Disallow whole-row variables in GENERATED expressions.

tglsfdc · tglsfdc · commit 4b10074453d1 · 2021-05-21T15:12:08.000-04:00
This was previously allowed, but I think that was just an oversight. It's a clear violation of the rule that a generated column cannot depend on itself or other generated columns. Moreover, because the code was relying on the assumption that no such cross-references exist, it was pretty easy to crash ALTER TABLE and perhaps other places. Even if you managed not to crash, you got quite unstable, implementation-dependent results. Per report from Vitaly Ustinov. Back-patch to v12 where GENERATED came in. Discussion: https://postgr.es/m/CAM_DEiWR2DPT6U4xb-Ehigozzd3n3G37ZB1+867zbsEVtYoJww@mail.gmail.com
diff --git a/src/backend/catalog/heap.c b/src/backend/catalog/heap.c
@@ -3020,15 +3020,26 @@ check_nested_generated_walker(Node *node, void *context)
 		AttrNumber	attnum;
 
 		relid = rt_fetch(var->varno, pstate->p_rtable)->relid;
+		if (!OidIsValid(relid))
+			return false;		/* XXX shouldn't we raise an error? */
+
 		attnum = var->varattno;
 
-		if (OidIsValid(relid) && AttributeNumberIsValid(attnum) && get_attgenerated(relid, attnum))
+		if (attnum > 0 && get_attgenerated(relid, attnum))
 			ereport(ERROR,
-					(errcode(ERRCODE_SYNTAX_ERROR),
+					(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
 					 errmsg("cannot use generated column \"%s\" in column generation expression",
 							get_attname(relid, attnum, false)),
 					 errdetail("A generated column cannot reference another generated column."),
 					 parser_errposition(pstate, var->location)));
+		/* A whole-row Var is necessarily self-referential, so forbid it */
+		if (attnum == 0)
+			ereport(ERROR,
+					(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
+					 errmsg("cannot use whole-row variable in column generation expression"),
+					 errdetail("This would cause the generated column to depend on its own value."),
+					 parser_errposition(pstate, var->location)));
+		/* System columns were already checked in the parser */
 
 		return false;
 	}
diff --git a/src/test/regress/expected/generated.out b/src/test/regress/expected/generated.out
@@ -46,6 +46,13 @@ ERROR:  cannot use generated column "b" in column generation expression
 LINE 1: ...AYS AS (a * 2) STORED, c int GENERATED ALWAYS AS (b * 3) STO...
                                                              ^
 DETAIL:  A generated column cannot reference another generated column.
+-- a whole-row var is a self-reference on steroids, so disallow that too
+CREATE TABLE gtest_err_2c (a int PRIMARY KEY,
+    b int GENERATED ALWAYS AS (num_nulls(gtest_err_2c)) STORED);
+ERROR:  cannot use whole-row variable in column generation expression
+LINE 2:     b int GENERATED ALWAYS AS (num_nulls(gtest_err_2c)) STOR...
+                                                 ^
+DETAIL:  This would cause the generated column to depend on its own value.
 -- invalid reference
 CREATE TABLE gtest_err_3 (a int PRIMARY KEY, b int GENERATED ALWAYS AS (c * 2) STORED);
 ERROR:  column "c" does not exist
diff --git a/src/test/regress/sql/generated.sql b/src/test/regress/sql/generated.sql
@@ -17,6 +17,9 @@ CREATE TABLE gtest_err_1 (a int PRIMARY KEY, b int GENERATED ALWAYS AS (a * 2) S
 -- references to other generated columns, including self-references
 CREATE TABLE gtest_err_2a (a int PRIMARY KEY, b int GENERATED ALWAYS AS (b * 2) STORED);
 CREATE TABLE gtest_err_2b (a int PRIMARY KEY, b int GENERATED ALWAYS AS (a * 2) STORED, c int GENERATED ALWAYS AS (b * 3) STORED);
+-- a whole-row var is a self-reference on steroids, so disallow that too
+CREATE TABLE gtest_err_2c (a int PRIMARY KEY,
+    b int GENERATED ALWAYS AS (num_nulls(gtest_err_2c)) STORED);
 
 -- invalid reference
 CREATE TABLE gtest_err_3 (a int PRIMARY KEY, b int GENERATED ALWAYS AS (c * 2) STORED);
