Re: Using defines for protocol characters

On Mon, 7 Aug 2023 at 16:50, Tatsuo Ishii <ishii(at)sraoss(dot)co(dot)jp> wrote:

> > On Mon, Aug 07, 2023 at 04:02:08PM -0400, Tom Lane wrote:
> >> Dave Cramer <davecramer(at)gmail(dot)com> writes:
> >>> On Mon, 7 Aug 2023 at 12:59, Robert Haas <robertmhaas(at)gmail(dot)com>
> wrote:
> >>>> PqMsgEmptyQueryResponse or something like that seems better, if we
> >>>> want to keep the current capitalization. I'm not a huge fan of the way
> >>>> we vary our capitalization conventions so much all over the code base,
> >>>> but I think we would at least do well to keep it consistent from one
> >>>> end of a certain identifier to the other.
> >>
> >>> I don't have a strong preference, but before I make the changes I'd
> like to
> >>> get consensus.
> >>> Can we vote or whatever it takes to decide on a naming pattern that is
> >>> acceptable ?
> >>
> >> I'm good with Robert's proposal above.
> >
> > +1
>
> +1.
>
> Also we need to decide what to do with them:
>
> > #define PQMSG_REQ_PREPARED 'S'
> > #define PQMSG_REQ_PORTAL 'P'
>
> If we go "PqMsgEmptyQueryResponse", probably we should go something
> like for these?
>
> #define PqMsgReqPrepared 'S'
> #define PqMsgReqPortal 'P'
>

I went with PqMsgPortalSubCommand and PqMsgPreparedSubCommand

See attached patch

Dave