BUG#21065746 RQG_PARTN_PRUNING_VALGRIND FAILED IN REM0REC.CC

Problem :
---------
This is a regression of Bug#19138298. In purge_node_t::validate_pcur
we are trying to get offsets for all columns of clustered index from
stored record in persistent cursor. This would fail when stored record
is not having all fields of the index. The stored record stores only
fields that are needed to uniquely identify the entry.

Solution :
----------
1. Use pcur.old_n_fields to get fields that are stored
2. Add comment to note dependency between stored fields in purge node
ref and stored cursor.
3. Return if the cursor record is not already stored as it is not safe
to access cursor record directly without latch.
Reviewed-by: Marko Makela <marko.makela@oracle.com>

RB: 9139

storage/innobase/row/row0purge.c

@@ -823,12 +823,12 @@ Validate the persisent cursor in the purge node. The purge node has two
 references to the clustered index record - one via the ref member, and the
 other via the persistent cursor.  These two references must match each
 other if the found_clust flag is set.
-@return true if the persistent cursor is consistent with the ref member.*/
+@return true if the stored copy of persistent cursor is consistent
+with the ref member.*/
 ibool
 row_purge_validate_pcur(
 	purge_node_t*	node)
 {
-	const rec_t*	rec ;
 	dict_index_t*	clust_index;
 	ulint*		offsets;
 	int		st;
@@ -841,23 +841,25 @@ row_purge_validate_pcur(
 		return(TRUE);
 	}
 
-	clust_index = node->pcur.btr_cur.index;
-
-	if (node->pcur.old_stored == BTR_PCUR_OLD_STORED) {
-		rec = node->pcur.old_rec;
-	} else {
-		rec = btr_pcur_get_rec(&node->pcur);
+	if (node->pcur.old_stored != BTR_PCUR_OLD_STORED) {
+		return(TRUE);
 	}
 
-	offsets = rec_get_offsets(rec,
-		clust_index, NULL, ULINT_UNDEFINED, &node->heap);
+	clust_index = node->pcur.btr_cur.index;
+
+	offsets = rec_get_offsets(node->pcur.old_rec, clust_index, NULL,
+				  node->pcur.old_n_fields, &node->heap);
 
-	st = cmp_dtuple_rec(node->ref, rec, offsets);
+	/* Here we are comparing the purge ref record and the stored initial
+	part in persistent cursor. Both cases we store n_uniq fields of the
+	cluster index and so it is fine to do the comparison. We note this
+	dependency here as pcur and ref belong to different modules. */
+	st = cmp_dtuple_rec(node->ref, node->pcur.old_rec, offsets);
 
 	if (st != 0) {
 		fprintf(stderr, "Purge node pcur validation failed\n");
 		dtuple_print(stderr, node->ref);
-		rec_print(stderr, rec, clust_index);
+		rec_print(stderr, node->pcur.old_rec, clust_index);
 		return(FALSE);
 	}