Re: TupleTableSlot abstraction

On Wed, 26 Sep 2018 at 05:15, Andres Freund <andres(at)anarazel(dot)de> wrote:
>
> Hi,
>
> On 2018-09-04 18:35:34 +0530, Amit Khandekar wrote:
> > Subject: [PATCH 05/14] Use tts_flags instead of previous bool members
> >
> > Pack the boolean members in TupleTableSlot into a 16 bit tts_flags.
> > This reduces the size of TupleTableSlot since each bool member takes
> > at least a byte on the platforms where bool is defined as a char.
> >
> > Ashutosh Bapat and Andres Freund
>
> > +
> > +/* true = slot is empty */
> > +#define TTS_ISEMPTY (1 << 1)
> > +#define IS_TTS_EMPTY(slot) ((slot)->tts_flags & TTS_ISEMPTY)
> > +#define SET_TTS_EMPTY(slot) ((slot)->tts_flags |= TTS_ISEMPTY)
> > +#define RESET_TTS_EMPTY(slot) ((slot)->tts_flags &= ~TTS_ISEMPTY)
>
> The flag stuff is the right way, but I'm a bit dubious about the
> accessor functions. I can see open-coding the accesses, using the
> macros, or potentially going towards using bitfields.
>
> Any comments?

I like this abstraction using macros, since this will allow us to
conveniently change the way this information is stored inside the
slot. I think for this same reason we have defined macros
(HeapTupleIsHotUpdated, etc) for each bit of heaptuple infomask.

--
Thanks,
-Amit Khandekar
EnterpriseDB Corporation
The Postgres Database Company