Re: Refactoring: join MakeSingleTupleTableSlot() and MakeTupleTableSlot() - Mailing list pgsql-hackers

From Alvaro Herrera
Subject Re: Refactoring: join MakeSingleTupleTableSlot() and MakeTupleTableSlot()
Date
Msg-id 202110261636.3wowatvoc4ub@alvherre.pgsql
Whole thread Raw
In response to Re: Refactoring: join MakeSingleTupleTableSlot() and MakeTupleTableSlot()  (Robert Haas <robertmhaas@gmail.com>)
List pgsql-hackers
On 2021-Oct-26, Robert Haas wrote:

> On Tue, Oct 26, 2021 at 7:54 AM Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
> > Another point that could be made is that perhaps
> > MakeSingleTupleTableSlot should always construct a slot using virtual
> > tuples rather than passing TTSOps as a parameter?
> 
> I haven't really looked at this issue deeply but that seems like it
> might be a bit confusing. Then "single" would end up being an alias
> for "virtual" which I don't suppose is what anyone is expecting.

Yeah -- another point against that idea is that most of the callers are
indeed not using virtual tuples, so it doesn't really work.  I was just
thinking that if something wants to process transient tuples they may
just be virtual and not be forced to make them heap tuples, but on
looking again, that's not how the abstraction works.

-- 
Álvaro Herrera              Valdivia, Chile  —  https://www.EnterpriseDB.com/
"Siempre hay que alimentar a los dioses, aunque la tierra esté seca" (Orual)



pgsql-hackers by date:

Previous
From: "Bossart, Nathan"
Date:
Subject: Re: Predefined role pg_maintenance for VACUUM, ANALYZE, CHECKPOINT.
Next
From: Mark Dilger
Date:
Subject: Re: src/port/snprintf.c: Optimize the common base=10 case in fmtint