Created by Oleg Tsarev on 2011-09-07 and last modified on 2011-09-16
Get this branch:
bzr branch lp:~tsarev/percona-server/rnt-5.1
Only Oleg Tsarev can upload to this branch. If you are Oleg Tsarev please log in for upload directions.

Branch merges

Related bugs

Related blueprints

Branch information

Recent revisions

197. By Oleg Tsarev on 2011-09-16

remove references to Makefile.in from mwl36.patch and wl5151.patch

196. By Oleg Tsarev on 2011-09-14

add ommited in revision 194 mysql-test/include/percona_row_based_replication_without_primary_key_delete.inc
fix bug #849849

195. By Alexey Kopytov on 2011-09-13

Bug #848703: innodb_fast_alter_column doesn't work with indexed columns

The problem was that when extending a column involved in an index,
mysql_alter_table() marked the corresponding index to be recreated
later (by adding it to both index_drop_buffer and
index_add_buffer). Such ALTER TABLE would normally be executed by
copying the table, because InnoDB would report incompatible metadata in
such a case.

However, due to innodb_fast_alter_column turned on, the metadata was
rightfully considered compatible, so the ALTER TABLE was executed
in-place, by first adding the indexes from index_add_buffer and then
dropping the ones from index_drop_buffer. Since in this case both drop
and add list contained the same index, InnoDB reported a duplicate index
name when trying to add the index. Reverting the order (i.e. dropping
before adding) would cause problems with foreign key constraints.

Fixed by changing compare_tables() so that, if innodb_fast_alter_column
is on, extending an index column does not lead to index being marked as

In addition, InnoDB no longer reports extending columns as compatible
metadata when innodb_fast_alter_column is OFF. This was the reasons why
the bug occurred even with disabled feature.

194. By Oleg Tsarev on 2011-09-12

1) move tests from mysql-test/ to patches
2) remove install_tests script and call of script from Makefile

193. By Oleg Tsarev on 2011-09-09

make regenerate

192. By Oleg Tsarev on 2011-09-09

remove mysql-test/mysqlbinlog_max_binlog_statement_size.patch/ because we have tests in the patch already

191. By Alexey Kopytov on 2011-09-08

Automerge from 5.1.

190. By Alexey Kopytov on 2011-09-08

Bug #827456: mysqlbinlog_max_binlog_statement_size.patch has a serious

The problem was that with multi-table UPDATE or DELETE queries,
splitting BINLOG statements into multiple ones did not handle multiple
table map event correctly: the second and subsequent BINLOG statement
would contain only one table map event.

Fixed by caching all table map events corresponding to the current
statement, so that BINLOG statements resulted from splitting contain
full list of table map events.

189. By Alexey Kopytov on 2011-09-08

Bug #844619: Compiler warning in rpl_record.cc in rnt-5.1

Fixed format string to avoid the warning.

188. By Oleg Tsarev on 2011-09-07

fix compiler warnings in wl47.patch and replication_slave_skip_columns.patch

Branch metadata

Branch format:
Branch format 7
Repository format:
Bazaar repository format 2a (needs bzr 1.16 or later)
Stacked on:
This branch contains Public information 
Everyone can see this information.