From: Dave Thomas <dave@...>
Date: 2008-03-18T22:44:26+09:00
Subject: Re: Range#member? semantics seem wrong


On Mar 17, 2008, at 11:14 PM, David Flanagan wrote:

> Actually your sample code also returns true in both 1.8 (at least  
> 1.8.6) and 1.9.0.

Checking, I agree. I wonder if this is a change since an earlier 1.8.

> And 1.9.0 has also introduced a new method named cover?.
>
> My understanding (and it takes me all of page 70 to explain this in  
> my book) is that in 1.9 include? and member? test what I call  
> "discrete membership" (and what you call membership) using succ when  
> the endpoints of the range are not numbers.  For numbers they test  
> continuous membership using <=>.  And the new method cover? always  
> uses the continuous membership test using <=>.

I use membership in the 'set' sense. And there doesn't seem to be a  
way of determining whether the range of integers  (1..3) contains 2  
(set membership) but not 2.5 (range inclusion). And that seems  
unfortunate. I'd like to see member? do what it's name suggests and  
work on a set-like basis, so that

   (1..3).member?(2) => true
   (1..3).member?(2.5) => false



Dave