Fixed memory leak when replication restarts in debug mode
include/my_sys.h:
Added option to not overwrite files to my_copy()
mysql-test/mysql-test-run.sh:
Fixed --ddd option
Fixed that mysqld is restarted if there is a testname-master.sh file
mysql-test/r/backup.result:
Update for security fix in BACKUP TABLE
mysql-test/t/backup.test:
Update for security fix in BACKUP TABLE
mysys/my_copy.c:
Added option to not overwrite files to my_copy()
sql/ha_myisam.cc:
Don't allow BACKUP TABLE to overwrite files
sql/slave.cc:
Fixed problem with --debug output from 'handle_slave'
Fixed memory leak when replication restarts in debug mode
mysql-test/r/join.result:
Test of range optimizer bug
mysql-test/t/join.test:
Test of range optimizer bug
sql/sql_rename.cc:
Added missing DEBUG_PRINT()
sql/lock.cc:
Added functions to handle list of table name locks
sql/mysql_priv.h:
Added functions to handle list of named locks
sql/sql_rename.cc:
Use new general table name lock functions
sql/sql_table.cc:
Require table name locks when doing drop table.
This fixed a deadlock problem when using LOCK TABLE in one thread and DROP TABLE in another
Add diagnostic prints if insert buffer merge is tried to a page whose type is not an index page, try to recover from the situation by discarding the insert buffer records
innobase/ibuf/ibuf0ibuf.c:
Add diagnostic prints if insert buffer merge is tried to a page whose type is not an index page, try to recover from the situation by discarding the insert buffer records
sql/sql_class.cc:
backporting Peter's fix for better RAND() initialization from 4.0
to see what it does try
perl -e 'while($N=shift) { printf "N=%-10d",$N; @b=@c=(); for (1..$N) { @a=split(" ",`mysql -BNe "select rand(), rand(), rand(), rand()"`); for(0..3) { $b[$_]+=$a[$_]; $c[$_]+=$a[$_]*$a[$_]; }} for (0..3) { printf "%15.14f ",($c[$_]-$b[$_]*$b[$_]/$N)/$N } print "\n" }' 10 10 10 100 100 10 100 1000
before and after the fix
libmysql/libmysql.c:
This is a fix for a bug in connect_timeout.
This bug manifested itself on operating systems that do support poll()
system call, which resulted in a timeout twice the value specified.
That is because timeout executed on both select() and poll().
mysql-test/r/type_datetime.result:
result for datetime fix
mysql-test/t/type_datetime.test:
test for datetime fix
sql/sql_select.cc:
Fix for a bug caused implicitely by MySQL having to comply with
ODBC standard on datetime comparisons.
This failed to update used_tables() for OR condition, which is why
query returned all rows on query as in .test above, while
none of types contains NULL nor 0 datetime (last one from ODBC)
- add version number to build log files (e.g. Logs/build-4.1-standard.log)
- add host name and uname to Subject line of build failure mails
(e.g. "Subject: build(Linux): mysql-4.1-standard compilation failed")
-> Please merge with all other trees!
Build-tools/Do-compile:
Small enhancements:
- add version number to build log files (e.g. Logs/build-4.1-standard.log)
- add host name and uname to Subject line of build failure mails
(e.g. "Subject: build(Linux): mysql-4.1-standard compilation failed")
the --password parameter (better add it to my.cnf instead) (Thanks to
Christian Hammers for the suggestion)
scripts/mysqlhotcopy.sh:
- added word of warning about providing the password as cleartext using
the --password parameter (better add it to my.cnf instead)
address to report bugs to bugs@mysql.com (thanks to Christian Hammers
for pointing this out) - please merge this into all other trees!
scripts/mysqlaccess.sh:
- Yves mail address does not seem to be valid anymore - changed mail
address to report bugs to bugs@mysql.com (thanks to Christian Hammers
for pointing this out)
Backport from 4.0: fix potential bug if store_lock is called with TL_IGNORE in the middle of a query
sql/ha_innobase.cc:
Backport from 4.0: fix potential bug if store_lock is called with TL_IGNORE in the middle of a query