On Mon, Jul 20, 2009 at 07:33:23PM -0700, Jon Lang wrote:
: A stronger argument against it would be to find comparison operators
: that exist at other precedence levels.  I don't think that there are
: any.  (Well, besides <=>, leg, and cmp.)

I think people would find it quite odd if an operator named "cmp" can't
be considered a comparison operator.

: Indeed, I'll be surprised if there are any other precedence levels
: that are chaining - which is another key point: what kind of operator
: would be chaining, but _not_ some sort of comparison?

Congratulations, with that argument you've completely persuaded me that
"Chaining" is the most precise description of that precedence level,
and we should leave it just the way it is now.

The reason I renamed "Nonchaining" is not that I dislike the concept
of chaining, but that all the *other* non-chaining precedence levels
are also, er, non-chaining. :)

Though the proximate cause was that it made for a stupid error message,
which indicated that the old name wasn't sufficiently accurate:

  Can't reduce a nonchaining operator because it's diffy and not chaining

It now reads:

  Can't reduce cmp because structural infix operators are diffy and not chaining

In short, "chaining" is a good concept, but "non-chaining" is bad
(and doubly bad when it means two different things).


Reply via email to