From: Yukihiro Matsumoto <matz@...>
Date: 2009-11-09T19:13:28+09:00
Subject: Re: Hash#count

Hi,

In message "Re: Hash#count"
    on Thu, 5 Nov 2009 01:19:13 +0900, Intransition <transfire@gmail.com> writes:

|Current:
|
|  {:a=>1,:b=2,:c=>1}.count([:a,1]) #=> 1
|  {:x=>1,:b=2,:c=>1}.count([:a,1]) #=> 0
|  {:a=>1,:b=2,:c=>1}.count(1)      #=> 0
|
|Proposed:
|
|  {:a=>1,:b=2,:c=>1}.count(1) #=> 2
|  {:x=>1,:b=2,:c=>1}.count(1) #=> 2
|
|It would be more useful that way.

That is based on viewpoint to see hashes are collections of values
indexed by arbitrary object.  But in reality, Ruby hashes are designed
as collections of key-value pair.  I am not particularly against the
object-indexed view, but I am not going to change the existing method
behavior.

|And, in the same vain, a #has_pair?
|seems like an obvious addition.

Let's address the issue one by one.

							matz.