Re: pageinspect: Hash index support

On Tue, Sep 20, 2016 at 11:14 PM, Michael Paquier <michael.paquier@gmail.com
> wrote:

>
> +     <para>
> +      The type information will be '<literal>m</literal>' for a metadata
> page,
> +      '<literal>v</literal>' for an overflow page,
> '<literal>b</literal>' for a bucket page,
> +      '<literal>i</literal>' for a bitmap page, and
> '<literal>u</literal>' for an unused page.
> +     </para>
>


> Other functions don't go into this level of details, so I would just
> rip out this paragraph.
>

I'd argue that the other functions should go into that level detail in some
places. Pageinspect is needlessly hard to use; not all precedent is good
precedent.  Some of them do refer you to header files or README files,
which can be useful. But the abbreviations used here are not explained in
any header file or README file, so I think the right place to explain them
is the documentation in that case. Or change from the single-letter strings
to full name strings so they are self-documenting.

Cheers,

Jeff