Fix an assertion failure in 64-bit computers if UNIV_MEM_DEBUG is defined
innobase/mem/mem0pool.c:
Fix an assertion failure in 64-bit computers if UNIV_MEM_DEBUG is defined
Fixing this bugs improves performance by 40% for very small read
statements and with 12-13 % for very simple updating transactions
(flexBench -o 10000) in single threaded application. Removes a fixed
cost of around 100.000 cycles every time the ndbd process wakes up to
execute some queries.
ndb/src/kernel/vm/FastScheduler.cpp:
Integrate sendPacked into doJob
The lack of integration meant that several loops in
ipControlLoop were executed each time the ndbd process
woke up, also for reads it meant that response was divided
in two TCP/IP packets
Is necessary to integrate this with overload protection
ndb/src/kernel/vm/ThreadConfig.cpp:
Integrate sendPacked into doJob
The lack of integration meant that several loops in
ipControlLoop were executed each time the ndbd process
woke up, also for reads it meant that response was divided
in two TCP/IP packets
Is necessary to integrate this with overload protection
useful than *.old.<PID> - use the mtime of configure.in instead.
This gives us a time stamp of when this last build directory was created,
e.g. "mysql-4.0.24-build-2005-01-31-16:47"
Build-tools/Bootstrap:
- Rename already existing build directories to something more useful than
*.old.<PID> - use the mtime of configure.in instead. This gives us a time
stamp of when this last build directory was created, e.g.
"mysql-4.0.24-build-2005-01-31-16:47"
+ ion shutdown - don't close files that are not open
+ abort in debug - if closing a fd == -1
BitKeeper/etc/logging_ok:
Logging to logging@openlogging.org accepted
innobase/row/row0mysql.c:
row_truncate_for_mysql(): Always lock the data dictionary.
row_truncate_for_mysql(): Improve comments and diagnostic messages.
sql/ha_innodb.cc:
delete_all_rows(): Replace innobase_commit_low()
with innobase_commit(). (partial fix to Bug #8151)
Fix replication for multi-update
new test - rpl_multi_update2
sql/mysql_priv.h:
Bug#7011
New function mysql_multi_update_lock()
sql/sql_parse.cc:
Bug#7011
New function check_multi_update_lock()
For multi-update on slave, perform an early open&lock
sql/sql_update.cc:
Bug#7011
Split out multi-update locking into its own function,
mysql_multi_update_lock()
mysys/my_handler.c:
Fix for a bug in multi table updates, when offset of one table is
5 bytes or more. This occured because in that case binary(5) indexed
column is space packed, so it was hit by this bug in the function
for binary comparison.
Bug #7788
mysqldump.c:
Trimmed some lines to be less than 80 chars. Using just NAME_LEN now for table name buffers
client/mysqldump.c:
Trimmed some lines to be less than 80 chars. Using just NAME_LEN now for table name buffers
into mysql.com:/home/jonas/src/mysql-5.0-ndb
ndb/docs/Makefile.am:
Auto merged
ndb/src/ndbapi/NdbDictionaryImpl.cpp:
Auto merged
sql/ha_ndbcluster.cc:
Auto merged
of strings of unequal length. (Bug #8134)
mysql-test/t/compare.test:
Add new regression test
strings/ctype-simple.c:
Fix value used for swapping negative/positive values using XOR
mysql-test/r/compare.result:
Add new test result
sql/examples/ha_archive.cc:
More comments, fixed issue with lost file descriptors.
BitKeeper/etc/logging_ok:
Logging to logging@openlogging.org accepted
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
innobase/ut/ut0ut.c:
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
innobase/os/os0file.c:
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
server on Windows service startup, and not just --defaults-file=file. This
appears to be an unnecessary limitation, but fixing that will require
a larger refactoring of the Windows service installation and startup code.
(Bug #7856)
sql/mysqld.cc:
Restore behavior from 4.1 of allowing any one parameter to the service on
install, and not just --defaults-file=file.
removed usage of SIGRTMIN temorarilly as it causes problems on some platforms
ndb/include/ndbapi/ndb_opt_defaults.h:
removed usage of SIGRTMIN temorarilly as it causes problems on some platforms
BitKeeper/etc/logging_ok:
Logging to logging@openlogging.org accepted