On 27/10/2025 04:32, Fujii Masao wrote:
> I think the "hs &&" and "ro &&" checks are no longer needed,
> since we've already confirmed they're not NULL at that point.

Right. Checks removed.

> Also, should "unknown" be marked for translation, as in the \conninfo code?
> I'm not sure whether showing a translated string in the psql prompt is
> desirable, though.

Here I am not sure if it's applicable. In this file there are other
strings that are not marked for translation, "abort", "on", "off". I
changed the string to _("unknown") for now, but we can remove it in
further iterations if we agree it's not desirable :)

v5 attached.

Thanks for the review.

Best, Jim

From 1323e4b555fbd63796a759d0b6741b8258ad88eb Mon Sep 17 00:00:00 2001
From: Jim Jones <[email protected]>
Date: Mon, 27 Oct 2025 08:36:48 +0100
Subject: [PATCH v5] Add %i prompt escape to indicate server read-only status

This patch introduces a new prompt escape `%i` for psql, which shows
whether the connected server is operating in read-only or read/write
mode. It expands to `read-only` if either the server is in hot standby
mode (`in_hot_standby = on`) or the session's default transaction
mode is read-only (`default_transaction_read_only = on`). Otherwise,
it displays `read/write`.

This is useful for distinguishing read-only sessions (e.g. connected
to a standby, or using a default read-only transaction mode) from
read/write ones at a glance, especially when working with multiple
connections in replicated or restricted environments.
---
 doc/src/sgml/ref/psql-ref.sgml | 14 ++++++++++++++
 src/bin/psql/prompt.c          | 17 +++++++++++++++++
 2 files changed, 31 insertions(+)

diff --git a/doc/src/sgml/ref/psql-ref.sgml b/doc/src/sgml/ref/psql-ref.sgml
index 1a339600bc..662dfb2ed5 100644
--- a/doc/src/sgml/ref/psql-ref.sgml
+++ b/doc/src/sgml/ref/psql-ref.sgml
@@ -5044,6 +5044,20 @@ testdb=&gt; <userinput>INSERT INTO my_table VALUES (:'content');</userinput>
         </listitem>
       </varlistentry>
 
+      <varlistentry id="app-psql-prompting-i">
+        <term><literal>%i</literal></term>
+        <listitem>
+          <para>
+            Displays the session's read-only status as <literal>read-only</literal>
+            if the server is in hot standby (<literal>in_hot_standby</literal> is
+            <literal>on</literal>) or the default transaction mode is read-only
+            (<literal>default_transaction_read_only</literal> is <literal>on</literal>),
+            or <literal>read-write</literal> otherwise. Useful for identifying
+            sessions that cannot perform writes, such as in replication setups.
+          </para>
+        </listitem>
+      </varlistentry>
+
       <varlistentry id="app-psql-prompting-x">
         <term><literal>%x</literal></term>
         <listitem>
diff --git a/src/bin/psql/prompt.c b/src/bin/psql/prompt.c
index b08d7328fb..fdcb779d43 100644
--- a/src/bin/psql/prompt.c
+++ b/src/bin/psql/prompt.c
@@ -43,6 +43,8 @@
  *			or a ! if session is not connected to a database;
  *		in prompt2 -, *, ', or ";
  *		in prompt3 nothing
+ * %i - displays "read-only" if in hot standby or default_transaction_read_only
+ *		is on, "read/write" otherwise.
  * %x - transaction status: empty, *, !, ? (unknown or no connection)
  * %l - The line number inside the current statement, starting from 1.
  * %? - the error code of the last query (not yet implemented)
@@ -247,7 +249,22 @@ get_prompt(promptStatus_t status, ConditionalStack cstack)
 							break;
 					}
 					break;
+				case 'i':
+					if (pset.db)
+					{
+						const char *hs = PQparameterStatus(pset.db, "in_hot_standby");
+						const char *ro = PQparameterStatus(pset.db, "default_transaction_read_only");
 
+						if (!hs || !ro)
+							strlcpy(buf, _("unknown"), sizeof(buf));
+						else if (strcmp(hs, "on") == 0 || strcmp(ro, "on") == 0)
+							strlcpy(buf, "read-only", sizeof(buf));
+						else
+							strlcpy(buf, "read/write", sizeof(buf));
+					}
+					else
+						buf[0] = '\0';
+					break;
 				case 'x':
 					if (!pset.db)
 						buf[0] = '?';
-- 
2.43.0

Reply via email to