Added a test case for bug #7769.
item_sum.h:
Fixed bug #7769: a crash for queries with group_concat and
having when the query table was empty.
The bug was due an unsafe dereferencing.
sql/item_sum.h:
Fixed bug #7769: a crash for queries with group_concat and
having when the query table was empty.
The bug was due an unsafe dereferencing.
mysql-test/t/func_gconcat.test:
Added a test case for bug #7769.
mysql-test/r/func_gconcat.result:
Added a test case for bug #7769.
include/my_sys.h:
Added function to call if IO_CACHE is moved
mysys/mf_iocache.c:
Added function to call if IO_CACHE is moved
sql/filesort.cc:
Tell that io_cache is moved
--with-cluster are provided. Yes, this sounds like a contradiction -
enabling debugging in NBD enables code parts that may still reveal
portability issues when compiled with non-gcc compilers. (request by TomasU)
The bug was that if you have two TL_WRITE_DELAYED at the same time,
mi_lock_databases() could be done in the wrong order and we could write the wrong header to the MyISAM index file.
sql/mysql_priv.h:
Fix bug in insert delayed with prepared statements
sql/sql_base.cc:
Fix bug in insert delayed with prepared statements
sql/sql_prepare.cc:
Fix bug in insert delayed with prepared statements
The bug was that if you have two TL_WRITE_DELAYED at the same time,
mi_lock_databases() could be done in the wrong order and we could write the wrong header to the MyISAM index file.
Merge from 4.0: Fix a theoretical hang over the adaptive hash latch in InnoDB if one runs INSERT ... SELECT ... (binlog not enabled), or a multi-table UPDATE or DELETE, and only the read tables are InnoDB type, the rest are MyISAM; this also fixes bug #7879 for InnoDB type tables
sql/ha_innodb.cc:
Merge from 4.0: Fix a theoretical hang over the adaptive hash latch in InnoDB if one runs INSERT ... SELECT ... (binlog not enabled), or a multi-table UPDATE or DELETE, and only the read tables are InnoDB type, the rest are MyISAM; this also fixes bug #7879 for InnoDB type tables
key algorithm for keys where they were explicitly specified.
mysql-test/r/show_check.result:
Add results
mysql-test/t/show_check.test:
Add tests for preservation of key algorithm in SHOW CREATE TABLE output
Fix a theoretical hang over the adaptive hash latch in InnoDB if one runs INSERT ... SELECT ... (binlog not enabled), or a multi-table UPDATE or DELETE, and only the read tables are InnoDB type, the rest are MyISAM; this also fixes bug #7879 for InnoDB type tables
sql/ha_innodb.cc:
Fix a theoretical hang over the adaptive hash latch in InnoDB if one runs INSERT ... SELECT ... (binlog not enabled), or a multi-table UPDATE or DELETE, and only the read tables are InnoDB type, the rest are MyISAM; this also fixes bug #7879 for InnoDB type tables
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.
innobase/include/srv0start.h:
Mac OS X: Add srv_have_fullfsync
innobase/os/os0file.c:
os_file_flush(): Use F_FULLFSYNC on Mac OS X 10.3, but not on 10.2
innobase/srv/srv0start.c:
innobase_start_or_create_for_mysql(): When compiled on OS X 10.2,
detect if the binary is running on OS X 10.3 or later, and set
srv_have_fullfsync accordingly.
