From: James Coglan <jcoglan@...>
Date: 2009-08-18T22:30:32+09:00
Subject: Re: Bizarre Range behavior

--001517448064a725cf04716a854e
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

2009/8/18 James Coglan <jcoglan@googlemail.com>

>
>
> 2009/8/18 Yukihiro Matsumoto <matz@ruby-lang.org>
>
> Hi,
>>
>> In message "Re: Bizarre Range behavior"
>>     on Tue, 18 Aug 2009 21:31:38 +0900, Daniel DeLorme <dan-ml@dan42.com>
>> writes:
>>
>> |That sounds nice and consistent, but why limit ourselves to numbers?
>> |     ("y".."aa").to_a
>> |should then generate ["y", "z", "aa"] since that's what String#succ
>> |generates. Sounds like this could snowball into a nightmare of
>> |incompatible changes.
>>
>> Indeed.  The issue is that it's not easily distinguishable whether
>> a sequence using #succ from a string reaches to another, or not.
>> If there's affordable scheme to implement, I'd love to check in.
>
>
> For user-defined objects, the `<=>` method is used to determine whether the
> end of a range has been reached or exceeded, and it seems like 1.8 does this
> for strings (and numbers, of course). I prefer 1.8's consistent application
> of this rule.
>

Actually I need to qualify this: the rule seems to be that if `(a <=> b) ==
1` and `a.length >= b.length` then b is not reachable using a `succ`
sequence from a.

--
James Coglan
http://jcoglan.com

--001517448064a725cf04716a854e--