diff options
author | unknown <kostja@bodhi.(none)> | 2007-07-14 02:04:48 +0400 |
---|---|---|
committer | unknown <kostja@bodhi.(none)> | 2007-07-14 02:04:48 +0400 |
commit | 31ea7d042ba48bd34a457bba138d2189f844731d (patch) | |
tree | ccedd61617353a865ce05b7b2dd6643ea3cda980 /sql/sql_class.h | |
parent | 6ba23b0ac9a4548de36386b4f892af9c7d471e97 (diff) | |
download | mariadb-git-31ea7d042ba48bd34a457bba138d2189f844731d.tar.gz |
A follow up after the patch for Bug#21074 - even though
we now have exclusive name lock on the table name in mysql_rm_table_part2,
we still should keep LOCK_open - some storage engines are not
ready for locking scope change and assume that LOCK_open is kept.
Still, the binary logging and query cache invalidation calls
moved out of LOCK_open scope.
Fixes some of the broken 5.1-runtime tests (tests break on asserts).
sql/ha_ndbcluster.cc:
Do not lock LOCK_open for mysql_rm_table_part2 - it does that
for us now.
sql/mysql_priv.h:
Remove an unused flag.
sql/sql_class.h:
Fix an unrelated compiler warning.
sql/sql_db.cc:
Adjust to the changed signature.
sql/sql_table.cc:
mysql_rm_table_part2: we need to keep LOCK_open while calling
storage engine functions, even though now
we have an exclusive lock on the table name. Some of them assume that it's
kept and attempt to unlock it.
Diffstat (limited to 'sql/sql_class.h')
-rw-r--r-- | sql/sql_class.h | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/sql/sql_class.h b/sql/sql_class.h index a8d62d93b21..0dad4966623 100644 --- a/sql/sql_class.h +++ b/sql/sql_class.h @@ -1999,8 +1999,8 @@ class select_insert :public select_result_interceptor { class select_create: public select_insert { ORDER *group; TABLE_LIST *create_table; - TABLE_LIST *select_tables; HA_CREATE_INFO *create_info; + TABLE_LIST *select_tables; Alter_info *alter_info; Field **field; public: |