Re: Fixing MSVC's inability to detect elog(ERROR) does not return

David Rowley <dgrowleyml(at)gmail(dot)com> writes:
> On Wed, 17 Sept 2025 at 16:03, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> It absolutely stands for "predicate". That's an ancient Lisp-ism.

> Thanks for the confirmation. I'm happy enough to leave the _p in
> there, but at the same time, I don't see the particular reason to
> follow some ancient Lisp rule.

I'm just saying that that's surely where the gcc crew got it from.
I do agree with Peter that we should follow that convention within
the narrow realm of compiler intrinsics; but I'm not arguing for
running around and renaming random PG functions to something_p.

As long as we're delving into weeds: we have another project
convention for "_P", which is terminal symbols in our grammar such
as NULL_P. Clearly, that "P" is not for "predicate"; I suppose it
should be read as "parse" or "parser". Given that large parts of
original POSTGRES were written in Lisp, it's a tad hard to believe
that whoever chose those names had not heard of the "predicate"
convention. I guess he/she figured that it didn't matter because
there'd be very little overlap or scope for confusion, and that
seems to have been borne out over time.

regards, tom lane