From: "Iñaki Baz Castillo" <ibc@...>
Date: 2011-06-09T20:29:37+09:00
Subject: Re: Where is the #hash method defined?

2011/6/9 Piotr Szotkowski <chastell@chastell.net>:
>> Hi, any object in Ruby has the method #hash which
>> returns an integer unique for the object (such value
>> doesn't change during the lifetime of the Ruby process).
>
> Note that #hash’s role is to fulfil the implication
> a.eql?(b) → a.hash == b.hash, but it’s as weak as possible
> (i.e., it would be perfectly legal, if very inefficient,
> for #hash to return the same result for all objects).
>
> In general, the idea is that it’s often possible to have a #hash
> method that computes a unique-ish result much faster than performing
> a full-blown #eql? check; in these cases it’s worthwhile to compare
> #hash results of two objects and call #eql? only if they’re the
> same (equal #hash results do not mean the objects are #eql?, but
> different #hash results do mean that they are not #eql?).
>
> By default, Kernel#hash uses #object_id to compute the hash,
> so it’s quite unique (I’m not sure what happens when #object_id
> rolls over the Fixnum limit; #hash seems to be returning
> a Fixnum anyway, so in theory it can be non-unique).


Thanks for the very good explanation.


>> I cannot find which class or module #hash method belongs to.
>> I expected it could be in Object class, but it's not. Where is it?
>
> It’s defined as Kernel#hash and in Ruby 1.9.2 implemented here:
> https://github.com/ruby/ruby/blob/ruby_1_9_2/object.c#L2516
> → https://github.com/ruby/ruby/blob/ruby_1_9_2/object.c#L99

I see. Ok. Thanks a lot.

-- 
Iñaki Baz Castillo
<ibc@aliax.net>