Tidy recent code bloat in pg_creatersubscriber::cleanup_objects_atexit

Started by Peter Smith10 months ago4 messages
#1Peter Smith
smithpb2250@gmail.com
1 attachment(s)

During a recent review of pg_creatersubscriber I saw that commit
e117cfb introduced a common 'dbinfos' struct to contain the array of
individual 'dbinfo[i]' infos. But, this now means that getting to each
dbinfo[i] requires another level of referencing.

In some places, e.g. function cleanup_objects_atexit, this caused
unnecessary code bloat. IMO this function is crying out for a local
variable to simplify the code again.

Please see the attached patch that implements this suggestion.

======
Kind Regards,
Peter Smith
Fujitsu Australia

Attachments:

v1-0001-Add-cleanup_objects_atexit-local-var.patchapplication/octet-stream; name=v1-0001-Add-cleanup_objects_atexit-local-var.patchDownload
From 92dbd7503211e604b72a438b186dce4e09c69cb8 Mon Sep 17 00:00:00 2001
From: Peter Smith <peter.b.smith@fujitsu.com>
Date: Tue, 11 Mar 2025 12:15:57 +1100
Subject: [PATCH v1] Add cleanup_objects_atexit local var

---
 src/bin/pg_basebackup/pg_createsubscriber.c | 26 ++++++++++++++------------
 1 file changed, 14 insertions(+), 12 deletions(-)

diff --git a/src/bin/pg_basebackup/pg_createsubscriber.c b/src/bin/pg_basebackup/pg_createsubscriber.c
index a5a2d61..6baf92e 100644
--- a/src/bin/pg_basebackup/pg_createsubscriber.c
+++ b/src/bin/pg_basebackup/pg_createsubscriber.c
@@ -184,17 +184,19 @@ cleanup_objects_atexit(void)
 
 	for (int i = 0; i < num_dbs; i++)
 	{
-		if (dbinfos.dbinfo[i].made_publication || dbinfos.dbinfo[i].made_replslot)
+		struct LogicalRepInfo *dbinfo = &dbinfos.dbinfo[i];
+
+		if (dbinfo->made_publication || dbinfo->made_replslot)
 		{
 			PGconn	   *conn;
 
-			conn = connect_database(dbinfos.dbinfo[i].pubconninfo, false);
+			conn = connect_database(dbinfo->pubconninfo, false);
 			if (conn != NULL)
 			{
-				if (dbinfos.dbinfo[i].made_publication)
-					drop_publication(conn, &dbinfos.dbinfo[i]);
-				if (dbinfos.dbinfo[i].made_replslot)
-					drop_replication_slot(conn, &dbinfos.dbinfo[i], dbinfos.dbinfo[i].replslotname);
+				if (dbinfo->made_publication)
+					drop_publication(conn, dbinfo);
+				if (dbinfo->made_replslot)
+					drop_replication_slot(conn, dbinfo, dbinfo->replslotname);
 				disconnect_database(conn, false);
 			}
 			else
@@ -204,18 +206,18 @@ cleanup_objects_atexit(void)
 				 * that some objects were left on primary and should be
 				 * removed before trying again.
 				 */
-				if (dbinfos.dbinfo[i].made_publication)
+				if (dbinfo->made_publication)
 				{
 					pg_log_warning("publication \"%s\" created in database \"%s\" on primary was left behind",
-								   dbinfos.dbinfo[i].pubname,
-								   dbinfos.dbinfo[i].dbname);
+								   dbinfo->pubname,
+								   dbinfo->dbname);
 					pg_log_warning_hint("Drop this publication before trying again.");
 				}
-				if (dbinfos.dbinfo[i].made_replslot)
+				if (dbinfo->made_replslot)
 				{
 					pg_log_warning("replication slot \"%s\" created in database \"%s\" on primary was left behind",
-								   dbinfos.dbinfo[i].replslotname,
-								   dbinfos.dbinfo[i].dbname);
+								   dbinfo->replslotname,
+								   dbinfo->dbname);
 					pg_log_warning_hint("Drop this replication slot soon to avoid retention of WAL files.");
 				}
 			}
-- 
1.8.3.1

#2Michael Paquier
michael@paquier.xyz
In reply to: Peter Smith (#1)
Re: Tidy recent code bloat in pg_creatersubscriber::cleanup_objects_atexit

On Tue, Mar 11, 2025 at 12:29:42PM +1100, Peter Smith wrote:

During a recent review of pg_creatersubscriber I saw that commit
e117cfb introduced a common 'dbinfos' struct to contain the array of
individual 'dbinfo[i]' infos. But, this now means that getting to each
dbinfo[i] requires another level of referencing.

In some places, e.g. function cleanup_objects_atexit, this caused
unnecessary code bloat. IMO this function is crying out for a local
variable to simplify the code again.

Right. This improves the clarity of the code, so agreed about the use
of a local variable here.
--
Michael

#3Michael Paquier
michael@paquier.xyz
In reply to: Michael Paquier (#2)
Re: Tidy recent code bloat in pg_creatersubscriber::cleanup_objects_atexit

On Tue, Mar 11, 2025 at 04:06:14PM +0900, Michael Paquier wrote:

Right. This improves the clarity of the code, so agreed about the use
of a local variable here.

More code paths of pg_createsubscriber.c have similar loops, but this
is the only one where LogicalRepInfo can be used. So, applied as you
have suggested, without touching the other loops.
--
Michael

#4Peter Smith
smithpb2250@gmail.com
In reply to: Michael Paquier (#3)
Re: Tidy recent code bloat in pg_creatersubscriber::cleanup_objects_atexit

Thanks for pushing it.

======
Kind Regards,
Peter Smith.
Fujitsu Australia