Thread: log rotatoin doc updates

This patch updates log rotation documentation:
    + Removed false statement that log_filename can only be changed on restart
(it is reloadable via sighup);
    + Added a couple of examples;
    + Cleaned up a few smgl tags;

"Ed L." <pgsql@bluepolka.net> writes:
> This patch updates log rotation documentation:

Applied, except for

>     + Removed false statement that log_filename can only be changed on restart
> (it is reloadable via sighup);

The statement was correct as given so I didn't remove it.  (If you think
it should mention that SIGHUP works, there are several dozen other uses
of the identical wording to describe SIGHUP variables ...)

>     + Added a couple of examples;

I moved these down a little, since it didn't seem to make sense to
provide an example before all the relevant variables had been discussed.

I'm not sure about the markup here; it might be better to break out of
the <variablelist> and format the examples as independent paragraphs.
Comments anyone?

            regards, tom lane

On Friday September 17 2004 4:44, Tom Lane wrote:
> "Ed L." <pgsql@bluepolka.net> writes:
> > This patch updates log rotation documentation:
>
> Applied, except for
>
> >     + Removed false statement that log_filename can only be changed on
> > restart (it is reloadable via sighup);
>
> The statement was correct as given so I didn't remove it.  (If you think
> it should mention that SIGHUP works, there are several dozen other uses
> of the identical wording to describe SIGHUP variables ...)

"This option can only be set at server start or in the
<filename>postgresql.conf</filename> configuration file."

My misunderstanding.  I misread the above to mean SIGHUP did not work, as
opposed to postgresql.conf variables for which SIGHUP *will* reload them.
I guess the word "only" refers to not being able to use SET?

Ed

"Ed L." <pgsql@bluepolka.net> writes:
> My misunderstanding.  I misread the above to mean SIGHUP did not work, as
> opposed to postgresql.conf variables for which SIGHUP *will* reload them.
> I guess the word "only" refers to not being able to use SET?

Or ALTER DATABASE, ALTER USER, etc.

I guess the wording assumes that you know that altering postgresql.conf
requires a SIGHUP to make anything happen.  I wouldn't object to
inserting some explicit mention of SIGHUP into these canned phrases,
if it can be done without making it much longer or more obscure.
Any thoughts?

            regards, tom lane

> !         <para>
> !           Example:  To keep

Why aren't 'example' tags used here? Seems it would be more appropriate
then "<para>Example: ...</para>"..

Example tag: (http://www.docbook.org/tdg/en/html/example.html)
(http://developer.postgresql.org/docs/postgres/runtime-config.html)


Tom:
Also, what do you mean by 'break out of the <variablelist>'?

Have the examples follow the <variablelist> or have the examples set in
such a way so as the <listitem>'s indentation doesn't affect the
example's text?

--
Regards,
        James William Pye

James William Pye <flaw@rhid.com> writes:
> Also, what do you mean by 'break out of the <variablelist>'?

I meant end the variablelist, format the examples, start a new
variablelist for the following variables.

Or we could put the examples at the end of the subsection, but that
seems awkwardly far away.

            regards, tom lane