The bug was caused by error in hash calculation function: it
always returned hash value for last field in a composite key, so
for keys like (a text, b char(1)) we were always
getting bad hash values.
myisam/mi_unique.c:
Fix for bug #3904:
We should take into account existing hash value when calculating hash for
next key in a composite unique index.
binlog even if they changed nothing, and a test for this.
This is useful when users use these commands to clean up their master and slave by issuing
one command on master (assume master and slave have slightly different data for some
reason and you want to clean up both).
Note that I have not changed multi-table DELETE and multi-table UPDATE because their
error-reporting mechanism is more complicated.
mysql-test/r/mysqlbinlog.result:
result update
mysql-test/r/rpl_charset.result:
result update
mysql-test/r/rpl_flush_log_loop.result:
result update
mysql-test/r/rpl_replicate_do.result:
result update
mysql-test/r/rpl_temporary.result:
result update
mysql-test/t/mysqlbinlog.test:
moving SET TIMESTAMP up as DROP shows up in binlog
sql/sql_db.cc:
DROP DATABASE IF EXISTS is now always logged to binlog, even if db did not exist
sql/sql_delete.cc:
DELETE FROM t is now always logged to binlog even if no rows deleted (but in this case, only if really no error).
sql/sql_table.cc:
DROP TABLE IF EXISTS is now always logged to binlog even if table did not exist
sql/sql_update.cc:
UPDATE is now always logged to binlog even if no rows updated (but in this case, only if really no error).
is broken (prepared statements)": fixed date handling in many places
of prepared statements code.
libmysql/libmysql.c:
Fix for Bug#4026:
- now buffer_length is defined for any buffer type. Network buffer
preallocation cleaned up.
- added constants for maximum buffer sizes necessary for MYSQL_TYPE_DATE,
MYSQL_TYPE_TIME, MYSQL_TYPE_DATETIME types.
- TIME/DATETIME packing/unpacking functions fixed
- now result set metadata is always updated from fields sent to COM_EXECUTE.
This is necessary to make 'SELECT ?' queries work without conversions.
sql/item.cc:
- added implementatoin of Item_param::get_date
sql/item.h:
- added enum_field_types Item_param::param_type. First step for proper
handling of placeholders.
- added get_date() implementation to prevent date -> string -> date
conversions when MYSQL_TYPE_DATE/DATETIME parameter is used in temporal
context.
sql/protocol.cc:
Fix for Bug#4026:
- PACKET_BUFFET_EXTRA_ALLOC -> PACKET_BUFFER_EXTRA_ALLOC.
The define itself was moved to .cc as it's used only in protocol.cc
- fixed Protocol_prep::store_time() call.
sql/protocol.h:
- PACKET_BUFFER_EXTRA_ALLOC moved to protocol.cc
sql/sql_prepare.cc:
Fix for Bug#4026:
- MYSQL_TYPE_TIME/DATETIME handling fixed.
- added initialization for Item_param::param_type in
setup_one_conversion_function
tests/client_test.c:
Test case for Bug#4026
In mysqlbinlog, there was a problem with how we escaped the content of a string user variable.
To be perfect, we should have escaped with character_set_client. But this charset is unknown
to mysqlbinlog. So the simplest is to print the string in hex. This is unreadable but
100% safe with any charset (checked with Bar), no more need to bother with character_set_client.
mysql-test/r/rpl_charset.result:
hex strings
mysql-test/r/rpl_user_variables.result:
hex strings
mysql-test/r/user_var.result:
hex strings
sql/log_event.cc:
In mysqlbinlog, there was a problem with how we escaped the content of a string user variable.
To be perfect, we should have escaped with character_set_client. But this charset is unknown
to mysqlbinlog. So the simplest is to print the string in hex. This is unreadable but
100% safe with any charset (checked with Bar), no more need to bother with character_set_client.
Fix Bug #4047: remove the improvement ported from 4.0 that made InnoDB to remember the original select_lock_type inside LOCK TABLES
sql/ha_innodb.cc:
Fix Bug #4047: remove the improvement ported from 4.0 that made InnoDB to remember the original select_lock_type inside LOCK TABLES
mysql-test/r/ctype_ucs.result:
Test for bug#3946: Error in LPAD() when padstring is longer than 1 character
mysql-test/t/ctype_ucs.test:
Test for bug#3946: Error in LPAD() when padstring is longer than 1 character
WL#1160.
Adding variable-conformant startup options for --default-character-set and --default-collation
sql/mysqld.cc:
WL#1160.
Adding variable-conformant startup options for --default-character-set and --default-collation
Dump could fail to load because of --default-character-set command line option.
More safe dump is now produces, --default-character-set doesn't matter.
client/mysqldump.c:
Dump could fail to load because of --default-character-set command line option.
More safe dump is now produces, --default-character-set doesn't matter.
A couple of small changes to ha_example (mainly comments).
sql/examples/ha_archive.cc:
Changes from Serg.
A few styles changes, fix for potential hash insert gone wrong, removed bad tables if create fails, gzflush used for flushing by default (had problems with OSX 10.2), and some spelling corrections.
sql/examples/ha_archive.h:
Removed fast_key_read() and changed the name of a conflicting variable.
sql/examples/ha_example.cc:
Added a few new comments.
sql/examples/ha_example.h:
Added pragma interface. Removed fast_key_read() and Serg's request
