Li-Bing, hello.
The patch is good, approved.
Just one remark on the test's comments.
> #At file:///home/anders/work/bzrwork/worktree1/mysql-5.0-bugteam/ based on
> revid:joro@stripped
>
> 2816 Li-Bing.Song@stripped 2009-10-28
> BUG#48216 Replication fails on all slaves after upgrade to 5.0.86 on master
>
> When a sessione is closed, all temporary tables of the session are
> automatically
> dropped and are binlogged. But it will be binlogged with wrong database names
> when
> the length of the temporary tables' database names are greater than the
> length of the current database name or the current database is not set.
>
> Query_log_event's db_len is forgot to set when Query_log_event's db is set.
> This patch wrote code to set db_len immediately after db has set.
>
> modified:
> mysql-test/r/drop_temp_table.result
> mysql-test/t/drop_temp_table.test
> sql/sql_base.cc
> === modified file 'mysql-test/r/drop_temp_table.result'
> --- a/mysql-test/r/drop_temp_table.result 2009-08-27 16:28:09 +0000
> +++ b/mysql-test/r/drop_temp_table.result 2009-10-28 09:42:04 +0000
> @@ -8,6 +8,7 @@ create temporary table shortn2 (a int);
> select get_lock("a",10);
> get_lock("a",10)
> 1
> +USE test;
> select get_lock("a",10);
> get_lock("a",10)
> 1
>
> === modified file 'mysql-test/t/drop_temp_table.test'
> --- a/mysql-test/t/drop_temp_table.test 2007-06-14 21:23:30 +0000
> +++ b/mysql-test/t/drop_temp_table.test 2009-10-28 09:42:04 +0000
> @@ -14,6 +14,17 @@ create temporary table shortn1 (a int);
> create temporary table `table:name` (a int);
> create temporary table shortn2 (a int);
> select get_lock("a",10);
> +
> +#
> +# BUG48216 Replication fails on all slaves after upgrade to 5.0.86 on master
> +#
> +# When the session is closed, any temporary tables of the session are dropped
> +# and are binlogged.
> # But it will be binlogged with a wrong database name when
> +# the length of the database name('drop-temp-table-test') is greater than the
> +# current database name('test').
The test's logics is clear. But not the last statement that really wants to say
what would happen without your fixes.
'will be binlogged' might sound like specifying what is expected to be seen with the
correct code.
I would remove the sentence altogether or to say that we test the bug's condition where
the actual db name's lengh is less than the current db.
> +#
> +USE test;
> +
> disconnect con1;
>
> connection con2;
>
> === modified file 'sql/sql_base.cc'
> --- a/sql/sql_base.cc 2009-08-28 15:51:31 +0000
> +++ b/sql/sql_base.cc 2009-10-28 09:42:04 +0000
> @@ -798,6 +798,7 @@ void close_temporary_tables(THD *thd)
> s_query.length() - 1 /* to remove trailing ',' */,
> 0, FALSE, THD::NOT_KILLED);
> qinfo.db= db.ptr();
> + qinfo.db_len= db.length();
> thd->variables.character_set_client= cs_save;
> DBUG_ASSERT(qinfo.error_code == 0);
> mysql_bin_log.write(&qinfo);
>
cheers,
Andrei
