Martijn van Oosterhout <kleptog@svana.org> writes:
> On Tue, Oct 28, 2008 at 01:50:26PM -0400, Tom Lane wrote:
>> ... So it might be possible to fix
>> by attaching some new precedence level to the ROWS token.
> Yes. Bison's default is to shift, which means that if you do nothing it
> will treat ROWS as part of the expression if it makes any sense at all.
> Given the requirement for a following UNBOUNDED or BETWEEN, the only
> problem is that you'll get a syntax error if the expr_list ends in a
> postfix operator, I don't see how you get hidden ambiguity.
Hmm, now I see what you meant; that's a little different than what I was
envisioning. I was thinking of trying to force a parse decision that
would support the windowing syntax, whereas you propose forcing a
parse decision that does the opposite, and making the user parenthesize
if he's got a conflict.
What the choice seems to come down to is making ROWS and RANGE reserved
(in some form or other) versus creating a corner case for users of
postfix operators. Phrased that way it does seem like the second
alternative is better.
Hitoshi: you can probably make this happen by including ROWS and RANGE
in the %nonassoc IDENT precedence declaration, but you'll want to test
to make sure the right things happen.
regards, tom lane