Fix oversight in new EvalPlanQual logic: the second loop over the ExecRowMark

list in ExecLockRows() forgot to allow for the possibility that some of the
rowmarks are for child tables that aren't relevant to the current row.
Per report from Kenichiro Tanaka.
This commit is contained in:
Tom Lane 2010-07-28 17:21:56 +00:00
parent 2dbbda02e7
commit 77c75076f3

View File

@ -8,7 +8,7 @@
*
*
* IDENTIFICATION
* $PostgreSQL: pgsql/src/backend/executor/nodeLockRows.c,v 1.5 2010/07/12 17:01:05 tgl Exp $
* $PostgreSQL: pgsql/src/backend/executor/nodeLockRows.c,v 1.6 2010/07/28 17:21:56 tgl Exp $
*
*-------------------------------------------------------------------------
*/
@ -195,6 +195,13 @@ lnext:
HeapTupleData tuple;
Buffer buffer;
/* ignore non-active child tables */
if (!ItemPointerIsValid(&(erm->curCtid)))
{
Assert(erm->rti != erm->prti); /* check it's child table */
continue;
}
if (EvalPlanQualGetTuple(&node->lr_epqstate, erm->rti) != NULL)
continue; /* it was updated and fetched above */