RESET MASTER, this breaks replication).
In the slave SQL thread, any DROP TEMPORARY TABLE is silently
converted to DROP TEMPORARY TABLE IF EXISTS, not to have irrelevant
"table does not exist". See comments in code or bug #186 for an example
of how irrelevant errors like this can arise.
Now it does not SEGV, but USER() is still badly replicated
(it is replicated to ""), which is a lower priority bug.
sql/item_strfunc.cc:
Fix for #178 Replicating INSERT VALUES(USER()) crashes (SEGV) the slave
Now it does not SEGV, but USER() is still badly replicated
(it is replicated to ""), which is a lower priority bug.
client/mysql.cc:
Remove unnecessary pthread_yield calls for NetWare
client/mysqldump.c:
Remove unnecessary pthread_yield calls for NetWare
include/my_pthread.h:
Special extern my_pthread_cond_timedwait no longer required for NetWare
mysql-test/t/backup.test:
replace_result added so test passes correctly on NetWare
mysys/default.c:
Add conditional to avoid non-applicable file checking on NetWare
mysys/my_init.c:
Remove working directory change from netware_init()
mysys/my_pthread.c:
Remove unnecessary NetWare version of my_pthread_cond_timedwait
mysys/thr_mutex.c:
Remove unnecessary NetWare block
scripts/make_binary_distribution.sh:
Changes to clear up extraneous EXTRA_BIN_FILES, NetWare sections
Use SESSION TABLE_HANDLER as default table handler if given table handler doesn't exists
Docs/internals.texi:
Updated description for how DATE, TIME and TIMESTAMP are stored
mysql-test/r/create.result:
New tests
mysql-test/t/create.test:
New tests
sql/handler.cc:
Use SESSION TABLE_HANDLER as default table handler if given table handler doesn't exists.
sql/sql_parse.cc:
Fix that we return hostname instead of IP in SHOW PROCESSLIST
sql/sql_show.cc:
Print right hostname or IP in SHOW PROCESSLIST
only do it if the user specifies --delete-master-logs (new option).
Safer + fixes bug #159.
client/client_priv.h:
New option --delete-master-logs for mysqldump
client/mysqldump.c:
New option --delete-master-logs for mysqldump
BitKeeper/etc/logging_ok:
Logging to logging@openlogging.org accepted
MEMORY is alias for HEAP for CREATE TABLE ... TYPE=HEAP
Fixed bug in multi-table-update where a row could be updated several times
mysql-test/r/heap.result:
Test of CREATE TABLE ... type=MEMORY
mysql-test/r/innodb.result:
Added multi-update-test
mysql-test/r/multi_update.result:
Added multi-update-test
mysql-test/t/heap.test:
Test of CREATE TABLE ... type=MEMORY
mysql-test/t/innodb.test:
Added multi-update-test
mysql-test/t/multi_update.test:
Added multi-update-test
sql/ha_innodb.h:
Allow optimzation of multi-table-update also for InnoDB tables
sql/handler.h:
Allow optimzation of multi-update also for InnoDB tables
sql/key.cc:
After merge fix
sql/lex.h:
MEMORY is alias for HEAP
sql/sql_test.cc:
Fixed wrong printf
sql/sql_update.cc:
Fixed bug in multi-table-update where a row could be updated several times
sql/sql_yacc.yy:
MEMORY is alias for HEAP
symbol in libmysqlclient) (thanks to Dennis Haney for the initial patch)
- cleanup: removed client/password.c (not used at all) and
libmysql/password.c (should rather be a symlink to sql/password.c instead)
- applied HPUX11 portability fix for char_val declaration to sql/password.c
(taken from libmysql/password.c)
BitKeeper/deleted/.del-password.c~c036d4f8b3280843:
Delete: client/password.c
BitKeeper/deleted/.del-password.c~76f30876e68eddb4:
Delete: libmysql/password.c
include/mysql_com.h:
- replaced "rnd" with "my_rnd"
libmysql/Makefile.am:
- removed dead piece of code ($qs was not defined)
- symlink password.c from the sql directory
libmysqld/Makefile.am:
- use password.c from the sql directory instead
sql/item_func.cc:
- replaced "rnd" with "my_rnd"
sql/mysqld.cc:
- replaced "rnd" with "my_rnd"
sql/password.c:
- replaced "rnd" with "my_rnd"
sql/sql_class.cc:
- replaced "rnd" with "my_rnd"
sql/sql_crypt.cc:
- replaced "rnd" with "my_rnd"
innobase/row/row0mysql.c:
Auto merged
innobase/row/row0sel.c:
Auto merged
sql/key.cc:
Auto merged
configure.in:
Merge with 3.23
mysql-test/r/innodb.result:
Merge with 3.23
mysql-test/t/innodb.test:
Merge with 3.23
strings/strto.c:
Merge with 3.23
a deadlock between I/O and SQL threads in replication
when relay_log_space is too small. This fixes bug #79.
sql/log.cc:
New variable rli->ignore_log_space_limit to resolve
a deadlock between I/O and SQL threads in replication
when relay_log_space is too small.
sql/slave.cc:
New variable rli->ignore_log_space_limit to resolve
a deadlock between I/O and SQL threads in replication
when relay_log_space is too small.
sql/slave.h:
New variable rli->ignore_log_space_limit to resolve
a deadlock between I/O and SQL threads in replication
when relay_log_space is too small.
sql/sql_repl.cc:
New variable rli->ignore_log_space_limit to resolve
a deadlock between I/O and SQL threads in replication
when relay_log_space is too small.
Ensure that \ is not allowed in filenames, even on Unix
include/my_dbug.h:
Rename: include/dbug.h -> include/my_dbug.h
include/Makefile.am:
Change dbug.h -> my_dbug.h
include/my_global.h:
Change dbug.h -> my_dbug.h
libmysql/get_password.c:
Change dbug.h -> my_dbug.h
libmysqld/lib_vio.c:
Change dbug.h -> my_dbug.h
sql/table.cc:
Ensure that \ is not allowed in filenames, even on Unix
Fix compiler warning in Windows about ib_longlong to ulint conversion
innobase/trx/trx0roll.c:
Fix compiler warning in Windows about ib_longlong to ulint conversion
Fix bug number 154: GROUP BY and DISTINCT could treat NULL values inequal
innobase/row/row0sel.c:
Fix bug number 154: GROUP BY and DISTINCT could treat NULL values inequal
Fix bug number 154: GROUP BY and DISTINCT could treat NULL values inequal
innobase/row/row0sel.c:
Fix bug number 154: GROUP BY and DISTINCT could treat NULL values inequal
