Re: [HACKERS] Patch to implement pg_current_logfile() function

On Thu, Jan 19, 2017 at 6:08 AM, Karl O. Pinc <kop@meme.com> wrote:
> On Wed, 18 Jan 2017 15:52:36 -0500
> Robert Haas <robertmhaas@gmail.com> wrote:
>
>> On Wed, Jan 18, 2017 at 12:08 PM, Karl O. Pinc <kop@meme.com> wrote:
>> > Seems to me that the file format should
>> > be documented if there's any intention that the end user
>> > look at or otherwise use the file's content.
>> >
>> > It's fine with me if the content of current_logfiles
>> > is supposed to be internal to PG and not exposed
>> > to the end user.  I'm writing to make sure that
>> > this is a considered decision.
>>
>> On the whole, documenting it seems better than documenting it,
>> provided there's a logical place to include it in the existing
>> documentation.
>>
>> But, anyway, Michael shouldn't remove it without some explanation or
>> discussion.
>
> He explained that where it was looked clunky, it being
> inside a table that otherwise has rows that are not tall.
>
> And, it looks like I'm wrong.  The format is documented
> by way of an example in section 19.8.1. Where To Log
> under log_destination (string).
>
> Sorry for the bother.

Er, well. I kept the same detail verbosity in the docs...

> I would like to see index entries for "current_logfiles"
> so this stuff is findable.

Why not.
-- 
Michael