From: Logan Capaldo <logancapaldo@...>
Date: 2005-12-20T01:04:56+09:00
Subject: Re: TheYAML change the id of Object


On Dec 19, 2005, at 10:47 AM, Ross Bamford wrote:

> On Mon, 19 Dec 2005 15:21:22 -0000, Logan Capaldo  
> <logancapaldo@gmail.com> wrote:
>
>>
>> On Dec 19, 2005, at 9:10 AM, Andrea wrote:
>>
>>> I'm using an hash object and i use a Termine Object for the key  
>>> and, for
>>> now, a String for the value.
>>>
>>
>> You need to overload the == operator and the hash method.
>>
>> class Termine
>>     alias eql? ==
>>     def ==(other)
>>            descrizione == other.descrizione and lingua == other.  
>> lingua
>>     end
>>     def hash
>>          (descrizione + lingua).hash
>>     end
>> end
>>
>> t = Termine.new("Hi")
>> t2 = Termine.new("Hi")
>>
>> t == t2 #=> true
>>
>> h = {}
>> h[t] = "short hello"
>> h #=> {#<Termine:0x24a22c @lingua="it", @descrizione="Hi">=>"short  
>> hello"}
>> h.has_key?(t2) #=> true
>>
>
> I was playing around with this, and now I'm doubly worried I have a  
> bug, or have missed something else. The last line in your code  
> gives out 'false' for me. I was experimenting with this:
>
> class Dclz
>   @str = "A string"
>   attr_reader :str
>
>   def ==(o)
>     self.str == o.str
>   end
>
>   def ===(o)
>     self.str === o.str
>   end
>
>   def hash
>     self.str.hash
>   end
> end
>
> case Dclz.new
> when Dclz.new
>   puts "Works"
> else
>   puts "doesn't"
> end                                   # => "Works"
>
> h = Hash.new { "Failure" }
> h[d = Dclz.new] = "Expected"
>
> puts Dclz.new == Dclz.new             # => true
> puts Dclz.new.hash == Dclz.new.hash   # => true
>
> puts h[d]                             # => "Expected"
> puts h[Dclz.new]                      # => "Failure"         
> huh????!!!!
>
> And I cannot understand what's happening here. Someone, please  
> confirm that I'm either sane or stupid?
>
> (Tried on both ruby 1.8.3 (2005-09-21) [i386-linux] and ruby 1.9.0  
> (2005-12-16) [i686-linux])
>
> -- 
> Ross Bamford - rosco@roscopeco.remove.co.uk
>

As ts already mentioned, you missed the eql? part. Me aliasing it to  
== and implementing == instead is a function of my C++ (as opposed to  
lisp or smalltalk(?) ) upbringing.