Date: Tue, 25 Sep 2018 18:42:51 -0700
From: Andres Freund <andres@anarazel.de>
To: Michael Paquier <michael@paquier.xyz>
Cc: Postgres hackers <pgsql-hackers@postgresql.org>
Subject: Re: pgsql: Remove absolete function TupleDescGetSlot().
Message-ID: <20180926014251.zrn27ljtqlqjzswx@alap3.anarazel.de>
References: <E1g4wv3-0001Q0-8n@gemulon.postgresql.org>
 <20180926000413.GC1659@paquier.xyz>
 <20180926001038.fh7rrhrqrwvpu43t@alap3.anarazel.de>
 <20180926013851.GG1659@paquier.xyz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20180926013851.GG1659@paquier.xyz>
Precedence: bulk

On 2018-09-26 10:38:51 +0900, Michael Paquier wrote:
> On Tue, Sep 25, 2018 at 05:10:39PM -0700, Andres Freund wrote:
> >> git grep TupleDescGetSlot
> >> doc/src/sgml/xfunc.sgml:     * user-defined SRFs that use the deprecated TupleDescGetSlot().
> >> src/include/funcapi.h:   * user-defined SRFs that use the deprecated TupleDescGetSlot().
> > 
> > But here I'm less convinced. It's not entirely clear to me that the only
> > real reason for this to exists actually was TupleDescGetSlot(). OTOH, I
> > can't really see any proper reason to have it either.
> 
> I have not been following the recent thread about the refactoring of
> TupleSlot and such very closely, but if you don't plan to use
> TupleTableSlot in FuncCallContext in the future, cannot this just go
> away?  The function is not here anymore, so my take would be to get rid
> of all things which relied on its presence.

My point is that FuncCallContext->slot isn't actually all that related
to TupleDescGetSlot() and could be used entirely independently.

Greetings,

Andres Freund