Add macros for looping through a List without a ListCell.

nathan-bossart · nathan-bossart · commit 14dd0f27d7cd · 2024-01-04T16:09:34.000-06:00
Many foreach loops only use the ListCell pointer to retrieve the content of the cell, like so: ListCell *lc; foreach(lc, mylist) { int myint = lfirst_int(lc); ... } This commit adds a few convenience macros that automatically declare the loop variable and retrieve the current cell's contents. This allows us to rewrite the previous loop like this: foreach_int(myint, mylist) { ... } This commit also adjusts a few existing loops in order to add coverage for the new/adjusted macros. There is presently no plan to bulk update all foreach loops, as that could introduce a significant amount of back-patching pain. Instead, these macros are primarily intended for use in new code. Author: Jelte Fennema-Nio Reviewed-by: David Rowley, Alvaro Herrera, Vignesh C, Tom Lane Discussion: https://postgr.es/m/CAGECzQSwXKnxGwW1_Q5JE%2B8Ja20kyAbhBHO04vVrQsLcDciwXA%40mail.gmail.com
diff --git a/src/backend/executor/execExpr.c b/src/backend/executor/execExpr.c
@@ -216,7 +216,6 @@ ExecInitQual(List *qual, PlanState *parent)
 	ExprState  *state;
 	ExprEvalStep scratch = {0};
 	List	   *adjust_jumps = NIL;
-	ListCell   *lc;
 
 	/* short-circuit (here and in ExecQual) for empty restriction list */
 	if (qual == NIL)
@@ -250,10 +249,8 @@ ExecInitQual(List *qual, PlanState *parent)
 	scratch.resvalue = &state->resvalue;
 	scratch.resnull = &state->resnull;
 
-	foreach(lc, qual)
+	foreach_ptr(Expr, node, qual)
 	{
-		Expr	   *node = (Expr *) lfirst(lc);
-
 		/* first evaluate expression */
 		ExecInitExprRec(node, state, &state->resvalue, &state->resnull);
 
@@ -265,9 +262,9 @@ ExecInitQual(List *qual, PlanState *parent)
 	}
 
 	/* adjust jump targets */
-	foreach(lc, adjust_jumps)
+	foreach_int(jump, adjust_jumps)
 	{
-		ExprEvalStep *as = &state->steps[lfirst_int(lc)];
+		ExprEvalStep *as = &state->steps[jump];
 
 		Assert(as->opcode == EEOP_QUAL);
 		Assert(as->d.qualexpr.jumpdone == -1);
diff --git a/src/backend/replication/logical/relation.c b/src/backend/replication/logical/relation.c
@@ -746,11 +746,9 @@ static Oid
 FindUsableIndexForReplicaIdentityFull(Relation localrel, AttrMap *attrmap)
 {
 	List	   *idxlist = RelationGetIndexList(localrel);
-	ListCell   *lc;
 
-	foreach(lc, idxlist)
+	foreach_oid(idxoid, idxlist)
 	{
-		Oid			idxoid = lfirst_oid(lc);
 		bool		isUsableIdx;
 		Relation	idxRel;
 		IndexInfo  *idxInfo;
diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c
@@ -1036,11 +1036,11 @@ fetch_remote_table_info(char *nspname, char *relname,
 
 		/* Build the pubname list. */
 		initStringInfo(&pub_names);
-		foreach(lc, MySubscription->publications)
+		foreach_node(String, pubstr, MySubscription->publications)
 		{
-			char	   *pubname = strVal(lfirst(lc));
+			char	   *pubname = strVal(pubstr);
 
-			if (foreach_current_index(lc) > 0)
+			if (foreach_current_index(pubstr) > 0)
 				appendStringInfoString(&pub_names, ", ");
 
 			appendStringInfoString(&pub_names, quote_literal_cstr(pubname));
diff --git a/src/backend/replication/pgoutput/pgoutput.c b/src/backend/replication/pgoutput/pgoutput.c
@@ -2234,7 +2234,6 @@ cleanup_rel_sync_cache(TransactionId xid, bool is_commit)
 {
 	HASH_SEQ_STATUS hash_seq;
 	RelationSyncEntry *entry;
-	ListCell   *lc;
 
 	Assert(RelationSyncCache != NULL);
 
@@ -2247,15 +2246,15 @@ cleanup_rel_sync_cache(TransactionId xid, bool is_commit)
 		 * corresponding schema and we don't need to send it unless there is
 		 * any invalidation for that relation.
 		 */
-		foreach(lc, entry->streamed_txns)
+		foreach_xid(streamed_txn, entry->streamed_txns)
 		{
-			if (xid == lfirst_xid(lc))
+			if (xid == streamed_txn)
 			{
 				if (is_commit)
 					entry->schema_sent = true;
 
 				entry->streamed_txns =
-					foreach_delete_current(entry->streamed_txns, lc);
+					foreach_delete_current(entry->streamed_txns, streamed_txn);
 				break;
 			}
 		}
diff --git a/src/include/nodes/pg_list.h b/src/include/nodes/pg_list.h
@@ -381,26 +381,26 @@ lnext(const List *l, const ListCell *c)
 /*
  * foreach_delete_current -
  *	  delete the current list element from the List associated with a
- *	  surrounding foreach() loop, returning the new List pointer.
+ *	  surrounding foreach() or foreach_*() loop, returning the new List
+ *	  pointer; pass the name of the iterator variable.
  *
- * This is equivalent to list_delete_cell(), but it also adjusts the foreach
- * loop's state so that no list elements will be missed.  Do not delete
- * elements from an active foreach loop's list in any other way!
+ * This is similar to list_delete_cell(), but it also adjusts the loop's state
+ * so that no list elements will be missed.  Do not delete elements from an
+ * active foreach or foreach_* loop's list in any other way!
  */
-#define foreach_delete_current(lst, cell)	\
-	(cell##__state.i--, \
-	 (List *) (cell##__state.l = list_delete_cell(lst, cell)))
+#define foreach_delete_current(lst, var_or_cell)	\
+	((List *) (var_or_cell##__state.l = list_delete_nth_cell(lst, var_or_cell##__state.i--)))
 
 /*
  * foreach_current_index -
- *	  get the zero-based list index of a surrounding foreach() loop's
- *	  current element; pass the name of the "ListCell *" iterator variable.
+ *	  get the zero-based list index of a surrounding foreach() or foreach_*()
+ *	  loop's current element; pass the name of the iterator variable.
  *
  * Beware of using this after foreach_delete_current(); the value will be
  * out of sync for the rest of the current loop iteration.  Anyway, since
  * you just deleted the current element, the value is pretty meaningless.
  */
-#define foreach_current_index(cell)  (cell##__state.i)
+#define foreach_current_index(var_or_cell)  (var_or_cell##__state.i)
 
 /*
  * for_each_from -
@@ -452,6 +452,57 @@ for_each_cell_setup(const List *lst, const ListCell *initcell)
 	return r;
 }
 
+/*
+ * Convenience macros that loop through a list without needing a separate
+ * "ListCell *" variable.  Instead, the macros declare a locally-scoped loop
+ * variable with the provided name and the appropriate type.
+ *
+ * Since the variable is scoped to the loop, it's not possible to detect an
+ * early break by checking its value after the loop completes, as is common
+ * practice.  If you need to do this, you can either use foreach() instead or
+ * manually track early breaks with a separate variable declared outside of the
+ * loop.
+ *
+ * Note that the caveats described in the comment above the foreach() macro
+ * also apply to these convenience macros.
+ */
+#define foreach_ptr(type, var, lst) foreach_internal(type, *, var, lst, lfirst)
+#define foreach_int(var, lst)	foreach_internal(int, , var, lst, lfirst_int)
+#define foreach_oid(var, lst)	foreach_internal(Oid, , var, lst, lfirst_oid)
+#define foreach_xid(var, lst)	foreach_internal(TransactionId, , var, lst, lfirst_xid)
+
+/*
+ * The internal implementation of the above macros.  Do not use directly.
+ *
+ * This macro actually generates two loops in order to declare two variables of
+ * different types.  The outer loop only iterates once, so we expect optimizing
+ * compilers will unroll it, thereby optimizing it away.
+ */
+#define foreach_internal(type, pointer, var, lst, func) \
+	for (type pointer var = 0, pointer var##__outerloop = (type pointer) 1; \
+		 var##__outerloop; \
+		 var##__outerloop = 0) \
+		for (ForEachState var##__state = {(lst), 0}; \
+			 (var##__state.l != NIL && \
+			  var##__state.i < var##__state.l->length && \
+			 (var = func(&var##__state.l->elements[var##__state.i]), true)); \
+			 var##__state.i++)
+
+/*
+ * foreach_node -
+ *	  The same as foreach_ptr, but asserts that the element is of the specified
+ *	  node type.
+ */
+#define foreach_node(type, var, lst) \
+	for (type * var = 0, *var##__outerloop = (type *) 1; \
+		 var##__outerloop; \
+		 var##__outerloop = 0) \
+		for (ForEachState var##__state = {(lst), 0}; \
+			 (var##__state.l != NIL && \
+			  var##__state.i < var##__state.l->length && \
+			 (var = lfirst_node(type, &var##__state.l->elements[var##__state.i]), true)); \
+			 var##__state.i++)
+
 /*
  * forboth -
  *	  a convenience macro for advancing through two linked lists

Original file line number	Diff line number	Diff line change
`@@ -746,11 +746,9 @@ static Oid`
`746`	`746`	`FindUsableIndexForReplicaIdentityFull(Relation localrel, AttrMap *attrmap)`
`747`	`747`	`{`
`748`	`748`	`List *idxlist = RelationGetIndexList(localrel);`
`749`		`- ListCell *lc;`
`750`	`749`
`751`		`- foreach(lc, idxlist)`
	`750`	`+ foreach_oid(idxoid, idxlist)`
`752`	`751`	`{`
`753`		`- Oid idxoid = lfirst_oid(lc);`
`754`	`752`	`bool isUsableIdx;`
`755`	`753`	`Relation idxRel;`
`756`	`754`	`IndexInfo *idxInfo;`
Original file line number	Diff line number	Diff line change
`@@ -2234,7 +2234,6 @@ cleanup_rel_sync_cache(TransactionId xid, bool is_commit)`
`2234`	`2234`	`{`
`2235`	`2235`	`HASH_SEQ_STATUS hash_seq;`
`2236`	`2236`	`RelationSyncEntry *entry;`
`2237`		`- ListCell *lc;`
`2238`	`2237`
`2239`	`2238`	`Assert(RelationSyncCache != NULL);`
`2240`	`2239`
`@@ -2247,15 +2246,15 @@ cleanup_rel_sync_cache(TransactionId xid, bool is_commit)`
`2247`	`2246`	`* corresponding schema and we don't need to send it unless there is`
`2248`	`2247`	`* any invalidation for that relation.`
`2249`	`2248`	`*/`
`2250`		`- foreach(lc, entry->streamed_txns)`
	`2249`	`+ foreach_xid(streamed_txn, entry->streamed_txns)`
`2251`	`2250`	`{`
`2252`		`- if (xid == lfirst_xid(lc))`
	`2251`	`+ if (xid == streamed_txn)`
`2253`	`2252`	`{`
`2254`	`2253`	`if (is_commit)`
`2255`	`2254`	`entry->schema_sent = true;`
`2256`	`2255`
`2257`	`2256`	`entry->streamed_txns =`
`2258`		`- foreach_delete_current(entry->streamed_txns, lc);`
	`2257`	`+ foreach_delete_current(entry->streamed_txns, streamed_txn);`
`2259`	`2258`	`break;`
`2260`	`2259`	`}`
`2261`	`2260`	`}`