Fix for bug#13479 "REPLACE activates UPDATE trigger, not the DELETE and

INSERT triggers".

In cases when REPLACE was internally executed via update and table had
on update (on delete) triggers defined we exposed the fact that such
optimization used by callng on update (not calling on delete) triggers.
Such behavior contradicts our documentation which describes REPLACE as
INSERT with optional DELETE.

This fix just disables this optimization for tables with on delete triggers.
The optimization is still applied for tables which have on update but have
no on delete triggers, we just don't invoke on update triggers in this case
and thus don't expose information about optimization to user.

Also added test coverage for values returned by ROW_COUNT() function (and
thus for values returned by mysql_affected_rows()) for various forms of
INSERT.

mysql-test/r/insert.result

@@ -305,3 +305,29 @@ insert delayed into v1 values (1);
 ERROR HY000: 'test.v1' is not BASE TABLE
 drop table t1;
 drop view  v1;
+create table t1 (id int primary key, data int);
+insert into t1 values (1, 1), (2, 2), (3, 3);
+select row_count();
+row_count()
+3
+insert ignore into t1 values (1, 1);
+select row_count();
+row_count()
+0
+replace into t1 values (1, 11);
+select row_count();
+row_count()
+2
+replace into t1 values (4, 4);
+select row_count();
+row_count()
+1
+insert into t1 values (2, 2) on duplicate key update data= data + 10;
+select row_count();
+row_count()
+2
+insert into t1 values (5, 5) on duplicate key update data= data + 10;
+select row_count();
+row_count()
+1
+drop table t1;

mysql-test/r/trigger.result

@@ -169,21 +169,22 @@ select @log;
 @log
 (BEFORE_INSERT: new=(id=1, data=2))
 set @log:= "";
-replace t1 values (1, 3), (2, 2);
+insert into t1 (id, data) values (1, 3), (2, 2) on duplicate key update data= data + 1;
 select @log;
 @log
-(BEFORE_INSERT: new=(id=1, data=3))(BEFORE_UPDATE: old=(id=1, data=1) new=(id=1, data=3))(AFTER_UPDATE: old=(id=1, data=1) new=(id=1, data=3))(BEFORE_INSERT: new=(id=2, data=2))(AFTER_INSERT: new=(id=2, data=2))
-alter table t1 add ts timestamp default now();
+(BEFORE_INSERT: new=(id=1, data=3))(BEFORE_UPDATE: old=(id=1, data=1) new=(id=1, data=2))(AFTER_UPDATE: old=(id=1, data=1) new=(id=1, data=2))(BEFORE_INSERT: new=(id=2, data=2))(AFTER_INSERT: new=(id=2, data=2))
 set @log:= "";
-replace t1 (id, data) values (1, 4);
+replace t1 values (1, 4), (3, 3);
 select @log;
 @log
-(BEFORE_INSERT: new=(id=1, data=4))(BEFORE_DELETE: old=(id=1, data=3))(AFTER_DELETE: old=(id=1, data=3))(AFTER_INSERT: new=(id=1, data=4))
+(BEFORE_INSERT: new=(id=1, data=4))(BEFORE_DELETE: old=(id=1, data=2))(AFTER_DELETE: old=(id=1, data=2))(AFTER_INSERT: new=(id=1, data=4))(BEFORE_INSERT: new=(id=3, data=3))(AFTER_INSERT: new=(id=3, data=3))
+drop trigger t1_bd;
+drop trigger t1_ad;
 set @log:= "";
-insert into t1 (id, data) values (1, 5), (3, 3) on duplicate key update data= data + 2;
+replace t1 values (1, 5);
 select @log;
 @log
-(BEFORE_INSERT: new=(id=1, data=5))(BEFORE_UPDATE: old=(id=1, data=4) new=(id=1, data=6))(AFTER_UPDATE: old=(id=1, data=4) new=(id=1, data=6))(BEFORE_INSERT: new=(id=3, data=3))(AFTER_INSERT: new=(id=3, data=3))
+(BEFORE_INSERT: new=(id=1, data=5))(AFTER_INSERT: new=(id=1, data=5))
 drop table t1;
 create table t1 (id int primary key, data varchar(10), fk int);
 create table t2 (event varchar(100));
@@ -493,15 +494,9 @@ select * from t1;
 i	k
 3	13
 replace into t1 values (3, 3);
-ERROR 42S22: Unknown column 'at' in 'NEW'
-select * from t1;
-i	k
-3	3
-alter table t1 add ts timestamp default now();
-replace into t1 (i, k) values (3, 13);
 ERROR 42S22: Unknown column 'at' in 'OLD'
 select * from t1;
-i	k	ts
+i	k
 drop table t1, t2;
 create table t1 (i int, bt int, k int, key(k)) engine=myisam;
 create table t2 (i int);
@@ -574,18 +569,11 @@ i	k
 1	1
 2	2
 replace into t1 values (2, 4);
-ERROR 42S22: Unknown column 'bt' in 'NEW'
+ERROR 42S22: Unknown column 'bt' in 'OLD'
 select * from t1;
 i	k
 1	1
 2	2
-alter table t1 add ts timestamp default now();
-replace into t1 (i, k) values (2, 11);
-ERROR 42S22: Unknown column 'bt' in 'OLD'
-select * from t1;
-i	k	ts
-1	1	0000-00-00 00:00:00
-2	2	0000-00-00 00:00:00
 drop table t1, t2;
 drop function if exists bug5893;
 create table t1 (col1 int, col2 int);

mysql-test/t/insert.test

@@ -187,3 +187,26 @@ create view  v1 as select * from t1;
 insert delayed into v1 values (1);
 drop table t1;
 drop view  v1;
+
+#
+# Test for values returned by ROW_COUNT() function
+# (and thus for values returned by mysql_affected_rows())
+# for various forms of INSERT
+#
+create table t1 (id int primary key, data int);
+insert into t1 values (1, 1), (2, 2), (3, 3);
+select row_count();
+insert ignore into t1 values (1, 1);
+select row_count();
+# Reports that 2 rows are affected (1 deleted + 1 inserted)
+replace into t1 values (1, 11);
+select row_count();
+replace into t1 values (4, 4);
+select row_count();
+# Reports that 2 rows are affected. This conforms to documentation.
+# (Useful for differentiating inserts from updates).
+insert into t1 values (2, 2) on duplicate key update data= data + 10;
+select row_count();
+insert into t1 values (5, 5) on duplicate key update data= data + 10;
+select row_count();
+drop table t1;

mysql-test/t/trigger.test

@@ -185,24 +185,26 @@ select @log;
 set @log:= "";
 insert ignore t1 values (1, 2);
 select @log;
-# REPLACE: before insert trigger should be called for both records,
-#          but then for first one update will be executed (and both update
-#          triggers should fire). For second after insert trigger will be
-#          called as for usual insert
+# INSERT ... ON DUPLICATE KEY UPDATE ...
 set @log:= "";
-replace t1 values (1, 3), (2, 2);
+insert into t1 (id, data) values (1, 3), (2, 2) on duplicate key update data= data + 1;
 select @log;
-# Now let us change table in such way that REPLACE on won't be executed
-# using update.
-alter table t1 add ts timestamp default now();
+# REPLACE (also test for bug#13479 "REPLACE activates UPDATE trigger,
+#          not the DELETE and INSERT triggers")
+# We define REPLACE as INSERT which DELETEs old rows which conflict with
+# row being inserted. So for the first record in statement below we will
+# call before insert trigger, then delete will be executed (and both delete
+# triggers should fire). Finally after insert trigger will be called.
+# For the second record we will just call both on insert triggers.
 set @log:= "";
-# This REPLACE should be executed via DELETE and INSERT so proper 
-# triggers should be invoked.
-replace t1 (id, data) values (1, 4);
+replace t1 values (1, 4), (3, 3);
 select @log;
-# Finally let us test INSERT ... ON DUPLICATE KEY UPDATE ...
+# Now we will drop ON DELETE triggers to test REPLACE which is internally
+# executed via update
+drop trigger t1_bd;
+drop trigger t1_ad;
 set @log:= "";
-insert into t1 (id, data) values (1, 5), (3, 3) on duplicate key update data= data + 2; 
+replace t1 values (1, 5);
 select @log;
 
 # This also drops associated triggers
@@ -531,14 +533,11 @@ alter table t1 add primary key (i);
 --error 1054
 insert into t1 values (3, 4) on duplicate key update k= k + 10;
 select * from t1;
+# The following statement will delete old row and won't
+# insert new one since after delete trigger will fail.
 --error 1054
 replace into t1 values (3, 3);
 select * from t1;
-# Change table in such way that REPLACE will delete row
-alter table t1 add ts timestamp default now();
---error 1054
-replace into t1 (i, k) values (3, 13);
-select * from t1;
 # Also drops all triggers
 drop table t1, t2;
 
@@ -594,11 +593,6 @@ select * from t1;
 --error 1054
 replace into t1 values (2, 4);
 select * from t1;
-# Change table in such way that REPLACE will delete row
-alter table t1 add ts timestamp default now();
---error 1054
-replace into t1 (i, k) values (2, 11);
-select * from t1;
 # Also drops all triggers
 drop table t1, t2;
 

sql/sql_insert.cc

@@ -1057,16 +1057,19 @@ int write_record(THD *thd, TABLE *table,COPY_INFO *info)
 	  to convert the latter operation internally to an UPDATE.
           We also should not perform this conversion if we have 
           timestamp field with ON UPDATE which is different from DEFAULT.
+          Another case when conversion should not be performed is when
+          we have ON DELETE trigger on table so user may notice that
+          we cheat here. Note that it is ok to do such conversion for
+          tables which have ON UPDATE but have no ON DELETE triggers,
+          we just should not expose this fact to users by invoking
+          ON UPDATE triggers.
 	*/
 	if (last_uniq_key(table,key_nr) &&
 	    !table->file->referenced_by_foreign_key() &&
             (table->timestamp_field_type == TIMESTAMP_NO_AUTO_SET ||
-             table->timestamp_field_type == TIMESTAMP_AUTO_SET_ON_BOTH))
+             table->timestamp_field_type == TIMESTAMP_AUTO_SET_ON_BOTH) &&
+            (!table->triggers || !table->triggers->has_delete_triggers()))
         {
-          if (table->triggers &&
-              table->triggers->process_triggers(thd, TRG_EVENT_UPDATE,
-                                                TRG_ACTION_BEFORE, TRUE))
-            goto before_trg_err;
           if (thd->clear_next_insert_id)
           {
             /* Reset auto-increment cacheing if we do an update */
@@ -1077,13 +1080,11 @@ int write_record(THD *thd, TABLE *table,COPY_INFO *info)
 					     table->record[0])))
             goto err;
           info->deleted++;
-          trg_error= (table->triggers &&
-                      table->triggers->process_triggers(thd, TRG_EVENT_UPDATE,
-                                                        TRG_ACTION_AFTER,
-                                                        TRUE));
-          /* Update logfile and count */
-          info->copied++;
-          goto ok_or_after_trg_err;
+          /*
+            Since we pretend that we have done insert we should call
+            its after triggers.
+          */
+          goto after_trg_n_copied_inc;
         }
         else
         {
@@ -1107,10 +1108,6 @@ int write_record(THD *thd, TABLE *table,COPY_INFO *info)
         }
       }
     }
-    info->copied++;
-    trg_error= (table->triggers &&
-                table->triggers->process_triggers(thd, TRG_EVENT_INSERT,
-                                                  TRG_ACTION_AFTER, TRUE));
   }
   else if ((error=table->file->write_row(table->record[0])))
   {
@@ -1118,14 +1115,14 @@ int write_record(THD *thd, TABLE *table,COPY_INFO *info)
 	(error != HA_ERR_FOUND_DUPP_KEY && error != HA_ERR_FOUND_DUPP_UNIQUE))
       goto err;
     table->file->restore_auto_increment();
+    goto ok_or_after_trg_err;
   }
-  else
-  {
+
+after_trg_n_copied_inc:
   info->copied++;
   trg_error= (table->triggers &&
               table->triggers->process_triggers(thd, TRG_EVENT_INSERT,
                                                 TRG_ACTION_AFTER, TRUE));
-  }
 
 ok_or_after_trg_err:
   if (key)