Push the patch of Jan Lindstrom: better comments
ha_innodb.cc:
Partial fix for Bug #12263 : we let InnoDB always to perform a rollback on the trx object if MySQL closes a connection; but we do print a warning to the .err log if an InnoDB transaction was active; we may remove that print later, since the situation really is not a bug; MySQL just is not aware that some cursor operation started an InnoDB transaction
cursors for InnoDB. The idea of the patch is that if MySQL requests
a consistent read view, we open one when open a cursor, set is as the
active view to a transaction when fetch from the cursor, and close
together with cursor close. This patch is associated to bugs #11813,
#11832, and #11833. Contains after review fixes.
cursors. This should fix Bug#11813 when InnoDB part is in
(tested with a draft patch).
The idea of the patch is that if a storage engine supports
consistent read views, we open one when open a cursor,
set is as the active view when fetch from the cursor, and close
together with cursor close.
The idea of the patch
is that every cursor gets its own lock id for table level locking.
Thus cursors are protected from updates performed within the same
connection. Additionally a list of transient (must be closed at
commit) cursors is maintained and all transient cursors are closed
when necessary. Lastly, this patch adds support for deadlock
timeouts to TLL locking when using cursors.
+ post-review fixes.
Changed defaults option --instance to --defaults-group-suffix
Changed option handling to allow --defaults-file, --defaults-extra-file and --defaults-group-suffix to be given in any order
Changed MYSQL_INSTANCE to MYSQL_GROUP_SUFFIX
mysql_print_defaults now understands --defaults-group-suffix
Remove usage of my_tempnam() (not safe function)
if( -> if ( and while( to while (
Let InnoDB use a consistent read when it initializes the auto-inc counter for a table: this will eliminate spurious deadlocks, but will ignore an UPDATE if that happens at the same time that we init the auto-inc counter; this has to be documented; this path also fixes most of Bug #11633, but not all: if ::external_lock() is not called on the table in SHOW TABLE STATUS, that might cause a crash if someone simultaneously DROPs the table
Since UNLOCK TABLES will commit the transaction (and thus release
all InnoDB locks held by it), it is unnecessary to release the
user-requested table locks separately.
Update the InnoDB internal auto-inc counter in ::write_row() if the statement is a REPLACE that fails in a duplicate key error: REPLACE will handle duplicate key error, and the insert will eventually succeed; note that we did NOT change InnoDB's behavior in an UPDATE, because updating the auto-inc counter in an UPDATE would require a bigger patch (Bug #11005)
Ensure that 'null_value' is not accessed before val() is called in FIELD() functions
Fixed initialization of key maps. This fixes some problems with keys when you have more than 64 keys
Fixed that ROLLUP don't always create a temporary table. This fix ensures that func_gconcat.test results are now predictable
Weaken InnoDB table locking at the start of a stored procedure call: this will make multi-transaction stored procedures less deterministic for binlogging, but since they are non-deterministic anyway, it is not that bad; this fix will remove the huge amount of InnoDB tranasctional deadlocks caused by table locking at the start of a stored procedure
The natural InnoDB type for NEWDECIMAL is FIXBINARY, not BINARY, since there probably is no benefit from stripping trailing bytes 0x20; note that tables created with 5.0.6 and containing DECIMAL do not need to be rebuilt, as BINARY works as well; note that because of Jani's fix to bug 10465, tables created with 5.0.? - 5.0.5 DO NEED TO BE REBUILT if they contain DECIMAL
Fix bug #10359 : the critical AUTO-INC bug in InnoDB; since reading of the auto-inc counter in an INSERT was not protected by the AUTO-INC table lock of InnoDB, two inserted rows could get the same value for the auto-inc counter, leading to a duplicate key error
remove FIX_GCC_LINKING_PROBLEM and -DDEFINE_CXA_PURE_VIRTUAL
replace echo in configure.in with AC_MSG_WARN/AC_MSG_ERROR
don't set -DUSE_MYSYS_NEW for gcc 2.95
set $USE_MYSYS_NEW even if CXX is g++
yassl bugfixes
instantiate all yassl templates explicitly
Add UNIV_UNLIKELY hint.
lock0lock.c:
Correct the prototypes of lock_deadlock_recursive() and
lock_deadlock_occurs().
data0data.c:
dtuple_convert_big_rec(): Add UNIV_UNLIKELY hint.
ha_innodb.cc:
ha_innobase::write_row(): Correct the type of 'mode'.
Semi-synchronous replication for InnoDB type tables; before telling the client that a commit has been processed, wait that the replication thread has returned from my_net_send() where it sends the binlog to the slave; note that TCP/IP, even with the TCP_NODELAY option does not guarantee that the slave has RECEIVED the data - this is just heuristic at the moment; this is useful in failover: in almost all cases, every transaction that has returned from the commit has been sent and processed in the slave, which makes failover to the slave simpler if the master crashes; the code does not work yet as is, because MySQL should call innobase_report_binlog_offset_and_commit() in a commit; we will most probably return that call to 5.0.x, to make InnoDB Hot Backup and group commit to work again; XA code broke them temporarily in 5.0.3
most InnoDB threads are not terminated properly and the buffer pool is not flushed
to disk. Still no committed transaction is lost as we flush the logs to disk.
InnoDB does crash recovery at startup after this shutdown.
Using this shutdown in testsuite (mysql-test-run --mysqld=--innodb_fast_shutdown=2) saved 3 minutes (13% of total time).
Fix part of bug #9670: if MySQL calls ::store_lock with TL_IGNORE, do not change prebuilt->select_lock_type; this fix may heal the assertion failures reported in UPDATE and multi-table UPDATE; it is not clear if this fixes the problems in OPTIMIZE TABLE
Fix Bug #9526 in 5.0: MySQL ENUM and SET columns are internally actually unsigned integer types; we must take care that old tables still treat ENUM and SET (incorrectly) as a character string, while new created tables treat it correctly as an unsigned integer
Correct the fix of Bug #9526 : InnoDB must use its own internal type info for old tables, so that old ENUMs and SETs still are (incorrectly) seen as char strings; we do not dare to allow InnoDB sometimes to see the type as an integer type for those old tables
Fix bug #9314 in InnoDB true VARCHAR: InnoDB stored the 'position' of a row wrong in a column prefix primary key index; this could cause MySQL to complain 'ERROR 1032: Can't find record in ...' in an update of the primary key, and also some ORDER BY or DISTINCT queries
Fix bug: InnoDB created all indexes on true VARCHARs as column prefix keys because field->pack_length() for a true VARCHAR is bigger than key_part->length; FOREIGN KEYs could not be created on true VARCHARs then
Fix a crash in a simple search with a key: the dtype->len of a true VARCHAR is the payload maximum len in bytes: it does not include the 2 bytes MySQL uses to store the string length
ha_innodb.cc:
Fix a crash in true VARCHARs in test-innodb: we passed a wrong pointer to the column conversion in an UPDATE
rowid_order_innodb.result, ps_3innodb.result, innodb.result, endspace.result:
Edit InnoDB test results to reflect the arrival of true VARCHARs
Add a settable session variable innodb_support_xa; setting it to 0 can save up to 10 % of CPU time and 150 bytes of space in each undo log
trx0trx.h, trx0undo.c, trx0trx.c, trx0roll.c:
Enable XA if innodb_support_xa is not set to 0; make prepare to do log fsync's according to innodb_flush_log_at_trx_commit
In TRUNCATE TABLE, MySQL may call innobase_xa_prepare() even though there is no active transaction because it was committed in ha_innobase::delete_all_rows(); in innobase_xa_prepare() InnoDB starts then a new transaction; let us adjust the trx->active_trans flag accordingly
Fix a glitch reported by Guilhem: release the memory allocated for innodb_internal_... if innobase_init() fails; cleanups: in ha_innodb, you should use the InnoDB formatting style for the source code
single table DELETE...SELECT clauses when innobase_locks_unsafe_for_binlog
is used and isolation level of the transaction is not serializable.
InnoDB uses consistent read in these cases for a selected table.
Backported from 5.0.x.
single table DELETE...SELECT clauses when innobase_locks_unsafe_for_binlog
is used and isolation level of the transaction is not serializable.
InnoDB uses consistent read in these cases for a selected table.
