Commit graph

3 commits

Author SHA1 Message Date
unknown
9fdbaf5195 added missing drop view at end of test 2005-11-24 09:56:48 +01:00
unknown
b323667ffc Prelocking-free SPs, post-review fixes:
* Don't activate prelocking mode for evaluating procedure arguments when it is not necessary.
* Code structure simplification and cleanup.
* Cleanup in .test files


mysql-test/r/sp-prelocking.result:
  Prelocking-free SPs, post-review fixes:
  Added comment, s/testdb/mysqltest/, fixed a wrong test (error wasnt reported because of known bug in mysqltestrun)
mysql-test/r/sp-security.result:
  Don't drop the table we're not using.
mysql-test/r/sp.result:
  Prelocking-free SPs, post-review fixes:
  remove redundant "drop table if exists t3" statements
mysql-test/t/sp-prelocking.test:
  Prelocking-free SPs, post-review fixes:
  Added comment, s/testdb/mysqltest/, fixed a wrong test (error wasnt reported because of known bug in mysqltestrun)
mysql-test/t/sp-security.test:
  Don't drop the table we're not using.
mysql-test/t/sp.test:
  Prelocking-free SPs, post-review fixes:
  remove redundant "drop table if exists t3" statements
sql/sp.cc:
  New, better defined, sp_get_prelocking_info() function to get info about
    statement prelocking options
sql/sp.h:
  Prelocking-free SPs, post-review fixes: New, better defined, sp_get_prelocking_info() 
  function to get info about statement prelocking options
sql/sp_cache.h:
  Prelocking-free SPs, post-review fixes: Amended the comments
sql/sp_head.cc:
  Prelocking-free SPs, post-review fixes: Amend the comments, simplify the code that 
  attaches removes statement's prelocking tables.
sql/sql_base.cc:
  Prelocking-free SPs, post-review fixes:
  * Use a better defined sp_get_prelocking_info() function to get info about 
    statement prelocking options
  * Don't activate prelocked mode for evaluation of SP arguments that use tables 
    but don't need prelocking.
sql/sql_class.cc:
  Prelocking-free SPs, post-review fixes: Initialize THD members in the order they are declared.
2005-08-03 03:37:32 +00:00
unknown
11abe15eab Added Non-prelocked SP execution: Now a PROCEDURE doesn't enter/leave prelocked mode for
its body, but lets each statement to get/release its own locks. This allows a broader set
of statements to be executed inside PROCEDUREs (but breaks replication)
This patch should fix BUG#8072, BUG#8766, BUG#9563, BUG#11126


mysql-test/r/sp-security.result:
  Drop tables this test attempts to create
mysql-test/r/sp-threads.result:
  Update test results
mysql-test/r/sp.result:
  Disabled a test that triggers BUG#11986, cleanup used tables when tests start.
mysql-test/r/view.result:
  Enabled a test case that now works with prelocking-free SPs
mysql-test/t/sp-security.test:
  Drop tables this test attempts to create
mysql-test/t/sp.test:
  Disabled a test that triggers BUG#11986, cleanup used tables when tests start.
mysql-test/t/view.test:
  Enabled a test case that now works with prelocking-free SPs
sql/handler.cc:
  Rename: thd->transaction.in_sub_stmt -> thd->in_sub_stmt
sql/item_func.cc:
  Rename: thd->transaction.in_sub_stmt -> thd->in_sub_stmt
sql/sp.cc:
  Non-prelocked SP execution: Added support for skipping prelocking of procedure body for
  "CALL proc(...)" statements.
sql/sp.h:
  Non-prelocked SP execution: Added support for skipping prelocking of procedure body for
  "CALL proc(...)" statements.
sql/sp_cache.h:
  Added comments
sql/sp_head.cc:
  Non-prelocked SP execution:
  * Try to unlock tables after PROCEDURE arguments have been evaluated.
  * Make sp_lex_keeper be able to execute in 2 modes: A) when already in prelocked mode
    B) when its statement enters/leaves prelocked mode itself.
sql/sp_head.h:
  Non-prelocked SP execution:  Make sp_lex_keeper to additionally keep list of tables it 
  needs to prelock when its statement enters/leaves prelocked mode on its own.
sql/sql_base.cc:
  Non-prelocked SP execution: Make open_tables() to
   * detect 'CALL proc(...)' and not to do prelocking for procedure body statements.
   * Make lex->query_tables_last to point precisely to a boundary in lex->query_tables 
     list where 'own' tables and views' tables end and added-for-prelocking tables begin.
     (it was not true before - view's tables could end up after query_tables_own_last)
sql/sql_class.cc:
  Rename: thd->transaction.in_sub_stmt -> thd->in_sub_stmt
sql/sql_class.h:
  Rename: thd->transaction.in_sub_stmt -> thd->in_sub_stmt
sql/sql_lex.cc:
  Non-prelocked SP execution: More rigourous cleanup in st_lex::cleanup_after_one_table_open()
sql/sql_parse.cc:
  Rename: thd->transaction.in_sub_stmt -> thd->in_sub_stmt, remove outdated comments
sql/sql_trigger.h:
  Rename: thd->transaction.in_sub_stmt -> thd->in_sub_stmt
2005-07-30 08:19:57 +00:00