From: Christoph <chr_news@...>
Date: 2002-04-30T19:45:17+09:00
Subject: Re: Numberic#prev

"Yukihiro Matsumoto" <matz@ruby-lang.org> wrote in
....
> Hi,
>
> In message "Re: Numberic#prev"
>     on 02/04/28, "Christoph" <chr_news@gmx.net> writes:
>
> |On the other hand what is useful about
>
> I'm not really sure what you meant.

I guess my point was that forming Ranges from non-Integer Numeric
arguments is (in my mind;-) a bit like declaring

class Numeric
     def prev
        to_i - 1
    end
    def succ
        to_i + 1
    end
end

- both construction might be occasionally useful but they don't seem to
be useful enough (imho;-) to elevate them to standard behavior.


>
> |    (2.2..4.3)  == (2..4)    # => false
> |    (2.2...4.3)  == (2...4)  # => false
> |    (2.2..4.3).to_a            # =>  [2,3,4]
> |    (2.2...4.3).to_a           # => [2,3].
>
> Even if they produce same integer arrays by "to_a", it does not mean
> they are same.

True - however

---
(2.2..4.3).to_a.member? 2 # => true

# but

(2.2..4.3).member? 2  # =>  false


require 'complex'

(2..Complex(-4)).member? 3  # => true
---

All in all:  The way Range and Numeric classes interact is plainly
counter intuitive too me.


/Christoph