Re: Backport of pg_statistics typos fix

On Wed, 8 Feb 2017 14:54:17 -0500
Robert Haas <robertmhaas(at)gmail(dot)com> wrote:

> On Tue, Feb 7, 2017 at 4:22 AM, Yugo Nagata <nagata(at)sraoss(dot)co(dot)jp> wrote:
> > I found typos "pg_statistics" in REL9_6_STABLE, but that has been
> > fixed in the master branch.
> >
> > Fix typo: pg_statistics -> pg_statistic
> > https://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=5a366b4ff4ceceb9793fcc13c3f097ee0d32c56d;hp=f7c62462402972b13d10e43f104ca0c0fecb6d08
> >
> > I think it would be better to backport this to other branches.
>
> We usually leave such decisions to the discretion of the committer,
> because back-porting such changes takes time and sometimes it just
> isn't that important. Nobody's likely to be confused by a few
> instances of writing pg_statistics rather than pg_statistic.
> Personally, I favor not back-porting such things in most cases,
> because I think patches that get back-ported should be strictly
> limited to bug fixes, and typos in code comments aren't bug fixes.
> But not everyone has the same opinion on this. What's your reason for
> wanting it back-ported?

I agree typos in code comments aren't bug fixes and not need
to get back-ported. However, there are typos also in the document.

The <function>scalarltsel</function> function retrieves the histogram for
<structfield>unique1</structfield> from
- <structname>pg_statistics</structname>. For manual queries it is more
+ <structname>pg_statistic</structname>. For manual queries it is more
convenient to look in the simpler <structname>pg_stats</structname>
view:

I think this might be a document bug, but if nobody cares of it,
I also don't mind.

Thanks,

>
> BTW, looking at that commit, this change looks to have adjusted this
> from being wrong to still being wrong:
>
> -<listitem><para>Allow pg_statistics to be reset by calling
> pg_stat_reset() (Christopher)</para></listitem>
> +<listitem><para>Allow pg_statistic to be reset by calling
> pg_stat_reset() (Christopher)</para></listitem>
>
> It's true that pg_stat_reset() doesn't reset the nonexistent
> pg_statistics table. But it doesn't reset pg_statistic either. IIUC,
> it resets the data gathered by the statistics collector, which is
> something else altogether.
>
> --
> Robert Haas
> EnterpriseDB: http://www.enterprisedb.com
> The Enterprise PostgreSQL Company

--
Yugo Nagata <nagata(at)sraoss(dot)co(dot)jp>