From: Peter Eisentraut <peter@eisentraut.org>
Date: Tue, 14 Sep 2021 07:34:50 +0000 (+0200)
Subject: Add WRITE_INDEX_ARRAY
X-Git-Tag: REL_15_BETA1~1500
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=bdeb2c4ec2700bfa561061ccd19181326ee01c3f;p=thirdparty%2Fpostgresql.git

Add WRITE_INDEX_ARRAY

We have a few WRITE_{name of type}_ARRAY macros, but the one case
using the Index type was hand-coded.  Wrap it into a macro as well.

This also changes the behavior slightly: Before, the field name was
skipped if the length was zero.  Now it prints the field name even in
that case.  This is more consistent with how other array fields are
handled.

Reviewed-by: Jacob Champion <pchampion@vmware.com>
Discussion: https://www.postgresql.org/message-id/c091e5cd-45f8-69ee-6a9b-de86912cc7e7@enterprisedb.com
---

diff --git a/src/backend/nodes/outfuncs.c b/src/backend/nodes/outfuncs.c
index 36e618611fd..f2a6a6e7a08 100644
--- a/src/backend/nodes/outfuncs.c
+++ b/src/backend/nodes/outfuncs.c
@@ -124,6 +124,13 @@ static void outChar(StringInfo str, char c);
 			appendStringInfo(str, " %u", node->fldname[i]); \
 	} while(0)
 
+#define WRITE_INDEX_ARRAY(fldname, len) \
+	do { \
+		appendStringInfoString(str, " :" CppAsString(fldname) " "); \
+		for (int i = 0; i < len; i++) \
+			appendStringInfo(str, " %u", node->fldname[i]); \
+	} while(0)
+
 #define WRITE_INT_ARRAY(fldname, len) \
 	do { \
 		appendStringInfoString(str, " :" CppAsString(fldname) " "); \
@@ -2510,14 +2517,7 @@ _outPathTarget(StringInfo str, const PathTarget *node)
 	WRITE_NODE_TYPE("PATHTARGET");
 
 	WRITE_NODE_FIELD(exprs);
-	if (node->sortgrouprefs)
-	{
-		int			i;
-
-		appendStringInfoString(str, " :sortgrouprefs");
-		for (i = 0; i < list_length(node->exprs); i++)
-			appendStringInfo(str, " %u", node->sortgrouprefs[i]);
-	}
+	WRITE_INDEX_ARRAY(sortgrouprefs, list_length(node->exprs));
 	WRITE_FLOAT_FIELD(cost.startup, "%.2f");
 	WRITE_FLOAT_FIELD(cost.per_tuple, "%.2f");
 	WRITE_INT_FIELD(width);