From: Robert Klemme <shortcutter@...>
Date: 2011-06-28T21:57:45+09:00
Subject: Re: Symbol#===

On Sat, Jun 25, 2011 at 9:15 PM, Intransition <transfire@gmail.com> wrote:
>
>
> On Jun 25, 1:00 pm, Robert Dober <robert.do...@gmail.com> wrote:
>> On Sat, Jun 25, 2011 at 7:57 AM, Intransition <transf...@gmail.com> wrote:
>> > As:
>>
>> >    class Symbol
>> >      def ===(object)
>> >        object.respond_to?(self)
>> >      end
>> >    end
>>
>> > Then:
>>
>> >    case foo
>> >    when :some_method
>> >      foo.some_method
>> >    end
>>
>> > Insanity?

Maybe, maybe not...  The two of me can't really agree... :-)

>> No, but too invasive...
>>
>> case my_token
>> when :integer
>> when :identifier
>> when....
>>
>> thanx for breaking that ;)
>
> lol. let me "fix" it for you
>
>    class Symbol
>      def ===(object)
>        return super(object) if Symbol === object
>        object.respond_to?(self)
>      end
>    end
>
> Of course, now no repsond_to? case checking for Symbols!

I'd rather keep this simpler:

class Symbol
  def ===(o)
    equal? o or o.respond_to? self
  end
end

irb(main):012:0> :s === :s
=> true
irb(main):013:0> :length === "foo"
=> true
irb(main):014:0> :nono === "bar"
=> false

>> Furthermore I already feel that Symbol has one of the worst kludges
>> ever conceived cast onto itself, which is #to_proc. On the same
>> account I feel #=== should not be concerned with receiver's methods.
>
> #to_proc that bad?

No, I like it.

Kind regards

robert


-- 
remember.guy do |as, often| as.you_can - without end
http://blog.rubybestpractices.com/