Re: COMMENT/SEC LABEL tab complete support suggests wrong SQL. - Mailing list pgsql-hackers

From Chao Li
Subject Re: COMMENT/SEC LABEL tab complete support suggests wrong SQL.
Date
Msg-id 8AF39AC6-0FFF-4365-A390-CDA993E84797@gmail.com
Whole thread
In response to COMMENT/SEC LABEL tab complete support suggests wrong SQL.  (Kirill Reshke <reshkekirill@gmail.com>)
List pgsql-hackers

> On Sep 18, 2026, at 16:01, Kirill Reshke <reshkekirill@gmail.com> wrote:
>
> HI!
>
> I noticed an issue on tab completion with LARGE OBJECT support. I am
> aware that tab complete support need not to be always-correct, but
> this cases are simple, so decided to fix.
>
> The psql tab completion for COMMENT ON and SECURITY LABEL ON suggests
> "IS" immediately after "LARGE OBJECT", this produces invalid SQL:
>
>  =# COMMENT ON LARGE OBJECT <TAB>
>  -- now COMMENT ON LARGE OBJECT IS
>
> The same bug exists for SECURITY LABEL. SEC LABEL also missed EVENT
> TRIGGER/  FOREIGN TABLE/MATERIALIZED VIEW cases.
>
> Same bug was fixed for GRANT/REVOKE in ff0bcb248e6.
>
>
>
> Quick sum up:
>  Before:
>    COMMENT ON LARGE OBJECT <TAB> → IS (invalid SQL)
>    SECURITY LABEL ON LARGE OBJECT <TAB> → IS (invalid SQL)
>    SECURITY LABEL ON EVENT TRIGGER <TAB> → IS (invalid SQL)
>    SECURITY LABEL ON FOREIGN TABLE <TAB> → IS (invalid SQL)
>    SECURITY LABEL ON MATERIALIZED VIEW <TAB> → IS (invalid SQL)
>
>  After:
>    COMMENT ON LARGE OBJECT <TAB> → OID list (correct)
>    SECURITY LABEL ON LARGE OBJECT <TAB> → OID list (correct)
>    SECURITY LABEL ON EVENT TRIGGER <TAB> → trigger name list (correct)
>    SECURITY LABEL ON FOREIGN TABLE <TAB> → foreign table list (correct)
>    SECURITY LABEL ON MATERIALIZED VIEW <TAB> → matview list (correct)
>
>
>
> --
> Best regards,
> Kirill Reshke
> <v1-0001-psql-Fix-tab-completion-for-COMMENT-SECURITY-LABE.patch>

Do we also need to support SECURITY LABEL FOR provider ON LARGE OBJECT …?

Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/







pgsql-hackers by date:

Previous
From: Nisha Moond
Date:
Subject: Re: Proposal: Conflict log history table for Logical Replication
Next
From: kedar anavardekar
Date:
Subject: Re: [PATCH] Release replication slot on error in SQL-callable slot functions