Bruce Momjian <bruce@momjian.us> writes:
> On Sun, Nov 26, 2023 at 11:13:39AM +0100, Shay Rojansky wrote:
>> Is there a missing line in the operator precedence table in the docs?
> I think the big question is whether AT TIME ZONE is significant enough
> to list there because there are many other clauses we could potentially
> add there.
Comparing the precedence list in the grammar with the doc table,
the only omissions I feel bad about are AT and COLLATE. There's
a group of keywords that have "almost the same precedence as IDENT"
which probably don't need documentation; but these are not in that
group.
I am, however, feeling a little bit on the warpath about the
grammar comments for the SQL/JSON keyword precedences:
/* SQL/JSON related keywords */
%nonassoc UNIQUE JSON
%nonassoc KEYS OBJECT_P SCALAR VALUE_P
%nonassoc WITH WITHOUT
Every other case where we're doing this has a para of explanation
in the block comment just below here. These not only have no
meaningful explanation, they are in the wrong place --- it looks
like they are unrelated to the block comment, whereas actually
(I think) they are another instance of it. I consider this
well below project standard.
regards, tom lane