Can avoid list_copy in recomputeNamespacePath() conditionally?

From: amul sul <sulamul(at)gmail(dot)com>
To: PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Can avoid list_copy in recomputeNamespacePath() conditionally?
Date: 2019-10-31 08:41:03
Message-ID: CAAJ_b94dGU4BYFtVCKkj58hNuPL9hwP4p1a9Eqq+e+v8GP8KoQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi all,

I wondered can we have a shortcut somewhat similar to following POC
in recomputeNamespacePath () when the recomputed path is the same as the
previous baseSearchPath/activeSearchPath :

== POC patch ==
diff --git a/src/backend/catalog/namespace.c
b/src/backend/catalog/namespace.c
index e251f5a9fdc..b25ef489e47 100644
--- a/src/backend/catalog/namespace.c
+++ b/src/backend/catalog/namespace.c
@@ -3813,6 +3813,9 @@ recomputeNamespacePath(void)
!list_member_oid(oidlist, myTempNamespace))
oidlist = lcons_oid(myTempNamespace, oidlist);

+ /* TODO: POC */
+ if (equal(oidlist, baseSearchPath))
+ return;
/*
* Now that we've successfully built the new list of namespace OIDs,
save
* it in permanent storage.
== POC patch end ==

It can have two advantages as:

1. Avoid unnecessary list_copy() in TopMemoryContext context &
2. Global pointers like activeSearchPath/baseSearchPath will not change if
some
implementation end up with cascaded call to recomputeNamespacePath().

Thoughts/Comments?

Regards,
Amul

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Antonin Houska 2019-10-31 08:43:47 Re: MarkBufferDirtyHint() and LSN update
Previous Message Fujii Masao 2019-10-31 08:38:32 Re: Problem with synchronous replication