Re: Document NULL

From: Junwang Zhao <zhjwpku(at)gmail(dot)com>
To: "David G(dot) Johnston" <david(dot)g(dot)johnston(at)gmail(dot)com>
Cc: Chao Li <li(dot)evan(dot)chao(at)gmail(dot)com>, Marcos Pegoraro <marcos(at)f10(dot)com(dot)br>, Álvaro Herrera <alvherre(at)kurilemu(dot)de>, jian he <jian(dot)universality(at)gmail(dot)com>, Jeff Davis <pgsql(at)j-davis(dot)com>, Yugo NAGATA <nagata(at)sraoss(dot)co(dot)jp>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Peter Eisentraut <peter(at)eisentraut(dot)org>, David Rowley <dgrowleyml(at)gmail(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: Document NULL
Date: 2026-02-01 15:11:51
Message-ID: CAEG8a3+wqhs3jUVQw3ETOkuwOaePp-ZU1Ru7PaWfvzuCE1Q7GQ@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Sat, Jan 31, 2026 at 9:23 AM David G. Johnston
<david(dot)g(dot)johnston(at)gmail(dot)com> wrote:
>
> On Thu, Jan 29, 2026 at 10:31 PM Chao Li <li(dot)evan(dot)chao(at)gmail(dot)com> wrote:
>>
>>
>> My only comment on v12 is:
>> ```
>> - Boolean (true/false) results.
>> + <link linkend="nullvalues">three-valued</link> boolean typed
>> ```
>>
>> I think it’s better to use capital “B” for “boolean”. See [1].
>>
>
> Thank you.
>
> Squashed v13 with that change (also removed the word 'typed') attached.
>
> David J.
>

+<!-- Agent Instructions for Interactive Example Review
+ - There are around 36 screen elements in this files.
+ - Show me each one along with its corresponding programlisting.
+ - Keep track of progress to avoid repeating or missing.
+ - When given new directives, update this block to remember them.
+ - Before presenting each example consult this block for how to behave.
+ - Also, using the psql program, run the programlisting
+ (you will need to incorporate the contents of the first program listing -
+ which has no screen output associated with it) to get them to
execute correctly.
+ - Show also the result of running psql.
+ - Point out if the script run and the screen contents do not match.
+ - Wait for me to say 'next' before proceeding to the next pair.
+ - Do not point out trailing whitespace differences; psql produces
such output.
+ - Do not tell me you are 'ignoring trailing whitespace as
instructruced' or similar.
+ - IMPORTANT: Point out when screen elements have extra blank lines
within table output that don't match psql output. These should be
removed to match psql exactly.
+ - IMPORTANT: BEFORE running each example, check the programlisting
for lowercase 'null' used as SQL NULL literals - these must be
uppercase NULL (e.g., 'NULL::text' not 'null::text', 'array[1, NULL]'
not 'array[1, null]'). Point out and fix these IMMEDIATELY.
+ - programlisting usage of literal NULL is uppercase.
+ - IMPORTANT: In programlisting elements, null literals should be
written as uppercase NULL (e.g., 'NULL', 'null::text' should be
'NULL::text', 'array[1, NULL]' not 'array[1, null]'). Point out when
programlistings use lowercase 'null' - this is incorrect.
+ - IMPORTANT: Column headers in screen elements should use titlecase
(e.g., 'Is Null' not 'IS NULL', 'Count Value' not 'Count VALUE').
Point out when headers are not titlecase.
+ - IMPORTANT: Exception to titlecase rule - when the header refers
to the SQL operator/predicate used in that column's expression, use
the exact operator syntax (e.g., 'IS NULL', 'IS DISTINCT FROM', 'IS
!=').
+ - IMPORTANT: SQL keywords AND/OR/IN should be uppercase in
programlisting expressions but titlecase in column headers (e.g.,
'True And Null' not 'True AND Null', 'In Present' not 'IN Present').
+ - IMPORTANT: Exception to titlecase rule - when referring to JSON
null values use lowercase 'null' (e.g., 'Cast jsonb null to SQL NULL'
not 'Cast jsonb NULL to SQL NULL'), and when referring to SQL NULL
values use uppercase 'NULL'.
+ - Point out if headers exceed 80 characters.
+ - NEVER manually replace parts of screen elements, always copy in
psql output.
+ - After making a program change re-run the entire example flow again.
+ - If there is a mis-match in the screen/psql output accept either
'fix' or 'skip' as commands, 'next' is invalid and ask whether to fix
or skip.
+Example shell:
+cd psql << 'EOF'
+\pset null Null
+\pset display_true true
+\pset display_false false
+\pset footer off
+
+SELECT
+ NULL OR NOT NULL AS "N OR !N";
+EOF
+-->

It seems to me that these are prompts for an LLM, should they be
committed in the tree?

--
Regards
Junwang Zhao

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message David E. Wheeler 2026-02-01 15:23:58 Re: PATCH: jsonpath string methods: lower, upper, initcap, l/r/btrim, replace, split_part
Previous Message Junwang Zhao 2026-02-01 14:55:53 Re: [PATCH] Fix error message in RemoveWalSummaryIfOlderThan to indicate file removal failure