Обсуждение: pg_input_error_info doc 2 exampled crammed together

hi.

select * from pg_input_error_info('42000000000', 'integer')
select message, detail from pg_input_error_info('1234.567', 'numeric(7,4)')
I found above two examples at [0] crammed together.


       <para>
        <literal>select * from pg_input_error_info('42000000000',
'integer')</literal>
        <returnvalue></returnvalue>
<programlisting>
                       message                        | detail | hint
| sql_error_code
------------------------------------------------------+--------+------+----------------
 value "42000000000" is out of range for type integer |        |      | 22003
</programlisting>
       </para>
       <para>
        <literal>select message, detail from
pg_input_error_info('1234.567', 'numeric(7,4)')</literal>
        <returnvalue></returnvalue>
<programlisting>
        message         |                                      detail
------------------------+----------------------------------&zwsp;-------------------------------------------------
 numeric field overflow | A field with precision 7, scale 4 must round
to an absolute value less than 10^3.
</programlisting>


after checking the definition of <programlisting>[1], <screen>[2],
maybe here we should use <screen> and also add `(1 row)` information.

or we can simply add a empty new line between
` value "42000000000" is out of range for type integer |        |      | 22003`
and
`</programlisting>`


[0] https://www.postgresql.org/docs/current/functions-info.html#FUNCTIONS-INFO-VALIDITY
[1] https://tdg.docbook.org/tdg/4.5/programlisting
[2] https://tdg.docbook.org/tdg/4.5/screen



-- 
 I recommend David Deutsch's <<The Beginning of Infinity>>

  Jian

after checking the definition of <programlisting>[1], <screen>[2],
maybe here we should use <screen>

and also add `(1 row)` information.

or we can simply add a empty new line between
` value "42000000000" is out of range for type integer |        |      | 22003`
and
`</programlisting>`

On Sun, Apr 28, 2024 at 06:45:30PM -0700, David G. Johnston wrote:
> My preference would be to limit this section to a single example.  The
> numeric one, as it provides values for more output columns.  I would change
> the output format to expanded from default, in order to show all columns
> and not overrun the length of a single line.

Agreed that having two examples does not bring much, so this could be
brought to a single one.  The first one is enough to show the point of
the function, IMO.  It is shorter in width and it shows all the output
columns.
--
Michael

On Sun, Apr 28, 2024 at 06:45:30PM -0700, David G. Johnston wrote:
> My preference would be to limit this section to a single example.  The
> numeric one, as it provides values for more output columns.  I would change
> the output format to expanded from default, in order to show all columns
> and not overrun the length of a single line.

Agreed that having two examples does not bring much, so this could be
brought to a single one.  The first one is enough to show the point of
the function, IMO.  It is shorter in width and it shows all the output
columns.

On Sun, Apr 28, 2024 at 10:07:49PM -0700, David G. Johnston wrote:
> Agreed.  The column names are self-explanatory if you’ve seen errors
> before.  The values are immaterial.  Plus we don’t generally use
> psql-specific features in our examples.

Okay, I've just cleaned up that a bit with f6ab942f5de0.
--
Michael