write to an IO_CACHE instead of a file. Changing row-based replication events to incrementally build one BINLOG statement for all events making up a statement. Adding a head_cache and a body_cache to cache statement comment and statement body respectively. In addition, the head_cache is used when printing other events than the RBR events. sql/sql_binlog.cc: Changing code to be able to decode several pieces of base64-encoded data for a BINLOG statement. The BINLOG statement now consists of several pieces of BASE64-encoded data, so once a block has been decoded and executed, the next block has to be read from the statement until there is no more data to read.
-
- 02 Oct, 2006 1 commit
-
-
unknown authored
Non-upper-level INSERTs (the ones in the body of stored procedure, stored function, or trigger) into a table that have AUTO_INCREMENT column didn't affected the result of LAST_INSERT_ID() on this level. The problem was introduced with the fix of bug 6880, which in turn was introduced with the fix of bug 3117, where current insert_id value was remembered on the first call to LAST_INSERT_ID() (bug 3117) and was returned from that function until it was reset before the next _upper-level_ statement (bug 6880). The fix for bug#21726 brings back the behaviour of version 4.0, and implements the following: remember insert_id value at the beginning of the statement or expression (which at that point equals to the first insert_id value generated by the previous statement), and return that remembered value from LAST_INSERT_ID() or @@LAST_INSERT_ID. Thus, the value returned by LAST_INSERT_ID() is not affected by values generated by current statement, nor by LAST_INSERT_ID(expr) calls in this statement. Version 5.1 does not have this bug (it was fixed by WL 3146). mysql-test/r/rpl_insert_id.result: Add results for bug#21726: Incorrect result with multiple invocations of LAST_INSERT_ID, and bug#20339: stored procedure using LAST_INSERT_ID() does not replicate statement-based. mysql-test/t/rpl_insert_id.test: Add test cases for bug#21726: Incorrect result with multiple invocations of LAST_INSERT_ID, and bug#20339: stored procedure using LAST_INSERT_ID() does not replicate statement-based. sql/item_func.cc: Add implementation of Item_func_last_insert_id::fix_fields(), where we remember in THD::current_insert_id the first value generated during execution of the previous statement, which is returned then from Item_func_last_insert_id::val_int(). sql/item_func.h: Add declaration of Item_func_last_insert_id::fix_fields(). sql/log_event.cc: Do not set THD::last_insert_id_used on LAST_INSERT_ID_EVENT. Though we know the statement will call LAST_INSERT_ID(), it wasn't called yet. sql/set_var.cc: In sys_var_last_insert_id::value_ptr() remember in THD::current_insert_id the first value generated during execution of the previous statement, and return this value for @@LAST_INSERT_ID. sql/sql_class.cc: Reset THD::last_insert_id_used after each statement execution. sql/sql_class.h: Rather then remember current insert_id value on first invocation of THD::insert_id(), remember it in Item_func_last_insert_id::fix_fields(), sys_var_last_insert_id::value_ptr(), or mysql_execute_command(). Remove THD::insert_id(), as it lost its value now. sql/sql_insert.cc: THD::insert_id() is removed, use THD::last_insert_id directly. sql/sql_load.cc: THD::insert_id() is removed, using THD::last_insert_id directly is OK. sql/sql_parse.cc: Remember in THD::current_insert_id first generated insert id value of the previous statement in mysql_execute_command(). No need to reset THD::last_insert_id_used in mysql_reset_thd_for_next_command(), it will be reset after each statement. sql/sql_select.cc: If "IS NULL" is replaced with "= <LAST_INSERT_ID>", use right value, which is THD::current_insert_id, and also set THD::last_insert_id_used to issue binary log LAST_INSERT_ID_EVENT. sql/sql_update.cc: THD::insert_id() is removed, use THD::last_insert_id directly. tests/mysql_client_test.c: Add test case for bug#21726: Incorrect result with multiple invocations of LAST_INSERT_ID.
-
- 21 Sep, 2006 1 commit
-
-
unknown authored
Using move_field_offset() to move field pointer to other record before calling pack()/unpack(). sql/log_event.cc: Using move_field_offset() to move the field pointer. It is necessary for unpack() and pack() of BIT fields to operate correctly. Adding debug printouts. sql/sql_class.cc: Using move_field_offset() to move the field pointer. It is necessary for unpack() and pack() of BIT fields to operate correctly. Adding debug printouts to packing of update rows event.
-
- 13 Sep, 2006 4 commits
-
-
unknown authored
-
unknown authored
Incorporating changes from review. Fixing one bug that surfaced. mysql-test/extra/rpl_tests/rpl_row_tabledefs.test: Adding tests that UPDATE and DELETE does not generate an error. mysql-test/r/rpl_row_tabledefs_2myisam.result: Result change. mysql-test/r/rpl_row_tabledefs_3innodb.result: Result change. mysql-test/t/disabled.def: Enabling rpl_sp_effects (even though it gives a result mismatch currently). sql/field.cc: Using constant to denote undefined last null byte. sql/field.h: Using constant to denote undefined last null byte. Adding documentation. sql/log_event.cc: Not generating error for non-NULL no-DEFAULT columns when updating or deleting row. Better documentation and comments. sql/rpl_utility.cc: Moving documentation to header file. sql/rpl_utility.h: Documenting class and members.
-
unknown authored
-
unknown authored
Cset exclude: kaa@polly.local|ChangeSet|20060908100829|09983 sql/item_cmpfunc.cc: Exclude sql/item_cmpfunc.h: Exclude sql/item_func.cc: Exclude sql/item_func.h: Exclude sql/log_event.cc: Exclude sql/sql_class.h: Exclude mysql-test/r/case.result: Exclude mysql-test/r/func_if.result: Exclude mysql-test/r/func_test.result: Exclude mysql-test/r/user_var.result: Exclude mysql-test/t/case.test: Exclude mysql-test/t/func_if.test: Exclude mysql-test/t/func_test.test: Exclude mysql-test/t/user_var.test: Exclude
-
- 12 Sep, 2006 1 commit
-
-
unknown authored
mysql-test/r/case.result: Post-review fix for bug #20924 mysql-test/r/func_if.result: Post-review fix for bug #20924 mysql-test/r/func_test.result: Post-review fix for bug #20924 mysql-test/r/user_var.result: Post-review fix for bug #20924 mysql-test/t/case.test: Post-review fix for bug #20924 mysql-test/t/func_if.test: Post-review fix for bug #20924 mysql-test/t/func_test.test: Post-review fix for bug #20924 mysql-test/t/user_var.test: Post-review fix for bug #20924 sql/item_func.cc: Post-review fix for bug #20924 sql/item_func.h: Post-review fix for bug #20924 sql/log_event.cc: Post-review fix for bug #20924
-
- 08 Sep, 2006 1 commit
-
-
unknown authored
sql/log_event.cc: Adding cast since byte is unsigned char on Windows sql/rpl_utility.cc: Adding missing return statement.
-
- 06 Sep, 2006 1 commit
-
-
unknown authored
Invalidating query cache when processing rows for a statement on the slave. mysql-test/r/rpl_row_basic_11bugs.result: Result file change mysql-test/t/rpl_row_basic_11bugs.test: Adding test to trigger failure sql/log_event.cc: Adding code to invalidate the query cache just after opening the tables for processing the rows of one statement.
-
- 22 Aug, 2006 1 commit
-
-
unknown authored
Post-merge fixes. mysql-test/r/rpl_row_tabledefs_3innodb.result: Result change mysql-test/t/disabled.def: Disabling test that appear to be fixed in other clones sql/field.cc: Post-merge fix sql/log_event.cc: Changes to unpack_row(): - Documentation - New parameter, the write/read set - R/W set now 0-indexed instead of 1-indexed - Removed column value printing Changes to replace_record(): - Documentation - Removed column value printing Removed unused function print_column_values()
-
- 13 Jul, 2006 1 commit
-
-
unknown authored
mysql-test/r/federated.result: A post-merge fix. mysql-test/r/show_check.result: A post-merge fix. sql/log_event.cc: Remove rewrite_db (it's refactored in 5.1 to rpl_filter) sql/share/errmsg.txt: Move ER_NO_TRIGGERS_ON_SYSTEM_SCHEMA to the end of the list. This breaks compatibility of errmsg.sys with 5.0 but preserves compatibility with 5.1.11. sql/sql_class.h: A post-merge fix.
-
- 10 Jul, 2006 1 commit
-
-
unknown authored
We now reset the THD members related to auto_increment+binlog in MYSQL_LOG::write(). This is better than in THD::cleanup_after_query(), which was not able to distinguish between SELECT myfunc1(),myfunc2() and INSERT INTO t SELECT myfunc1(),myfunc2() from a binlogging point of view. Rows_log_event::exec_event() now calls lex_start() instead of mysql_init_query() because the latter now does too much (it resets the binlog format). mysql-test/extra/rpl_tests/rpl_insert_id.test: fix after merge mysql-test/mysql-test-run.pl: -v does not bring useful information when running valgrind; I remove it; if you think it's useful add it back. mysql-test/r/binlog_stm_mix_innodb_myisam.result: Position columns of SHOW BINLOG EVENTS are replaced by # (more robust if the size of an event changes). mysql-test/r/rpl_insert_id.result: fix after merge mysql-test/r/rpl_loaddata.result: The binlog positions change, because one event disappeared; indeed there was this in the binlog (in the current 5.1!): SET INSERT_ID=2; SET INSERT_ID=1; SET TIMESTAMP=1152540671; load data LOCAL INFILE '/tmp/SQL_LOAD_MB-1-2' INTO table t1; Two INSERT_ID events, useless and a bug. Goes away afer cleaning up auto_increment handling. mysql-test/r/rpl_switch_stm_row_mixed.result: INSERT_ID=5 appears, it's a consequence of having merged the fix for BUG#20341 "stored function inserting into one auto_increment puts bad data in slave". In mixed mode, if one substatement of a stored procedure requires row-based, the entire procedure uses row-based (was already true for stored functions); this is a consequence of not doing the resetting of binlog format inside lock_tables() (which didn't work with how the slave thread executes row-based binlog events). mysql-test/t/rpl_switch_stm_row_mixed.test: removing the multi-row delayed insert because in RBR the number of events which it generates, is not repeatable (probably depends on how the delayed thread groups rows, i.e. dependent on timing). sql/ha_partition.cc: update to new prototype sql/ha_partition.h: update to new prototype of the handler:: method. sql/handler.cc: after-merge fixes (manually merging part which was hard to merge in fmtool) sql/log.cc: When we write to the binary log, THD's parameters which influenced this write are reset: stmt_depends_on_first_successful_insert_id_in_prev_stmt and auto_inc_intervals_in_cur_stmt_for_binlog. This is so that future writes are not influenced by those and can write their own values. As a consequence, when we don't write to the binlog we do not reset. This is to abide by the rule that in a complex statement (using triggers etc), the first top- or substatement to generate auto_increment ids wins their writing to the binlog (that writing may be done by the statement itself or by the caller); so for example for INSERT INTO t SELECT myfunc() where myfunc() inserts into auto_increment and INSERT INTO t does not, myfunc() will fill auto_inc_intervals_in_cur_stmt_for_binlog, which will not be reset when myfunc() ends, then INSERT INTO t will write to the binlog and thus write the preserved auto_inc_intervals_in_cur_stmt_for_binlog. sql/log_event.cc: mysql_init_query() does too much now to be called in Rows_log_event::exec_event (it call mysql_reset_thd_for_next_command() which may switch the binlog format now). It's ok to call it in Table_map_log_event::exec_event() but its call must be before setting the binlog format to "row". sql/sql_base.cc: Resetting the binlog format in lock_tables() was a bad idea of mine; it causes problems in execution of row-based binlog events, where the thread sets the binlog format by itself and does not want a next lock_tables() to reset the binlog format. It is also misleading, for a function named lock_tables(), to reset the binlog format. As a consequence of this change, in mixed binlogging mode, a routine is logged either entirely statement-based or entirely row-based, we don't switch in the middle (this was already true for prelocked routines, now it's also true for stored procedures). sql/sql_class.cc: resetting of auto_increment variables used for binlogging is now done when writing to the binary log, no need to do the resetting at the end of the statement. It is also more correct this way; consider SELECT myfunc1(),myfunc2(); where both functions insert into the same auto_increment column. Binlogging is done in 2 events: "SELECT myfunc1()" and "SELECT myfunc2()". So each of those needs to have, in binlog, the INSERT_ID which it inserted. But as the 2 function calls are executed under prelocked mode, the old code didn't reset auto_inc_intervals_in_cur_stmt_for_binlog after the first SELECT was binlogged, and so the INSERT_ID of the first SELECT was binlogged for the first SELECT and (wrong) also for the 2nd SELECT event. stmt_depends_on_first_... has the same logic. sql/sql_class.h: clearer comment sql/sql_delete.cc: unneeded #ifdef. As we temporarily change the binlog format to "statement" before calling mysql_delete(), we must restore it afterwards. sql/sql_insert.cc: after-merge fixes. No need to reset auto_inc_intervals_in_cur_stmt_for_binlog for every row in the delayed insert system thread, because we already reset it when writing to the binlog. sql/sql_parse.cc: unneeded #ifdef
-
- 09 Jul, 2006 1 commit
-
-
unknown authored
this is a cleanup patch for our current auto_increment handling: new names for auto_increment variables in THD, new methods to manipulate them (see sql_class.h), some move into handler::, causing less backup/restore work when executing substatements. This makes the logic hopefully clearer, less work is is needed in mysql_insert(). By cleaning up, using different variables for different purposes (instead of one for 3 things...), we fix those bugs, which someone may want to fix in 5.0 too: BUG#20339 "stored procedure using LAST_INSERT_ID() does not replicate statement-based" BUG#20341 "stored function inserting into one auto_increment puts bad data in slave" BUG#19243 "wrong LAST_INSERT_ID() after ON DUPLICATE KEY UPDATE" (now if a row is updated, LAST_INSERT_ID() will return its id) and re-fixes: BUG#6880 "LAST_INSERT_ID() value changes during multi-row INSERT" (already fixed differently by Ramil in 4.1) Test of documented behaviour of mysql_insert_id() (there was no test). The behaviour changes introduced are: - LAST_INSERT_ID() now returns "the first autogenerated auto_increment value successfully inserted", instead of "the first autogenerated auto_increment value if any row was successfully inserted", see auto_increment.test. Same for mysql_insert_id(), see mysql_client_test.c. - LAST_INSERT_ID() returns the id of the updated row if ON DUPLICATE KEY UPDATE, see auto_increment.test. Same for mysql_insert_id(), see mysql_client_test.c. - LAST_INSERT_ID() does not change if no autogenerated value was successfully inserted (it used to then be 0), see auto_increment.test. - if in INSERT SELECT no autogenerated value was successfully inserted, mysql_insert_id() now returns the id of the last inserted row (it already did this for INSERT VALUES), see mysql_client_test.c. - if INSERT SELECT uses LAST_INSERT_ID(X), mysql_insert_id() now returns X (it already did this for INSERT VALUES), see mysql_client_test.c. - NDB now behaves like other engines wrt SET INSERT_ID: with INSERT IGNORE, the id passed in SET INSERT_ID is re-used until a row succeeds; SET INSERT_ID influences not only the first row now. Additionally, when unlocking a table we check that the thread is not keeping a next_insert_id (as the table is unlocked that id is potentially out-of-date); forgetting about this next_insert_id is done in a new handler::ha_release_auto_increment(). Finally we prepare for engines capable of reserving finite-length intervals of auto_increment values: we store such intervals in THD. The next step (to be done by the replication team in 5.1) is to read those intervals from THD and actually store them in the statement-based binary log. NDB will be a good engine to test that. mysql-test/extra/binlog_tests/binlog.test: Testing that if INSERT_ID is set to a value too big for the column's type, the binlogged INSERT_ID is the truncated value (important if slave has a column of a "wider" numeric type). Testing binlogging of INSERT_ID with INSERT DELAYED, to be sure that we binlog an INSERT_ID event only for the delayed rows which use one. mysql-test/extra/rpl_tests/rpl_insert_id.test: Testcase for BUG#20339 "stored procedure using LAST_INSERT_ID() does not replicate statement-based". Testcase for BUG#20341 "stored function inserting into one auto_increment puts bad data in slave". mysql-test/extra/rpl_tests/rpl_loaddata.test: Test that LOAD DATA INFILE sets a value for a future LAST_INSERT_ID(). mysql-test/r/auto_increment.result: behaviour change: when INSERT totally fails (not even succeeds partially and then rolls back), don't change last_insert_id(). Behaviour change: LAST_INSERT_ID() is now the first successfully inserted, autogenerated, id. Behaviour change: if INSERT ON DUPLICATE KEY UPDATE, if the table has auto_increment and a row is updated, then LAST_INSERT_ID() returns the id of this row. mysql-test/r/binlog_row_binlog.result: result update mysql-test/r/binlog_stm_binlog.result: result update mysql-test/r/insert.result: result update mysql-test/r/rpl_insert_id.result: result update mysql-test/r/rpl_loaddata.result: result update mysql-test/r/rpl_ndb_auto_inc.result: ndb's behaviour is now like other engines wrt SET INSERT_ID in a multi-row INSERT: - with INSERT IGNORE: the id passed in SET INSERT_ID is re-used until a row succeeds. - generally, SET INSERT_ID sets the first value and other values are simply computed from this first value, instead of previously where the 2nd and subsequent values where not influenced by SET INSERT_ID; this good change is due to the removal of "thd->next_insert_id=0" from ha_ndbcluster. mysql-test/t/auto_increment.test: A testcase of BUG#19243: if ON DUPLICATE KEY UPDATE updates a row, LAST_INSERT_ID() now returns the id of the row. Test of new behaviour of last_insert_id() when no autogenerated value was inserted, or when only some autogenerated value (not the first of them) was inserted. mysql-test/t/insert.test: testing INSERT IGNORE re-using generated values sql/ha_federated.cc: update for new variables. sql/ha_ndbcluster.cc: handler::auto_increment_column_changed not needed, equivalent to (insert_id_for_cur_row > 0). thd->next_insert_id=0 not needed anymore; it was used to force handler::update_auto_increment() to call ha_ndbcluster::get_auto_increment() for each row of a multi-row INSERT, now this happens naturally because NDB says "I have reserved you *one* value" in get_auto_increment(), so handler::update_auto_increment() calls again for next row. sql/handler.cc: More comments, use of new methods and variables. Hopes to be clearer than current code. thd->prev_insert_id not in THD anymore: it is managed locally by inserters (like mysql_insert()). THD::clear_next_insert_id is now equivalent to handler::next_insert_id > 0. get_auto_increment() reserves an interval of values from the engine, uses this interval for next rows of the statement, until interval is exhausted then it asks for another interval (of a bigger size than the first one; size doubles until reaching 65535 then it stays constant). If doing statement-based binlogging, intervals are remembered in a list for storage in the binlog. For "forced" insert_id values (SET INSERT_ID or replication slave), forced_auto_inc_intervals is non-empty and the handler takes its intervals from there, without calling get_auto_increment(). ha_release_auto_increment() resets the handler's auto_increment variables; it calls release_auto_increment() which is handler-dependent and serves to return to the engine any unused tail of the last used interval. If ending a statement, next_insert_id>0 means that autoinc values have been generated or taken from the master's binlog (in a replication slave) so we clear those values read from binlog, so that next top- or sub- statement does not use them. sql/handler.h: handler::auto_increment_changed can be replaced by (handler::insert_id_for_cur_row > 0). THD::next_insert_id moves into handler (more natural, and prepares for the day when we'll support a single statement inserting into two tables - "multi-table INSERT" like we have UPDATE - will this happen?). This move makes the backup/restore of THD::next_insert_id when entering a substatement unneeded, as each substatement has its own handler objects. sql/item_func.cc: new names for variables. For the setting of what mysql_insert_id() will return to the client, LAST_INSERT_ID(X) used to simply pretend that the generated autoinc value for the current row was X, but this led to having no reliable way to know the really generated value, so we now have a bool: thd->arg_of_last_insert_id_function which enables us to know that LAST_INSERT_ID(X) was called (and then X can be found in thd->first_successful_insert_id_in_prev_stmt). sql/log.cc: new variable names for insert_ids. Removing some unused variables in the slow log. sql/log_event.cc: new variable names, comments. Preparing for when master's won't binlog LAST_INSERT_ID if it was 0. sql/set_var.cc: new variable names. The last change repeats how Bar fixed BUG#20392 "INSERT_ID session variable has weird value" in 5.0. sql/sql_class.cc: new variables for insert_id. In THD::cleanup_after_query() we fix BUG#20339 "stored procedure using LAST_INSERT_ID() does not replicate statement-based" (will one want to fix it in 5.0?). Many comments about what stored functions do to auto_increment. In reset|restore_sub_statement_state(), we need to backup less auto_inc variables as some of them have moved to the handler; we backup/restore those which are about the current top- or sub- statement, *not* those about the statement-based binlog (which evolve as the top- and sub-statement execute). Because we split THD::last_insert_id into THD::first_successful_insert_id_in_prev_stmt and THD::auto_inc_intervals_for_binlog (among others), we fix BUG#20341 "stored function inserting into one auto_increment puts bad data in slave": indeed we can afford to not backup/restore THD::auto_inc_intervals_for_binlog (which fixes the bug) while still backing up / restoring THD::first_successful_insert_id_in_prev_stmt (ensuring that the top-level LAST_INSERT_ID() is not affected by INSERTs done by sub-statements, as is desirable and tested in rpl_insert_id.test). sql/sql_class.h: new variables and methods for auto_increment. Some THD members move into handler (those which are really about the table being inserted), some stay in THD (those which are about what a future LAST_INSERT_ID() should return, or about what should be stored into the statement-based binlog). THD::next_insert_id moves to handler::. THD::clear_next_insert_id removed (had become equivalent to next_insert_id > 0). THD::last_insert_id becomes four: THD::first_successful_insert_id_in_cur_stmt, THD::auto_inc_intervals_for_binlog, handler::insert_id_for_cur_row, THD::first_successful_insert_id_in_prev_stmt. THD::current_insert_id becomes: THD::first_successful_insert_id_in_prev_stmt_for_binlog THD::prev_insert_id is removed, handler can just use handler::insert_id_for_cur_row instead (which is more accurate: for the first row, prev_insert_id was set before get_auto_increment was called, so was 0, causing a call to get_auto_increment() for the 2nd row if the 1st row fails; here we don't need the call as insert_id_for_cur_row has the value of the first row). THD::last_insert_id_used becomes: stmt_depends_on_first_row_in_prev_stmt THD::insert_id_used is removed (equivalent to auto_inc_intervals_for_binlog non empty). The interval returned by get_auto_increment() and currently being consumed is handler::auto_inc_interval_for_cur_row. Comments to explain each of them. select_insert::last_insert_id becomes autoinc_value_of_last_inserted_row. sql/sql_insert.cc: the "id" variable is not changed for each row now; it used to compensate for this contradiction: - thd->last_insert_id supposed job was to keep the id of the first row - but it was updated for every row - so mysql_insert() made sure to catch its first value and restore it at the end of stmt. Now THD keeps the first value in first_successful_insert_id_in_cur_stmt, and value of the row in insert_id_for_cur_row. So "id" only serves to fill mysql_insert_id(), as depending on some conditions, "id" must be different values. Prev_insert_id moves from THD to write_record(). We now set LAST_INSERT_ID() in ON DUPLICATE KEY UPDATE too (BUG#19243). In an INSERT DELAYED, we still "reset auto-increment caching" but differently (by calling ha_release_auto_increment()). sql/sql_load.cc: no need to fiddle with "id", THD maintains THD::first_successful_insert_id_in_cur_stmt by itself and correctly now. ha_release_auto_increment() is now (logically) called before we unlock the table. sql/sql_parse.cc: update to new variable names. Assertion that reset_thd_for_next_command() is not called for every substatement of a routine (I'm not against it, but if we do this change, statement-based binlogging needs some adjustments). sql/sql_select.cc: update for new variable names sql/sql_table.cc: next_insert_id not needed in mysql_alter_table(), THD manages. sql/sql_update.cc: update for new variable names. Even though this is UPDATE, an insert id can be generated (by LAST_INSERT_ID(X)) and should be recorded because mysql_insert_id() wants to know about it. sql/structs.h: A class for "discrete" intervals (intervals of integer numbers with a certain increment between them): Discrete_interval, and a class for a list of such intervals: Discrete_intervals_list tests/mysql_client_test.c: tests of behaviour of mysql_insert_id(): there were no such tests, while in our manual we document its behaviour. In comments you'll notice the behaviour changes introduced (there are 5).
-
- 07 Jul, 2006 2 commits
- 06 Jul, 2006 1 commit
-
-
unknown authored
before update trigger on NDB table". Two main changes: - We use TABLE::read_set/write_set bitmaps for marking fields used by statement instead of Field::query_id in 5.1. - Now when we mark columns used by statement we take into account columns used by table's triggers instead of marking all columns as used if table has triggers. mysql-test/r/federated.result: Changed test in order to make it work with RBR. RBR changes the way in which we execute "DELETE FROM t1" statement - we don't use handler::delete_all_rows() method if RBR is enabled (see bug#19066). As result federated engine produces different sequences of statements for remote server in non-RBR and in RBR cases. And this changes order of the rows inserted by following INSERT statements. mysql-test/t/federated.test: Changed test in order to make it work with RBR. RBR changes the way in which we execute "DELETE FROM t1" statement - we don't use handler::delete_all_rows() method if RBR is enabled (see bug#19066). As result federated engine produces different sequences of statements for remote server in non-RBR and in RBR cases. And this changes order of the rows inserted by following INSERT statements. sql/ha_partition.cc: Added handling of HA_EXTRA_WRITE_CAN_REPLACE/HA_EXTRA_WRITE_CANNOT_REPLACE to ha_partition::extra(). sql/item.cc: Adjusted comment after merge. In 5.1 we use TABLE::read_set/write_set bitmaps instead of Field::query_id for marking columns used. sql/log_event.cc: Write_rows_log_event::do_before_row_operations(): Now we explicitly inform handler that we want to replace rows so it can promote operation done by write_row() to replace. sql/mysql_priv.h: Removed declaration of mark_fields_used_by_triggers_for_insert_stmt() which is no longer used (we have TABLE::mark_columns_needed_for_insert() instead). sql/sql_insert.cc: Adjusted code after merge. Get rid of mark_fields_used_by_triggers_for_insert_stmt() as now we use TABLE::mark_columns_needed_for_insert() for the same purprose. Aligned places where we call this method with places where we call mark_fields_used_by_triggers_for_insert() in 5.0. Finally we no longer need to call handler::extra(HA_EXTRA_WRITE_CAN_REPLACE) in case of REPLACE statement since in 5.1 write_record() marks all columns as used before doing actual row replacement. sql/sql_load.cc: Adjusted code after merge. In 5.1 we use TABLE::mark_columns_needed_for_insert() instead of mark_fields_used_by_triggers_for_insert_stmt() routine. We also no longer need to call handler::extra(HA_EXTRA_RETRIEVE_ALL_COLS) if we execute LOAD DATA REPLACE since in 5.1 write_record() will mark all columns as used before doing actual row replacement. sql/sql_trigger.cc: Table_triggers_list::mark_fields_used(): We use TABLE::read_set/write_set bitmaps for marking fields used instead of Field::query_id in 5.1. sql/sql_trigger.h: TABLE::mark_columns_needed_for_* methods no longer need to be friends of Table_triggers_list class as intead of dirrectly accessing its private members they can use public Table_triggers_list::mark_fields_used() method. Also Table_triggers)list::mark_fields_used() no longer needs THD argument. sql/table.cc: TABLE::mark_columns_needed_for_*(): Now we mark columns which are really used by table's triggers instead of marking all columns as used if table has triggers.
-
- 28 Jun, 2006 2 commits
-
-
unknown authored
No test case as the bug is in an existing test case (rpl_trigger.test when it is run under valgrind). The warning was caused by memory corruption in replication slave: thd->db was pointing at a stack address that was previously used by sp_head::execute()::old_db. This happened because mysql_change_db behaved differently in replication slave and did not make a copy of the argument to assign to thd->db. The solution is to always free the old value of thd->db and allocate a new copy, regardless whether we're running in a replication slave or not. sql/log_event.cc: Move rewrite_db to log_event.cc, the only place where it is used. sql/slave.cc: Move rewrite_db to log_event.cc sql/slave.h: Remove an unneeded declaration. sql/sql_class.h: Fix set_db to always free the old db, even if the argument is NULL. Add a comment. sql/sql_db.cc: Always make a deep copy of the argument in mysql_change_db, even if running in a replication slave. This is necessary because sp_use_new_db (stored procedures) assumes that mysql_change_db always makes a deep copy of the argument, and thus passes a pointer to stack into it. This assumption was true for all cases except the replication slave thread.
-
unknown authored
(Found by dgcov) mysql-test/r/show_check.result: Test SHOW PLUGIN. (Was not covered by test suite before) mysql-test/t/show_check.test: Test SHOW PLUGIN. (Was not covered by test suite before) mysys/my_handler.c: Added purecov comment for new, not executed code sql/filesort.cc: Added purecov comment for new, not executed code Removed 'purecov: inspected' from code covered by current testsuite sql/ha_ndbcluster_binlog.cc: Added purecov comment for new, not executed code sql/log_event.cc: Added purecov comment for new, not executed code sql/mysqld.cc: Added purecov comment for new, not executed code sql/sql_parse.cc: Added purecov comment for new, not executed code
-
- 27 Jun, 2006 1 commit
-
-
unknown authored
Fix a minor issue with Bug#16206 (bdb.test failed if the tree is compiled without blackhole). include/my_sys.h: Change declaration of my_strdup_with_length to accept const char *, not const byte *: in 5 places out of 6 where this function is used, it's being passed char *, not byte * mysql-test/r/bdb.result: Remove dependency on an optional engine (updated test results). mysql-test/t/bdb.test: Remove dependency on an optional engine. mysys/my_malloc.c: my_strdup_with_length: const byte * -> const char * mysys/safemalloc.c: my_strdup_with_length: const byte * -> const char * sql/ha_federated.cc: my_strdup_with_length: const byte * -> const char * sql/log_event.cc: my_strdup_with_length: const byte * -> const char * sql/set_var.cc: my_strdup_with_length: const byte * -> const char * sql/sql_class.h: Change db_length type to uint from uint32 (see also table.h) sql/table.h: Change the type of db_length to uint from uint32: LEX_STRING uses uint for length, we need a small and consistent set of types to store length to minimize cast and compile failures.
-
- 26 Jun, 2006 1 commit
-
-
unknown authored
Bug#19022 "Memory bug when switching db during trigger execution" Bug#17199 "Problem when view calls function from another database." Bug#18444 "Fully qualified stored function names don't work correctly in SELECT statements" Documentation note: this patch introduces a change in behaviour of prepared statements. This patch adds a few new invariants with regard to how THD::db should be used. These invariants should be preserved in future: - one should never refer to THD::db by pointer and always make a deep copy (strmake, strdup) - one should never compare two databases by pointer, but use strncmp or my_strncasecmp - TABLE_LIST object table->db should be always initialized in the parser or by creator of the object. For prepared statements it means that if the current database is changed after a statement is prepared, the database that was current at prepare remains active. This also means that you can not prepare a statement that implicitly refers to the current database if the latter is not set. This is not documented, and therefore needs documentation. This is NOT a change in behavior for almost all SQL statements except: - ALTER TABLE t1 RENAME t2 - OPTIMIZE TABLE t1 - ANALYZE TABLE t1 - TRUNCATE TABLE t1 -- until this patch t1 or t2 could be evaluated at the first execution of prepared statement. CURRENT_DATABASE() still works OK and is evaluated at every execution of prepared statement. Note, that in stored routines this is not an issue as the default database is the database of the stored procedure and "use" statement is prohibited in stored routines. This patch makes obsolete the use of check_db_used (it was never used in the old code too) and all other places that check for table->db and assign it from THD::db if it's NULL, except the parser. How this patch was created: THD::{db,db_length} were replaced with a LEX_STRING, THD::db. All the places that refer to THD::{db,db_length} were manually checked and: - if the place uses thd->db by pointer, it was fixed to make a deep copy - if a place compared two db pointers, it was fixed to compare them by value (via strcmp/my_strcasecmp, whatever was approproate) Then this intermediate patch was used to write a smaller patch that does the same thing but without a rename. TODO in 5.1: - remove check_db_used - deploy THD::set_db in mysql_change_db See also comments to individual files. mysql-test/r/create.result: Modify the result file: a database can never be NULL. mysql-test/r/ps.result: Update test results (Bug#17199 et al) mysql-test/r/sp.result: Update test results (Bug#17199 et al) mysql-test/t/create.test: Update the id of the returned error. mysql-test/t/ps.test: Add test coverage for prepared statements and current database. In scope of work on Bug#17199 "Problem when view calls function from another database." mysql-test/t/sp.test: Add a test case for Bug#17199 "Problem when view calls function from another database." and Bug#18444 "Fully qualified stored function names don't work correctly in SELECT statements". Test a complementary problem. sql/item_strfunc.cc: Touch the code that reads thd->db (cleanup). sql/log_event.cc: While we are at it, replace direct access to thd->db with a method. Should simplify future conversion of THD::db to LEX_STRING. sql/slave.cc: While we are at it, replace direct access to thd->db with a method. Should simplify future conversion of THD::db to LEX_STRING. sql/slave.h: Remove a declaration for a method that is used only in one module. sql/sp.cc: Rewrite sp_use_new_db: this is a cleanup that I needed in order to understand this function and ensure that it has no bugs. sql/sp.h: Add a new declaration for sp_use_new_db (uses LEX_STRINGs) and a comment. sql/sp_head.cc: - drop sp_name_current_db_new - a creator of sp_name class that was used when sp_name was created for an identifier without an explicitly initialized database. Now we pass thd->db to constructor of sp_name right in the parser. - rewrite sp_head::init_strings: name->m_db is always set now - use the new variant of sp_use_new_db - we don't need to update thd->db with SP MEM_ROOT pointer anymore when parsing a stored procedure, as noone will refer to it (yes!) sql/sp_head.h: - remove unneded methods and members sql/sql_class.h: - introduce 3 THD methods to work with THD::db: .set_db to assign the current database .reset_db to reset the current database (temporarily) or set it to NULL .opt_copy_db_to - to deep-copy thd->db to a pointer if it's not NULL sql/sql_db.cc: While we are at it, replace direct access to thd->db with a method. Should simplify future conversion of THD::db to LEX_STRING. sql/sql_insert.cc: - replace checks with asserts: table_list->db must be always set in the parser. sql/sql_lex.h: - add a comment sql/sql_parse.cc: - implement the invariant described in the changeset comment. - remove juggling with lex->sphead in SQLCOM_CREATE_PROCEDURE: now db_load_routine uses its own LEX object and doesn't damage the main LEX. - add DBUG_ASSERT(0) to unused "check_db_used" sql/sql_table.cc: - replace a check with an assert (table_ident->db) sql/sql_trigger.cc: While we are at it, replace direct access to thd->db with a method. Should simplify future conversion of THD::db to LEX_STRING. sql/sql_udf.cc: - use thd->set_db instead of direct modification of to thd->db sql/sql_view.cc: - replace a check with an assert (view->db) sql/sql_yacc.yy: - make sure that we always copy table->db or name->db or ident->db or select_lex->db from thd->db if the former is not set. If thd->db is not set but is accessed, return an error. sql/tztime.cc: - be nice, never copy thd->db by pointer.
-
- 23 Jun, 2006 1 commit
-
-
unknown authored
sql/handler.cc: Made code more portable
-
- 19 Jun, 2006 1 commit
-
-
unknown authored
Added missing copyright Indentation and prototype fixes Fixed compiler warning include/my_sys.h: Fix prototype for my_print_open_files. Indentation fixes mysys/mf_keycache.c: Fix that key cache can be compiled without #define THREAD mysys/my_error.c: Indentation cleanups mysys/my_memmem.c: Added missing copyright. Fix include files sql/event.cc: unsigned char -> uchar sql/field.cc: unsigned char -> uchar sql/filesort.cc: Indentation fix sql/ha_ndbcluster.cc: unsigned char -> uchar sql/ha_ndbcluster_binlog.cc: unsigned char -> uchar sql/item_strfunc.cc: unsigned char -> uchar sql/log_event.cc: unsigned char -> uchar sql/opt_range.cc: Indentation fixes sql/sql_crypt.cc: unsigned char -> uchar sql/sql_string.cc: unsigned char -> uchar sql/tztime.cc: unsigned char -> uchar storage/heap/hp_extra.c: Clear also next_block (See heap_scan_init) storage/heap/hp_test2.c: Set required struct memeber max_table_size (Fixes that hp_test2 works again) storage/myisam/mi_rsamepos.c: Fixed bug in mi_rsame_with_pos strings/decimal.c: Fixed compiler warning
-
- 12 Jun, 2006 1 commit
-
-
unknown authored
mode This is a modification of serg's and guilhem's suggestion in the bug report, in that it also causes the transaction log to be written to disc. mysql-test/r/bdb.result: Add result. mysql-test/t/bdb.test: Add test sql/log.cc: Create a log-committing event that itself won't be written to the log when we're in autocommit mode. sql/log_event.cc: Add a new subclass of Query_log_event that doesn't write itself to the log, for cases where we only want to flush out the transaction and not also write about this event. sql/log_event.h: Add a new subclass of Query_log_event that doesn't write itself to the log, for cases where we only want to flush out the transaction and not also write about this event.
-
- 05 Jun, 2006 1 commit
-
-
unknown authored
The bug was caused by ignoring failure when opening a a table in Table_map_log_event::exec_event(). mysql-test/r/rpl_row_inexist_tbl.result: Fixed test case result. mysql-test/t/disabled.def: Enabled rpl_row_inexist_table test. mysql-test/t/rpl_row_inexist_tbl.test: Extended test case with "replicate-ignore-table" t2. sql/log_event.cc: Table_map_log_event::exec_event(): Remove code which ignores opening table failure. Previously, this code was needed because of extraneous table maps in the binary log. Currently, for any table that should be replicated to slave, a filter is needed (bug#18948).
-
- 04 Jun, 2006 1 commit
-
-
unknown authored
This changeset is largely a handler cleanup changeset (WL#3281), but includes fixes and cleanups that was found necessary while testing the handler changes Changes that requires code changes in other code of other storage engines. (Note that all changes are very straightforward and one should find all issues by compiling a --debug build and fixing all compiler errors and all asserts in field.cc while running the test suite), - New optional handler function introduced: reset() This is called after every DML statement to make it easy for a handler to statement specific cleanups. (The only case it's not called is if force the file to be closed) - handler::extra(HA_EXTRA_RESET) is removed. Code that was there before should be moved to handler::reset() - table->read_set contains a bitmap over all columns that are needed in the query. read_row() and similar functions only needs to read these columns - table->write_set contains a bitmap over all columns that will be updated in the query. write_row() and update_row() only needs to update these columns. The above bitmaps should now be up to date in all context (including ALTER TABLE, filesort()). The handler is informed of any changes to the bitmap after fix_fields() by calling the virtual function handler::column_bitmaps_signal(). If the handler does caching of these bitmaps (instead of using table->read_set, table->write_set), it should redo the caching in this code. as the signal() may be sent several times, it's probably best to set a variable in the signal and redo the caching on read_row() / write_row() if the variable was set. - Removed the read_set and write_set bitmap objects from the handler class - Removed all column bit handling functions from the handler class. (Now one instead uses the normal bitmap functions in my_bitmap.c instead of handler dedicated bitmap functions) - field->query_id is removed. One should instead instead check table->read_set and table->write_set if a field is used in the query. - handler::extra(HA_EXTRA_RETRIVE_ALL_COLS) and handler::extra(HA_EXTRA_RETRIEVE_PRIMARY_KEY) are removed. One should now instead use table->read_set to check for which columns to retrieve. - If a handler needs to call Field->val() or Field->store() on columns that are not used in the query, one should install a temporary all-columns-used map while doing so. For this, we provide the following functions: my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->read_set); field->val(); dbug_tmp_restore_column_map(table->read_set, old_map); and similar for the write map: my_bitmap_map *old_map= dbug_tmp_use_all_columns(table, table->write_set); field->val(); dbug_tmp_restore_column_map(table->write_set, old_map); If this is not done, you will sooner or later hit a DBUG_ASSERT in the field store() / val() functions. (For not DBUG binaries, the dbug_tmp_restore_column_map() and dbug_tmp_restore_column_map() are inline dummy functions and should be optimized away be the compiler). - If one needs to temporary set the column map for all binaries (and not just to avoid the DBUG_ASSERT() in the Field::store() / Field::val() methods) one should use the functions tmp_use_all_columns() and tmp_restore_column_map() instead of the above dbug_ variants. - All 'status' fields in the handler base class (like records, data_file_length etc) are now stored in a 'stats' struct. This makes it easier to know what status variables are provided by the base handler. This requires some trivial variable names in the extra() function. - New virtual function handler::records(). This is called to optimize COUNT(*) if (handler::table_flags() & HA_HAS_RECORDS()) is true. (stats.records is not supposed to be an exact value. It's only has to be 'reasonable enough' for the optimizer to be able to choose a good optimization path). - Non virtual handler::init() function added for caching of virtual constants from engine. - Removed has_transactions() virtual method. Now one should instead return HA_NO_TRANSACTIONS in table_flags() if the table handler DOES NOT support transactions. - The 'xxxx_create_handler()' function now has a MEM_ROOT_root argument that is to be used with 'new handler_name()' to allocate the handler in the right area. The xxxx_create_handler() function is also responsible for any initialization of the object before returning. For example, one should change: static handler *myisam_create_handler(TABLE_SHARE *table) { return new ha_myisam(table); } -> static handler *myisam_create_handler(TABLE_SHARE *table, MEM_ROOT *mem_root) { return new (mem_root) ha_myisam(table); } - New optional virtual function: use_hidden_primary_key(). This is called in case of an update/delete when (table_flags() and HA_PRIMARY_KEY_REQUIRED_FOR_DELETE) is defined but we don't have a primary key. This allows the handler to take precisions in remembering any hidden primary key to able to update/delete any found row. The default handler marks all columns to be read. - handler::table_flags() now returns a ulonglong (to allow for more flags). - New/changed table_flags() - HA_HAS_RECORDS Set if ::records() is supported - HA_NO_TRANSACTIONS Set if engine doesn't support transactions - HA_PRIMARY_KEY_REQUIRED_FOR_DELETE Set if we should mark all primary key columns for read when reading rows as part of a DELETE statement. If there is no primary key, all columns are marked for read. - HA_PARTIAL_COLUMN_READ Set if engine will not read all columns in some cases (based on table->read_set) - HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS Renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION. - HA_DUPP_POS Renamed to HA_DUPLICATE_POS - HA_REQUIRES_KEY_COLUMNS_FOR_DELETE Set this if we should mark ALL key columns for read when when reading rows as part of a DELETE statement. In case of an update we will mark all keys for read for which key part changed value. - HA_STATS_RECORDS_IS_EXACT Set this if stats.records is exact. (This saves us some extra records() calls when optimizing COUNT(*)) - Removed table_flags() - HA_NOT_EXACT_COUNT Now one should instead use HA_HAS_RECORDS if handler::records() gives an exact count() and HA_STATS_RECORDS_IS_EXACT if stats.records is exact. - HA_READ_RND_SAME Removed (no one supported this one) - Removed not needed functions ha_retrieve_all_cols() and ha_retrieve_all_pk() - Renamed handler::dupp_pos to handler::dup_pos - Removed not used variable handler::sortkey Upper level handler changes: - ha_reset() now does some overall checks and calls ::reset() - ha_table_flags() added. This is a cached version of table_flags(). The cache is updated on engine creation time and updated on open. MySQL level changes (not obvious from the above): - DBUG_ASSERT() added to check that column usage matches what is set in the column usage bit maps. (This found a LOT of bugs in current column marking code). - In 5.1 before, all used columns was marked in read_set and only updated columns was marked in write_set. Now we only mark columns for which we need a value in read_set. - Column bitmaps are created in open_binary_frm() and open_table_from_share(). (Before this was in table.cc) - handler::table_flags() calls are replaced with handler::ha_table_flags() - For calling field->val() you must have the corresponding bit set in table->read_set. For calling field->store() you must have the corresponding bit set in table->write_set. (There are asserts in all store()/val() functions to catch wrong usage) - thd->set_query_id is renamed to thd->mark_used_columns and instead of setting this to an integer value, this has now the values: MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE Changed also all variables named 'set_query_id' to mark_used_columns. - In filesort() we now inform the handler of exactly which columns are needed doing the sort and choosing the rows. - The TABLE_SHARE object has a 'all_set' column bitmap one can use when one needs a column bitmap with all columns set. (This is used for table->use_all_columns() and other places) - The TABLE object has 3 column bitmaps: - def_read_set Default bitmap for columns to be read - def_write_set Default bitmap for columns to be written - tmp_set Can be used as a temporary bitmap when needed. The table object has also two pointer to bitmaps read_set and write_set that the handler should use to find out which columns are used in which way. - count() optimization now calls handler::records() instead of using handler->stats.records (if (table_flags() & HA_HAS_RECORDS) is true). - Added extra argument to Item::walk() to indicate if we should also traverse sub queries. - Added TABLE parameter to cp_buffer_from_ref() - Don't close tables created with CREATE ... SELECT but keep them in the table cache. (Faster usage of newly created tables). New interfaces: - table->clear_column_bitmaps() to initialize the bitmaps for tables at start of new statements. - table->column_bitmaps_set() to set up new column bitmaps and signal the handler about this. - table->column_bitmaps_set_no_signal() for some few cases where we need to setup new column bitmaps but don't signal the handler (as the handler has already been signaled about these before). Used for the momement only in opt_range.cc when doing ROR scans. - table->use_all_columns() to install a bitmap where all columns are marked as use in the read and the write set. - table->default_column_bitmaps() to install the normal read and write column bitmaps, but not signaling the handler about this. This is mainly used when creating TABLE instances. - table->mark_columns_needed_for_delete(), table->mark_columns_needed_for_delete() and table->mark_columns_needed_for_insert() to allow us to put additional columns in column usage maps if handler so requires. (The handler indicates what it neads in handler->table_flags()) - table->prepare_for_position() to allow us to tell handler that it needs to read primary key parts to be able to store them in future table->position() calls. (This replaces the table->file->ha_retrieve_all_pk function) - table->mark_auto_increment_column() to tell handler are going to update columns part of any auto_increment key. - table->mark_columns_used_by_index() to mark all columns that is part of an index. It will also send extra(HA_EXTRA_KEYREAD) to handler to allow it to quickly know that it only needs to read colums that are part of the key. (The handler can also use the column map for detecting this, but simpler/faster handler can just monitor the extra() call). - table->mark_columns_used_by_index_no_reset() to in addition to other columns, also mark all columns that is used by the given key. - table->restore_column_maps_after_mark_index() to restore to default column maps after a call to table->mark_columns_used_by_index(). - New item function register_field_in_read_map(), for marking used columns in table->read_map. Used by filesort() to mark all used columns - Maintain in TABLE->merge_keys set of all keys that are used in query. (Simplices some optimization loops) - Maintain Field->part_of_key_not_clustered which is like Field->part_of_key but the field in the clustered key is not assumed to be part of all index. (used in opt_range.cc for faster loops) - dbug_tmp_use_all_columns(), dbug_tmp_restore_column_map() tmp_use_all_columns() and tmp_restore_column_map() functions to temporally mark all columns as usable. The 'dbug_' version is primarily intended inside a handler when it wants to just call Field:store() & Field::val() functions, but don't need the column maps set for any other usage. (ie:: bitmap_is_set() is never called) - We can't use compare_records() to skip updates for handlers that returns a partial column set and the read_set doesn't cover all columns in the write set. The reason for this is that if we have a column marked only for write we can't in the MySQL level know if the value changed or not. The reason this worked before was that MySQL marked all to be written columns as also to be read. The new 'optimal' bitmaps exposed this 'hidden bug'. - open_table_from_share() does not anymore setup temporary MEM_ROOT object as a thread specific variable for the handler. Instead we send the to-be-used MEMROOT to get_new_handler(). (Simpler, faster code) Bugs fixed: - Column marking was not done correctly in a lot of cases. (ALTER TABLE, when using triggers, auto_increment fields etc) (Could potentially result in wrong values inserted in table handlers relying on that the old column maps or field->set_query_id was correct) Especially when it comes to triggers, there may be cases where the old code would cause lost/wrong values for NDB and/or InnoDB tables. - Split thd->options flag OPTION_STATUS_NO_TRANS_UPDATE to two flags: OPTION_STATUS_NO_TRANS_UPDATE and OPTION_KEEP_LOG. This allowed me to remove some wrong warnings about: "Some non-transactional changed tables couldn't be rolled back" - Fixed handling of INSERT .. SELECT and CREATE ... SELECT that wrongly reset (thd->options & OPTION_STATUS_NO_TRANS_UPDATE) which caused us to loose some warnings about "Some non-transactional changed tables couldn't be rolled back") - Fixed use of uninitialized memory in ha_ndbcluster.cc::delete_table() which could cause delete_table to report random failures. - Fixed core dumps for some tests when running with --debug - Added missing FN_LIBCHAR in mysql_rm_tmp_tables() (This has probably caused us to not properly remove temporary files after crash) - slow_logs was not properly initialized, which could maybe cause extra/lost entries in slow log. - If we get an duplicate row on insert, change column map to read and write all columns while retrying the operation. This is required by the definition of REPLACE and also ensures that fields that are only part of UPDATE are properly handled. This fixed a bug in NDB and REPLACE where REPLACE wrongly copied some column values from the replaced row. - For table handler that doesn't support NULL in keys, we would give an error when creating a primary key with NULL fields, even after the fields has been automaticly converted to NOT NULL. - Creating a primary key on a SPATIAL key, would fail if field was not declared as NOT NULL. Cleanups: - Removed not used condition argument to setup_tables - Removed not needed item function reset_query_id_processor(). - Field->add_index is removed. Now this is instead maintained in (field->flags & FIELD_IN_ADD_INDEX) - Field->fieldnr is removed (use field->field_index instead) - New argument to filesort() to indicate that it should return a set of row pointers (not used columns). This allowed me to remove some references to sql_command in filesort and should also enable us to return column results in some cases where we couldn't before. - Changed column bitmap handling in opt_range.cc to be aligned with TABLE bitmap, which allowed me to use bitmap functions instead of looping over all fields to create some needed bitmaps. (Faster and smaller code) - Broke up found too long lines - Moved some variable declaration at start of function for better code readability. - Removed some not used arguments from functions. (setup_fields(), mysql_prepare_insert_check_table()) - setup_fields() now takes an enum instead of an int for marking columns usage. - For internal temporary tables, use handler::write_row(), handler::delete_row() and handler::update_row() instead of handler::ha_xxxx() for faster execution. - Changed some constants to enum's and define's. - Using separate column read and write sets allows for easier checking of timestamp field was set by statement. - Remove calls to free_io_cache() as this is now done automaticly in ha_reset() - Don't build table->normalized_path as this is now identical to table->path (after bar's fixes to convert filenames) - Fixed some missed DBUG_PRINT(.."%lx") to use "0x%lx" to make it easier to do comparision with the 'convert-dbug-for-diff' tool. Things left to do in 5.1: - We wrongly log failed CREATE TABLE ... SELECT in some cases when using row based logging (as shown by testcase binlog_row_mix_innodb_myisam.result) Mats has promised to look into this. - Test that my fix for CREATE TABLE ... SELECT is indeed correct. (I added several test cases for this, but in this case it's better that someone else also tests this throughly). Lars has promosed to do this. BitKeeper/etc/ignore: added mysys/test_bitmap include/base64.h: Removed my_global.h, as this must be included first in any program include/heap.h: Added heap_reset() (Required by new handler interface) include/my_base.h: Removed HA_EXTRA_RESET. MySQL will now call ::reset() instead of ::extra(HA_EXTRA_RESET). HA_EXTRA_RETRIVE_ALL_COLS and HA_EXTRA_RETRIVE_PRIMARY key are deleted as the column bitmaps makes these unnecessary include/my_bitmap.h: Remove my_pthread.h (should be included at upper level) Introduced my_bitmap_map typedef to make it the bitmap handling more like a black box Added bitmap_is_overlapping(), bitmap_test_and_clear(), bitmap_copy() and bitmap_cmp() Made bitmap_set_bit(), bitmap_flip_bit(), bitmap_clear_bit() return void include/myisam.h: Added mi_reset() (Required by new handler interface) include/myisammrg.h: Added myrg_reset() (Required by new handler interface) include/mysql_com.h: Added flag FIELD_IN_ADD_INDEX to be able to remove Field->add_index mysql-test/extra/binlog_tests/mix_innodb_myisam_binlog.test: Added testing of CREATE ... SELECT in a mixed environment (This found some bugs that Mats is going to fix shortly) mysql-test/install_test_db.sh: Simplify ldata usage Added --tmpdir=. option to mysqld bootstrap (Removed some warnings when TMPDIR was wrongly set) mysql-test/mysql-test-run.pl: Added --tmpdir=. to bootstrap mysql-test/mysql-test-run.sh: Use copy instead of INSTALL_DB for master and slave databases. (Speeds up startup time a lot!) Remove snapshot directories at startup (removes some strange warnings) mysql-test/r/binlog_row_mix_innodb_myisam.result: Added testing of CREATE ... SELECT in a mixed environment (This found some bugs that Mats is going to fix shortly) mysql-test/r/binlog_stm_mix_innodb_myisam.result: Added testing of CREATE ... SELECT in a mixed environment mysql-test/r/create.result: Some extra tests of warnings and number of tables opened by CREATE ... SELECT mysql-test/r/federated.result: Drop some left over tables Added testing of multiple table update and multiple table delete (with and without keys) mysql-test/r/func_gconcat.result: Enable some disabled tests (converted them slightly to be predictable) mysql-test/r/func_time.result: Added drop of test function mysql-test/r/innodb_mysql.result: Added tests for CREATE ... SELECT mysql-test/r/insert.result: More tests Added testing of duplicate columns in insert mysql-test/r/loaddata.result: Added testing LOAD DATA ... SET ... mysql-test/r/multi_update.result: Test multi updates and deletes using primary key and without mysql-test/r/ndb_index_unique.result: Better error message mysql-test/r/ndb_replace.result: New correct result after fixing REPLACE handling with NDB mysql-test/r/rpl_ddl.result: Now we don't get these (wrong) warnings anymore mysql-test/r/view_grant.result: Drop used views mysql-test/t/create.test: Some extra tests of warnings and number of tables opened by CREATE ... SELECT mysql-test/t/federated.test: Drop some left over tables Added testing of multiple table update and multiple table delete (with and without keys) mysql-test/t/func_gconcat.test: Enable some disabled tests (converted them slightly to be predictable) mysql-test/t/func_time.test: Added drop of test function mysql-test/t/innodb_mysql.test: Added tests for CREATE ... SELECT mysql-test/t/insert.test: More tests Added testing of duplicate columns in insert mysql-test/t/loaddata.test: Added testing LOAD DATA ... SET ... mysql-test/t/multi_update.test: Test multi updates and deletes using primary key and without mysql-test/t/view_grant.test: Drop used views mysql-test/valgrind.supp: Added supression of not needed warnings when printing stack trace mysys/base64.c: Include my_global.h first mysys/my_bitmap.c: Added bitmap_is_overlapping(), bitmap_test_and_clear() and bitmap_copy() Changed logic of bitmap handling to be a bit more efficent (Did this together with Mikael Ronström) Now the 'extra, not used bits' in the bitmap are assumed to have a 'random value' and the bitmap functions are free to change them whenever needed. Changed how mutex is allocated to make 'bitmap_free()' function simpler. mysys/thr_lock.c: Added 0x before thread pointers (for easier comparison of DBUG traces) sql/event.cc: Ensure 'use_all_columns()' is used for event tables Don't print warning that event table is damaged if it doesn't exists. sql/field.cc: Added ASSERT_COLUMN_MARKED_FOR_WRITE in all store() methods and ASSERT_COLUMN_MARKED_FOR_READ in all val() methods to catch wrong setting if table->read_set and table->write_set (Rest of changes are only indentation cleanups) sql/field.h: Removed Field->query_id (replaced by table->read_set and table->write_set) Removed Field->fieldnr (use Field->field_index instead) Removed Field->add_index (Use Field->flags instead) Add Field->part_of_key_not_clustered (for usage in opt_range.cc) sql/filesort.cc: Added paramater sort_postion to filesort() to force sorting by position instead of storing all fields in the result set. This allowed me to remove checking of sql_command. Create a temporary column bitmap for fields that are used by the sorting process. Use column bitmaps instead of query_id sql/ha_berkeley.cc: Update to 'newer' table handler interface sql/ha_berkeley.h: Update to 'newer' table handler interface sql/ha_federated.cc: Update to 'newer' table handler interface Only read columns that are needed from remote server. In case of eq ranges, don't generate two conditions in the WHERE clause (this can still be optimized, but would require a bigger code change) Use 'simpler to use' XXXX_LEN' macros A bit simpler logic in ::write_row() when creating statements. In update, only include test of fields actually read. (This greatly simplifies the queries sent by the federated engine) Similar changes done for delete_row() sql/ha_federated.h: Update to 'newer' table handler interface Changed XXX_LEN macros to use sizeof(...)-1, to simplify usage in ha_federated.cc Added HA_PRIMARY_KEY_REQUIRED_FOR_DELETE to tell MySQL to read all primary key columns in case of DELETE sql/ha_heap.cc: Update to 'newer' table handler interface sql/ha_heap.h: Update to 'newer' table handler interface sql/ha_innodb.cc: Update to 'newer' table handler interface - Update innobase_create_handler() to new interface - Removed HA_NOT_EXACT_COUNT (not needed) - Renamed HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION. - Prefixed base status variables with 'stats' - Use table column bitmaps instead of ha_get_bit_in_read_set() - Added ::reset(), with code from ::extra(HA_EXTRA_RESET) - Removed HA_EXTRA_RETRIVE_ALL_COLS and HA_EXTRA_RETRIEVE_PRIMARY_KEY as the table->read_set and table->write_set bitmaps now are accurate sql/ha_innodb.h: Update to 'newer' table handler interface - table_flags are now ulonglong - Added reset() method - Removed not needed ha_retrieve_all_cols() and ha_retrieve_all_pk() columns. - Made build_template() a class function to be able to easier access class variables sql/ha_myisam.cc: Update to 'newer' table handler interface sql/ha_myisam.h: Update to 'newer' table handler interface sql/ha_myisammrg.cc: Update to 'newer' table handler interface sql/ha_myisammrg.h: Update to 'newer' table handler interface sql/ha_ndbcluster.cc: Update to 'newer' table handler interface Fixed use_blob_value() to be accurate In ::complemented_read() we have to check both the read and write bitmap as the old code did mark all changed columns also in the read map Correct dumping of field data with DBUG_DUMP Prefix addresses in DBUG_PRINT with 0x Fixed usage of not initialized memory Update to use field->flags & FIELD_IN_ADD_INDEX instead of field->add_index. sql/ha_ndbcluster.h: Update to 'newer' table handler interface sql/ha_ndbcluster_binlog.cc: Mark usage of all columns in ndbcluster binlog tables false -> FALSE, true -> TRUE Use table->s->all_set instead of creating a temporary bitmap. sql/ha_partition.cc: Update to 'newer' table handler interface Added memroot to initialise_partitions() and related functions to get faster memory allocation. partition_create_handler() is now responsible for initialisation of the partition object Some trivial optimizations and indentation fixes Ensure that table_flags() are up to date Removed documentation for removed HA_EXTRA flags Fixed 'strange' usage of m_file[i] in new_handlers_from_part_info()that worked in current code 'by chance' sql/ha_partition.h: Update to 'newer' table handler interface sql/handler.cc: create_xxx handler now takes MEMROOT as an argument to simplify memory allocation. Much simpler get_new_handler() (Initialization of the object is now handled by the create method for the engine) Moved all allocation of bitmap handling to the TABLE object (in table.cc) Added column_bitmaps_signal() to signal column usage changes. Changed binlog_log_row() to use the exiusting all_set bitmap in the table object. Added ha_reset() function to test that the file object is ok at end of statement and call handler::reset() Added use_hidden_primary_key() to signal handler that we we are going to read and update + delete the row and the handler should thus remember the position for the row sql/handler.h: Added HA_NO_TRANSACTIONS, HA_PARTIAL_COLUMN_READ, HA_REQUIRES_KEY_COLUMNS_FOR_DELETE,HA_PRIMARY_KEY_REQUIRED_FOR_DELETE and HA_HAS_RECORDS Removed HA_NOT_EXACT_COUNT, HA_READ_RND_SAME HA_DUPP_POS -> HA_DUPLICATE_POS HA_NOT_EXACT_COUNT replaced by HA_STATS_RECORDS_IS_EXACT, HA_HAS_RECORDS and records() HA_PRIMARY_KEY_ALLOW_RANDOM_ACCESS renamed to HA_PRIMARY_KEY_REQUIRED_FOR_POSITION Added future row type 'ROW_TYPE_PAGES' Added MEM_ROOT to handlerton 'create' function Added ha_statistics, a structure for all status variable in the base handler class. Moved all status variables in the handler class into a stats structs to improve readability. ha_table_flags() is now a cached (not virtual) version of table_flags() reset() doesn't anymore call extra(HA_EXTRA_RESET) but is a function of it's own. Renamed dupp_ref to dup_ref Renamed not used handler::sortkey Moved read_set and write_set to TABLE structure handler::init() function added for cacheing of virtual constants from engine. sql/item.cc: Added register_field_in_read_map() for marking used columns in expression. This is used by filesort() for creating an optimal column bitmap while retrieving columns for sorting. Initalize value.cs_info.character_set_client to fix core dump bug with --debug set_query_id -> mark_used_columns Mark used columns in read_set OR write_set. sql/item.h: Removed reset_query_id_processor() as it's not needed anymore. Added register_field_in_read_map() Added extra argument to Item::walk() to indicate if we should also traverse sub queries. sql/item_cmpfunc.cc: Temporary mark used columns to be read/writable Update Item::walk to new interface sql/item_cmpfunc.h: Added extra argument to Item::walk() to indicate if we should also traverse sub queries. sql/item_func.cc: Update Item::walk() to new interface table_flags() -> ha_table_flags() sql/item_func.h: Update Item::walk() to new interface sql/item_row.cc: Update Item::walk() to new interface sql/item_row.h: Update Item::walk() to new interface sql/item_strfunc.h: Update Item::walk() to new interface sql/item_subselect.cc: Added Item_subselect::walk() (It was a bug it was missing before. Not sure what kind of bugs this could have caused) sql/item_subselect.h: Update Item::walk() to new interface sql/item_sum.cc: Update Item::walk() to new interface Updates for new handler interace sql/item_sum.h: Update Item::walk() to new interface sql/key.cc: Updates for new handler interace sql/log.cc: Mark all columns used for log tables Split options flag Ensured that second argument to trans_register_ha is a bool sql/log_event.cc: Fixed comments to be withing 79 characters Use OPTION_KEEP_LOG instead of OPTION_STATUS_NO_TRANS_UPDATE to remove wrong warnings Updates for new handler interface Use 0x%lx instead of %p (portability problem) sql/mysql_priv.h: Added OPTION_KEEP_LOG to indicate that we should replicate the binlog even on rollback Removed not used 'conds' argument to setup_tables sql/mysqld.cc: Indentation fixes and removed old comment sql/opt_range.cc: Update to new handler and bitmap interface. Fixed calls to cp_buffer_from_ref() and walk() (new argument). Create new temporary bitmaps for ror scans. (Needed because of handler changes and to get more accurate column bitmaps than before) Remove not needed file->ha_reset() call before file->close(). Some trivial optimization and indentation fixes. Use Field->part_of_key_not_clustered() to check if field is part of a key, instead of looping over all key parts. Added flag 'in_ror_merged_scan' to allow ::get_next() to know that we need a special column bitmap to only fetch pointer to record. This is needed because ror scan uses the same TABLE object but different file objects, which creates problem for the column bitmap handling. (This is a temporary solution. A better one would be to allocate an own TABLE object for ROR scans) Optimized bitmap handling in ror scans: - Start bitmap at position 0, not 1 - Use same bitmap size as in TABLE - Use table->read_set and table->write_set to create column bitmaps instead of looping over all fields in table sql/opt_range.h: Added 'in_ror_merged_scan' to indicate if we are doing a ROR scan Added temporary column bitmaps used in ROR scans sql/opt_sum.cc: Added get_ext_record_count() which is used in COUNT() optimization if handler has HA_HAS_RECORDS Note that we don't call this if handler has HA_STATS_RECORDS_IS_EXACT set. sql/protocol.cc: We need to mark columns as readable in ::store() as we sometimes return default value for fields to the user sql/records.cc: Updates for new handler interface sql/set_var.cc: Handle splitting OPTION_STATUS_NO_TRANS_UPDATE to two flags sql/share/errmsg.txt: Fixed wrong sql/sp.cc: Mark that we are using all columns for the proc table Update call to setup_tables() to use new prototype sql/sp_head.cc: Removed QQ comment sql/spatial.cc: Removed wrong QQ comment sql/sql_acl.cc: Mark that we need all columns for acl tables Supply memroot to some 'new' calls. Indentation fixes sql/sql_base.cc: set_query_id removed Ensure we call ha_reset() at end of each statement Mark read columns in read_set and changed columns in write_set (Before all columns was marked in read set) Fixed marking of some columns that was not proplerly marked before Maintain in TABLE->merge_keys set of all keys that are used in some way Removed not used 'conds' argument from setup_tables() Remove not used setting of 'dupp_field' in insert_fields() Added missing FN_LIBCHAR in mysql_rm_tmp_tables() (This has probably caused us to not properly remove temporary files after crash) sql/sql_bitmap.h: Added is_overlapping() sql/sql_class.cc: Slow_logs was not properly initialized, which could maybe cause extra/lost entries in slow log. set_query_id -> mark_used_columns Simpler variable usage in pack_row() (cleanup) Moved some variable declartion at start of function for better code readability sql/sql_class.h: Added enum_mark_columns Updated comments Renamed dupp_field -> dup_field Added virtual function 'can_rollback_data()' to select_insert() to be used in CREATE ... SELECT to optimize use of OPTION_STATUS_NO_TRANS_UPDATE. (This fixes a bug in CREATE ... SELECT where we did give wrong warnings when using non transacational tables) sql/sql_delete.cc: Updates to new handler interface Call table->mark_columns_needed_for_delete() to allow us to put additional columns in column usage maps if handler so requires. Call table->prepare_for_position() to tell handler that we are going to call ha_position(). Removed call to free_io_cache(). (io_cache is now removed in ha_reset()). Fixed calls to setup_tables() sql/sql_do.cc: Update call to setup_fields() sql/sql_handler.cc: Tell handler tables to always read all columns. Use temporary column map when storing value in field for later index usage sql/sql_help.cc: Makr all used fields to be read Update call to setup_fields() sql/sql_insert.cc: Tell handler we are going to update the auto_increment column dupp_field -> dup_field Set column usage bits for timestamp field. Call table->mark_columns_needed_for_insert() and table->mark_auto_increment_column() Removed not used argument from mysql_prepare_insert_check_table(). If we get an duplicate row on insert, change column map to read and write all columns while retrying the operatation. This is required by the definition of REPLACE and also ensures that fields that are only part of UPDATE are properly handled. This fixed a bug in NDB and REPLACE where REPLACE wrongly copied some column values from the replaced row. Setup new bitmaps for delayed insert rows Remove reseting of next_number_fields as it will be reset on next call to handler_insert() Fixed usage of thd->options and OPTION_STATUS_NO_TRANS_UPDATE. The issue was that one should not to reset this flag as it may be set by a previous statement. The way it was now used caused us to loose some warnings and get other wrong warnings when using non transactional tables mixed with transactional. I fixed it by introducing 'select_insert::can_rollback_data' to inform send_error() that the given statement can be rolled back (which in case of CREATE TABLE can always be done) Don't close tables created with CREATE ... SELECT but keep them in the table cache. Moved out MY_HOOKS from inside function (better readability) sql/sql_load.cc: Update to use new handler and column marking interface Update using setup_tables() sql/sql_olap.cc: Update calls to setup_tables Use enums instead of constants to setup_fields() sql/sql_parse.cc: Handle OPTION_KEEP_LOG: - Set it on CREATE TEMPORARY TABLE / DROP TABLE - Reset it when OPTION_STATUS_NO_TRANS_UPDATE is reset - Don't set it for CREATE ... SELECT (this is handled in select_create class) Remove reseting of OPTION_STATUS_NO_TRANS_UPDATE in begin_trans() as this should already be reset. If in autocommit mode, reset OPTION_KEEP_LOG and OPTION_STATUS_NO_TRANS_UPDATE to not give warnings in future commands sql/sql_partition.cc: Update walk() usage Trivial indentation fixes sql/sql_plugin.cc: Mark all columns as used for plugins sql/sql_prepare.cc: Added assert to find out hidden bugs in character_set_client (got an error in debug binary when this not set correctly) Updates for new handler interface Update calls to setup_fields() sql/sql_repl.cc: Indentation fixes sql/sql_select.cc: Update call to setup_tables() and setup_fields() Remove some old disabled code Update to new hadler interface Indentation cleanups Added column bitmaps for temporary tables. Remove updating of the removed slots in the Field class Added TABLE argument to cp_buffer_from_ref() (To be able to install temporary column maps) For internal temporary tables, use handler::write_row(), handler::delete_row() and handler::update_row() instead of handler::ha_xxxx() for faster execution. sql/sql_select.h: Indentaition fixes. Install temporary column usage maps when needed Added TABLE element to cp_buffer_from_ref() sql/sql_show.cc: Update to new handler interface Mark all columns used for internal tables. Style fixes. Added support for 'future' ROW_TYPE_PAGES. Don't allocate TMP_TABLE_PARAM with calloc. The 'init()' function will initialize the structure properly. sql/sql_table.cc: Update to new handler interface Simple my_snprintf -> strmake() Changed some constants to defines Don't test for NULL in primary key (as we a couple of line above force the PRIMARY KEY to be NOT NULL) Change field->add_index to use field->flags & FIELD_IN_ADD_INDEX Mark all columns as used for ALTER TABLE Style fixes Update call to filesort() sql/sql_trigger.h: Added friend functions to be able to test if triggers exists for table we are going to insert/update or delete in. sql/sql_udf.cc: Mark all columns as used for udf system table. sql/sql_union.cc: Update call to walk() Update to new handler interface sql/sql_update.cc: Remove query_id argument from compare_record() Use column bitmaps instead of query_id. We can't use compare_records() to skip updates for handlers that returns a partial column set and the read_set doesn't cover all columns in the write set, because compare_record() can't in this case know if a not read column changed value. Update call to setup_fields() Using separate column read and write sets allows for easier checking of timestamp field was set by statement. Removed call to free_io_cache() as this is now done in ha_reset() Call table->mark_columns_needed_for_update() and table->prepare_for_position() Style fixes sql/sql_view.cc: Style fixes sql/table.cc: Remove implicitely include 'errno.h' Remove code for building normalized path, as this is now identical to 'path' Remove field->fieldnr Added update of field->part_of_key_not_clustered() Create column bitmaps in TABLE and TABLE_SHARE Don't setup a temporary MEM_ROOT object as a thread specific variable for the handler. Instead we send the to-be-used MEMROOT to get_new_handler() Update to new handler interface Update call to walk() Added new functions: - st_table::clear_column_bitmaps() - st_table::prepare_for_position() - st_table::mark_columns_used_by_index() - st_table::restore_column_maps_after_mark_index() - st_table::mark_columns_used_by_index_no_reset() - st_table::mark_auto_increment_column() - st_table::mark_columns_needed_for_delete() - st_table::mark_columns_needed_for_update() - st_table::mark_columns_needed_for_insert() sql/table.h: Moved column usage bitmaps from handler to TABLE Added to TABLE_SHARE all_set and column_bitmap_size Added to TABLE merge_keys, bitmap_init_values, def_read_set, def_write_set, tmp_set, read_set and write_set. Declared all new table column bitmap functions Added TABLE functions column_bitmaps_set(), column_bitmaps_set_no_signal(), use_all_columns() and default_column_bitmaps() Added functions: tmp_use_all_columns() and tmp_restore_column_map() to temporarly switch column bitmaps Added functions: dbug_tmp_use_all_columns() and dbug_tmp_restore_column_map() to temporarly switch column bitmaps to avoid asserts in Field::store() and Field::val(). sql/tztime.cc: Mark all columns as used for timezone tables storage/archive/ha_archive.cc: Update to new handler interface storage/archive/ha_archive.h: Update to new handler interface storage/blackhole/ha_blackhole.cc: Update to new handler interface storage/blackhole/ha_blackhole.h: Update to new handler interface removed not needed flag HA_DUPP_POS storage/csv/ha_tina.cc: Update to new handler interface storage/csv/ha_tina.h: Update to new handler interface storage/example/ha_example.cc: Update to new handler interface storage/example/ha_example.h: Update to new handler interface storage/heap/hp_extra.c: Added heap_reset() (Required by new handler interface) storage/heap/hp_test2.c: Use heap_reset() storage/myisam/ft_boolean_search.c: Fixed compiler warning storage/myisam/mi_extra.c: Added mi_reset() (Required by new handler interface) storage/myisam/mi_search.c: Fixed DBUG_PRINT messages to use 0x%lx instead of %lx storage/myisam/mi_test2.c: Use mi_reset() storage/myisam/myisampack.c: Use mi_reset() storage/myisammrg/myrg_extra.c: Added myrg_reset() (Required by new handler interface) unittest/mysys/base64.t.c: Include my_global.h Don't include implictely include file 'stdlib.h'
-
- 02 Jun, 2006 1 commit
-
-
unknown authored
New prototype for get_auto_increment() (but new arguments not yet used), to be able to reserve a finite interval of auto_increment values from cooperating engines. A hint on how many values to reserve is found in handler::estimation_rows_to_insert, filled by ha_start_bulk_insert(), new wrapper around start_bulk_insert(). NOTE: this patch changes nothing, for all engines. But it makes the API ready for those engines which will want to do reservation. More csets will come to complete WL#3146. sql/ha_berkeley.cc: update to new prototype of get_auto_increment sql/ha_berkeley.h: update to new prototype of get_auto_increment sql/ha_heap.cc: update to new prototype of get_auto_increment sql/ha_heap.h: update to new prototype of get_auto_increment sql/ha_innodb.cc: update to new prototype of get_auto_increment sql/ha_innodb.h: update to new prototype of get_auto_increment sql/ha_myisam.cc: update to new prototype of get_auto_increment sql/ha_myisam.h: update to new prototype of get_auto_increment sql/ha_ndbcluster.cc: update to new prototype of get_auto_increment sql/ha_ndbcluster.h: update to new prototype of get_auto_increment sql/ha_partition.cc: update to new prototype of get_auto_increment sql/ha_partition.h: update to new prototype of get_auto_increment sql/handler.cc: new prototype of get_auto_increment, comments, preparation for when the MySQL layer is capable of getting finite auto_increment intervals from cooperating engines. sql/handler.h: a wrapper around start_bulk_insert(): ha_bulk_insert(), which stores the argument (number of rows expected for insertion) into a member of handler: estimation_rows_to_insert. This member will soon be used to decide how much auto_increment values we want to reserve from cooperating engines. New prototype for get_auto_increment, preparing for cooperating engines to reserve finite auto_increment intervals. release_auto_increment() will be used by the MySQL layer to inform the engine that it has not used all of the interval (engine can reclaim it). sql/log.cc: note for the future sql/log_event.cc: name of wrapper sql/sql_insert.cc: name of wrapper. When done with inserting, return unused auto_inc values to engine. sql/sql_load.cc: name of wrapper. When done with inserting, return unused auto_inc values to engine. sql/sql_select.cc: name of wrapper sql/sql_table.cc: name of wrapper storage/archive/ha_archive.cc: update to new prototype of get_auto_increment. Archive reserves only one value (Archive's god - Brian - told me that's the truth). storage/archive/ha_archive.h: update to new prototype of get_auto_increment()
-