Address Copilot review: pre-init ExprState, add ON MATCH SET predecessor test

- Move ExecInitExpr for ON CREATE/MATCH SET items from per-row
  execution in apply_update_list() to plan initialization in
  begin_cypher_merge(). Follows the established pattern used by
  cypher_target_node (id_expr_state, prop_expr_state).
- Add prop_expr_state field to cypher_update_item with serialization
  support in outfuncs/readfuncs/copyfuncs.
- apply_update_list() uses pre-initialized state when available,
  falls back to per-row init for plain SET callers.
- Fix misleading comment: "ON MATCH SET" → "ON CREATE SET" for Case 1
  first-run test.
- Add Case 1 second-run test that triggers ON MATCH SET with a
  predecessor clause (MATCH ... MERGE ... ON MATCH SET).

Author: gregfelice

.gitignore

@@ -15,3 +15,4 @@ __pycache__
 **/apache_age_python.egg-info
 
 drivers/python/build
+*.bc

regress/expected/cypher_merge.out

@@ -1944,7 +1944,7 @@ $$) AS (name agtype, created agtype, matched agtype);
  "Bob" | true    | true
 (1 row)
 
--- ON MATCH SET with MERGE after MATCH (Case 1: has predecessor)
+-- ON CREATE SET with MERGE after MATCH (Case 1: has predecessor, first run = create)
 SELECT * FROM cypher('merge_actions', $$
   MATCH (a:Person {name: 'Alice'})
   MERGE (a)-[:KNOWS]->(b:Person {name: 'Charlie'})
@@ -1956,6 +1956,18 @@ $$) AS (a agtype, b agtype, source agtype);
  "Alice" | "Charlie" | "merge_create"
 (1 row)
 
+-- ON MATCH SET with MERGE after MATCH (Case 1: has predecessor, second run = match)
+SELECT * FROM cypher('merge_actions', $$
+  MATCH (a:Person {name: 'Alice'})
+  MERGE (a)-[:KNOWS]->(b:Person {name: 'Charlie'})
+    ON MATCH SET b.visited = true
+  RETURN a.name, b.name, b.visited
+$$) AS (a agtype, b agtype, visited agtype);
+    a    |     b     | visited 
+---------+-----------+---------
+ "Alice" | "Charlie" | true
+(1 row)
+
 -- Multiple SET items in a single ON CREATE SET
 SELECT * FROM cypher('merge_actions', $$
   MERGE (n:Person {name: 'Dave'})

regress/sql/cypher_merge.sql

@@ -903,14 +903,22 @@ SELECT * FROM cypher('merge_actions', $$
   RETURN n.name, n.created, n.matched
 $$) AS (name agtype, created agtype, matched agtype);
 
--- ON MATCH SET with MERGE after MATCH (Case 1: has predecessor)
+-- ON CREATE SET with MERGE after MATCH (Case 1: has predecessor, first run = create)
 SELECT * FROM cypher('merge_actions', $$
   MATCH (a:Person {name: 'Alice'})
   MERGE (a)-[:KNOWS]->(b:Person {name: 'Charlie'})
     ON CREATE SET b.source = 'merge_create'
   RETURN a.name, b.name, b.source
 $$) AS (a agtype, b agtype, source agtype);
 
+-- ON MATCH SET with MERGE after MATCH (Case 1: has predecessor, second run = match)
+SELECT * FROM cypher('merge_actions', $$
+  MATCH (a:Person {name: 'Alice'})
+  MERGE (a)-[:KNOWS]->(b:Person {name: 'Charlie'})
+    ON MATCH SET b.visited = true
+  RETURN a.name, b.name, b.visited
+$$) AS (a agtype, b agtype, visited agtype);
+
 -- Multiple SET items in a single ON CREATE SET
 SELECT * FROM cypher('merge_actions', $$
   MERGE (n:Person {name: 'Dave'})

src/backend/executor/cypher_merge.c

@@ -191,6 +191,35 @@ static void begin_cypher_merge(CustomScanState *node, EState *estate,
         }
     }
 
+    /*
+     * Pre-initialize ExprStates for ON CREATE SET / ON MATCH SET items.
+     * This must happen once at plan init time, not per-row.
+     */
+    if (css->on_create_set_info != NULL)
+    {
+        foreach(lc, css->on_create_set_info->set_items)
+        {
+            cypher_update_item *item = (cypher_update_item *)lfirst(lc);
+            if (item->prop_expr != NULL)
+            {
+                item->prop_expr_state = ExecInitExpr(
+                    (Expr *)item->prop_expr, (PlanState *)node);
+            }
+        }
+    }
+    if (css->on_match_set_info != NULL)
+    {
+        foreach(lc, css->on_match_set_info->set_items)
+        {
+            cypher_update_item *item = (cypher_update_item *)lfirst(lc);
+            if (item->prop_expr != NULL)
+            {
+                item->prop_expr_state = ExecInitExpr(
+                    (Expr *)item->prop_expr, (PlanState *)node);
+            }
+        }
+    }
+
     /*
      * Postgres does not assign the es_output_cid in queries that do
      * not write to disk, ie: SELECT commands. We need the command id

src/backend/executor/cypher_set.c

@@ -504,8 +504,20 @@ void apply_update_list(CustomScanState *node,
             Datum val;
             bool isnull;
 
-            expr_state = ExecInitExpr((Expr *)update_item->prop_expr,
-                                      (PlanState *)node);
+            /*
+             * Use the pre-initialized ExprState if available (set during
+             * plan init in begin_cypher_merge). Fall back to per-row init
+             * for callers that haven't pre-initialized (e.g. plain SET).
+             */
+            if (update_item->prop_expr_state != NULL)
+            {
+                expr_state = update_item->prop_expr_state;
+            }
+            else
+            {
+                expr_state = ExecInitExpr((Expr *)update_item->prop_expr,
+                                          (PlanState *)node);
+            }
             val = ExecEvalExpr(expr_state, econtext, &isnull);
             remove_property = isnull;
 

src/backend/nodes/cypher_copyfuncs.c

@@ -137,6 +137,7 @@ void copy_cypher_update_item(ExtensibleNode *newnode, const ExtensibleNode *from
     COPY_SCALAR_FIELD(remove_item);
     COPY_SCALAR_FIELD(is_add);
     COPY_NODE_FIELD(prop_expr);
+    COPY_NODE_FIELD(prop_expr_state);
 }
 
 /* copy function for cypher_delete_information */

src/backend/nodes/cypher_outfuncs.c

@@ -430,6 +430,7 @@ void out_cypher_update_item(StringInfo str, const ExtensibleNode *node)
     WRITE_BOOL_FIELD(remove_item);
     WRITE_BOOL_FIELD(is_add);
     WRITE_NODE_FIELD(prop_expr);
+    WRITE_NODE_FIELD(prop_expr_state);
 }
 
 /* serialization function for the cypher_delete_information ExtensibleNode. */

src/backend/nodes/cypher_readfuncs.c

@@ -270,6 +270,7 @@ void read_cypher_update_item(struct ExtensibleNode *node)
     READ_BOOL_FIELD(remove_item);
     READ_BOOL_FIELD(is_add);
     READ_NODE_FIELD(prop_expr);
+    READ_NODE_FIELD(prop_expr_state);
 }
 
 /*

src/include/nodes/cypher_nodes.h

@@ -455,6 +455,7 @@ typedef struct cypher_update_item
     bool is_add;
     Node *prop_expr;    /* SET value expression, used by MERGE ON CREATE/MATCH SET
                          * where the expression is not in the plan's target list */
+    ExprState *prop_expr_state;  /* initialized at plan init, not per-row */
 } cypher_update_item;
 
 typedef struct cypher_delete_information