If MySQL wrote to its binlog, but for some reason trx->update_undo and trx->insert_undo were NULL in InnoDB, then trx->commit_lsn was garbage, and InnoDB could assert in the log flush of trx_commit_complete_for_mysql() (Bug #9277)
If MySQL wrote to its binlog, but for some reason trx->update_undo and trx->insert_undo were NULL in InnoDB, then trx->commit_lsn was garbage, and InnoDB could assert in the log flush of trx_commit_complete_for_mysql() (Bug #9277)
Change the sorting order of TEXT columns in InnoDB: pad with spaces at the end in comparisons; we MUST UPDATE the MySQL MANUAL to warn about the need to rebuild tables (also MyISAM) in certain cases in an upgrade to 5.0.3; TODO: study what complications the missing DATA_BINARY_TYPE flag in InnoDB < 4.0.14 causes: we would compare then also BLOBs with space padding, not just TEXT
Revert the change to the space padding of BLOB and TEXT in comparisons; we cannot change the sorting order, because that would make old tables to appear corrupt; better to change LIKE 'a%' processing in MySQL, so that the lower end of the search interval would be 'a', not 'a '
Fix that 'a' LIKE 'a%' was not true for an InnoDB BLOB or TEXT type column prefix index search: InnoDB assumed that the BLOB and TEXT should not be padded with space in comparisons; this bug is also in 4.1; it may have worked in 4.1 because MySQL may have asked there for strings >= 'a', while in 5.0 it asks for strings >= 'a '
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
instead of SYS_TABLES.MIX_LEN, because the latter was not
initialized to zero in old MySQL 3.23 releases. This will break
existing MySQL/InnoDB 5.0.3-bk databases for which
SHOW TABLE STATUS displays Row_format=Compact.
which is always false. In a very fast InnoDB shutdown, we just ensure that
no more transactions are running, flush InnoDB log, signal InnoDB threads to die,
and then return from InnoDB (from innobase_end()) without waiting for those threads
to actually die. I have tested on a 4CPU machine that even with --innodb_flush_log_at_trx_commit=0,
this optimized InnoDB very fast shutdown loses no committed transactions. Patch pre-approved by Heikki.
dict_load_table(): Remove the check for row_format=compact for now,
because the flag bit we used (high-order bit of mix_len)
has not been zero for at least two customers.
AIX 5.1 after security patch ML7 seems to contain a bug that instead of EEXIST it sets errno to 0 if a file creation fails because the file already exists. Work around that bug by interpreting errno 0 in AIX as EEXIST.
Make InnoDB-5.0.3 to process log records of 4.1 undo log header create and reuse just like in 4.1; storing XID to the start of the undo log is a separately logged operation; this conforms to Rule 3 of InnoDB redo logging
Fix bug #8677: if one used LOCK TABLES, created an InnoDB temp table, and did a multi-table update where a MyISAM table was the update table and the temp table was a read table, then InnoDB aserted in row0sel.c because n_mysql_tables_in_use was 0. Also, we remove the assertion altogether and just print an error to the .err log if this important consistency check fails. Then it is up to the user to read the .err log and notice the problem if there still are errors in MySQL's table locking.
Fix memory corruption bug reported by Georg Richter: if one created a table whose primary key contained at least 2 column prefixes, that could corrupt the memory heap of prebuilt_t
Use a 32-bit right-shift implementation through a 64-bit integer that should work ok for both gcc and Visual C++; the code depended on defining SIZEOF_LONG, and that is not set when compiling ibbackup
Fix Windows porting bugs that broke ibbackup: 1) wrong error check in for CreateDirectory(), 2) wrong error check if the file did not exist in DeleteFile(), 3) too strict sharing restrictions in os_file_create_simple(): when ibbackup called that function, it would not allow mysqld to write to the file
We accidentally checked if the DIRECTORY is of type OS_FILE_TYPE_UNKNOWN; our intention was to check if the FILE is that; best to remove the check altogether, as in crash recovery it is safest to try to open also files whose type is unknown
os0file.c:
Fix a bug: in Windows, os_file_readdir_next_file() returned OS_FILE_TYPE_UNKNOWN as the type of a regular file; this did not break mysqld, but did break ibbackup on Windows
Corrected formatting of comments.
rem0rec.ic:
rec_set_info_and_status_bits(): pass arguments to rec_set_info_bits()
in the correct order.
Corrected formatting of comments.
It is not safe to infer the status bits from the B-tree page
level, because after MLOG_COMP_LIST_END_COPY_CREATED, the
level will not be initialized before the records have been inserted.
(Bug #7973)
Fix a race condition that could cause the assertion space->n_pending_flushes == 0 to fail in fil0fil.c, in fil_space_free(), in DROP TABLE or in ALTER TABLE
at run-time, so that an executable compiled on Mac OS X 10.2 can
be run on Mac OS X 10.2 (without the work-around) and Mac OS X 10.3
and later with the work-aroud enabled.
Add comments about why the InnoDB latching order is obeyed also for the MySQL query cache mutex; add an error printf if that is not the case
sync0sync.h:
Assign sync0sync.h ranks also for the MySQL query cache mutex and the MySQL binlog mutex; the latching order must be obeyed also for these
row0ins.c:
Add a comment why the query cache invalidate operation cannot deadlock in a cascaded FOREIGN KEY operation
Use the fcntl() file flush method on OS X; Apple disabled fsync() for internal disk drives, which caused corruption in power outages; the patch was recommended by an Apple engineer
Fix a wrong memset in InnoDB Hot Backup code; the bug probably did not affect anything since we do not assume that the header of a log file is filled with zeros before writing the header info there; the bug found by Felix von Leitner
Fix a little bug in InnoDB: we looked at the physical size of a stored SQL NULL value from a wrong field in the index; this has probably caused no bugs visible to the user, only caused some extra space usage in some rare cases; we may later backport the fix to 4.0
rem0rec.c:
rec_init_offsets(): Make the function comment more accurate.
rem0rec.ic:
rec_get_nth_field(): Return pointer to the field, even to NULL fields.
rec_set_nth_field(): Make more accurate debug assertions.
Print a more descriptive error and refuse to start InnoDB if the size of ibdata files is smaller than what is stored in the tablespace header; innodb_force_recovery will override this
Fix the previous bug fix: dropping a table with FOREIGN KEY checks running on it caused a cascade of failed drops while the foreign key check was waiting for a lock
Fix bug: if we dropped a table where an INSERT was waiting for a lock to check a FOREIGN KEY constraint, then an assertion would fail in lock_reset_all_on_table(), since that operation assumes no waiting locks on the table or its records
row0mysql.c:
Fix bug: InnoDB failed to drop a table in the background drop queue if the table was referenced by a foreign key constraint
Fix InnoDB critical bug #7496; we scan the InnoDB data dictionary also at a normal mysqld startup, and create the spaces, so that we know the mapping space id -> .ibd file name; fix an infinite loop if DISCARD TABLESPACE coincides with INSERT or some other table operation; fix a potential crash if DISCARD TABLESPACE coincides with a cascaded FOREIGN KEY operation in the same table; do not allow DISCARD TABLESPACE of a referenced table if FOREIGN_KEY_CHECKS=1
Logging to logging@openlogging.org accepted
sql_yacc.yy, sql_parse.cc, sql_lex.h, lex.h:
Implements the SHOW MUTEX STATUS command
set_var.cc, mysqld.cc, mysql_priv.h:
Added new GLOBAL variable timed_mutexes
ha_innodb.h:
New function innodb_mutex_show_status
ha_innodb.cc:
Added new innodb variables in SHOW STATUS
Implements the SHOW MUTEX STATUS command
innodb.test, innodb.result:
Added new row_lock_waits status variables tests.
variables.test, variables.result:
test new variable timed_mutexes
ut0ut.c:
New function ut_usectime.
sync0sync.c:
Mutex counting.
sync0rw.c:
New mutex parameters initialization.
srv0srv.c:
Counting row lock waits
row0sel.c, row0mysql.c:
Setting row_lock or table_lock state to thd.
que0que.c:
Added default no_lock_state to thd.
univ.i:
Added UNIV_SRV_PRINT_LATCH_WAITS debug define
sync0sync.ic:
Count mutex using.
sync0sync.h:
Added new parameters to mutex structure for counting.
sync0rw.h:
Added new parameters to rw_create_func.
srv0srv.h:
Added new innodb varuables to SHOW STATUS.
que0que.h:
Added thread lock states.
Fix InnoDB bug: on HP-UX, with a 32-bit binary, InnoDB was only able to read or write <= 2 GB files; the reason was that InnoDB treated the return value of lseek() as a 32-bit integer; lseek was used on HP-UX-11 as a replacement for pread() and pwrite() because HAVE_BROKEN_PREAD was defined on that platform
Print progress of background rollback of transactions with more than 1000 undo log entries
srv0start.c, trx0roll.c, log0recv.c, trx0roll.h:
Cleanup background rollback code in crash recovery; do not flush all modified pages from the buffer pool after a crash recovery: this makes mysqld accesible for users more quickly
Do not use short int in rem0rec.ic, since its size is not fixed in ANSI C; improve comments of the relative offset field in a record; use mach_read_from_2() to read the relative offset field to save CPU time, if the compiler does not optimize a more complex access function
Fix for the 0xA0 character problem in the InnoDB FOREIGN KEY parser: if my_isspace() treats 0xA0 as space, then let InnoDB do the same; this might break some multi-byte charset id's, though for big5, ujis, sjis this seems not to change the current behavior (I checked the tables in /share/charsets); this fix must NOT be merged to 4.1 because in 4.1 everything is in UTF-8
Fix the bug that the character 0xA0 that EMS MySQL Manager in ALTER TABLE adds after a table name confuses the InnoDB FOREIGN KEY parser, causing an error 121 when we try to add a new constraint; a full fix would require the lexer to be aware of thd->charset_info() and UTF-8
srv_printf_innodb_monitor(): Removed extraneous rewind() call.
The caller of this function calls rewind() if necessary.
In this way, we avoid rewind()ing stderr in
srv_lock_timeout_and_monitor_thread().
transactional table locks to tables mentioned in the query. These locks
are released at the end of the transaction automatically.
This is fix for bugs #5655, #5998 and issue #3762.
Remove parameter from call to btr_search_validate().
buf0buf.c:
Initialize member "index" of buf_block_t.
buf0buf.h:
Add member "index" to buf_block_t.
btr0sea.h:
Remove parameter of btr_search_validate()
btr0sea.c:
Make use of the added member "index" of buf_block_t.
Let MySQL check the existence of readdir_r with 3 arguments; Solaris seems to have just 2 args
Check the existence of readdir_r and localtime_r; even though MySQL does check these too, we need our own check for Hot Backup code
os0file.c:
Use re-entrant readdir_r where available
ut0ut.c:
Make a function to use thread-safe localtime_r where available; that particular function was not called from anywhere, though
