From: Brian Candler <B.Candler@...>
Date: 2004-10-09T20:51:30+09:00
Subject: Re: Range behavior (Re: [RCR] New [] Semantics)

On Sat, Oct 09, 2004 at 08:14:42PM +0900, gabriele renzi wrote:
> Randy W. Sims ha scritto:
> 
> >Or better. Some classes could implement an optional method (a mixin 
> >interface?), say find_in_range, so that if the method exists for the 
> >type of objects in the range it is called for an efficient lookup. If 
> >the object does not support the method, it can be done the old way using 
> >obj#succ. 
> 
> isn't <=>+checking for same class enough for this?

At the moment,

  (3..5).member?(3.5)    #=> false
  (3..5).member?(4.0)    #=> true

so following your rule exactly wouldn't duplicate the current behaviour.

You'd need some way of saying "is there some Integer x where x == 4.0"

Or I suppose more generally:

  (a..b).member?(c)
is true if a <= b <[=] c
and there exists some d such that c == d and d.instance_of?(a.class)

Except - some ranges change class as they iterate, e.g. if 'a' is a Fixnum
and 'b' is a Bignum. Ugh!

Regards,

Brian.