Re: zero? & 0 | 1

• To: "Tim Bradshaw" <tfb@cley.com>
• Subject: Re: zero? & 0 | 1
• From: "Gabor Greif" <gabor@no.netopia.com>
• Date: 13 Mar 2000 12:22:03 +0100
• Cc: info-dylan@ai.mit.edu
• Sender: owner-info-dylan@ai.mit.edu

On Mon, Mar 13, 2000 12:11 Uhr, Tim Bradshaw <mailto:tfb@cley.com> wrote:
>* Gabor Greif wrote:
>> The reasoning is that binary operators are evaluated from left to right
and
>> &, | only evaluate their rhs if necessary.
>
>I don't think this has to follow.  I don't know dylan that well (and
>I've no implementation to hand to try this on), but, for instance,
>what does 1 + 2 * 3 evaluate to?  My guess is 7, not 9.  So
>bool-value? & 0 | 1 could easily be (bool-value? & (0 | 1)).
>
>--tim
>

Thanks for the critique, I should have added that DRM defines | & and := as
having same precedence and lower than any other (binary) operator.

This is probably why the Dylan folks did not cry out loud :-)

Also everything ~== #f on the lhs of & triggers eval of its rhs.

Now the reasoning should infer the identity of

bool-value? & 0 | 1

with

(bool-value? & 0) | 1

	Gabor

• Prev: Re: zero? & 0 | 1
• Next: How can I..
• Index(es):
  • Main
  • Thread