From: David Flanagan <david@...>
Date: 2008-06-12T03:59:27+09:00
Subject: [ruby-core:17225] Re: Enumerable::Enumerator#with_memo

David A. Black wrote:
>> I'm beginning to feel that the method we're discussing here is one 
>> particular member of a set of methods for defining custom iterations. 
>> Imagine something like:
>>
>> a1.each.with_index.and_next_of(a2).and_const(a3).and_returning([]) do
>> |elt1, index, elt2, a3, retval|
>> end
>>
>> Or:
>>
>> [1,2,3].each.alternating_with([4,5,6]).and_returning([]) {|x,ret|...}
>>
>> Does anyone like the use of "and" instead of "with" to indicate the 
>> addition of a new value to be passed to the block?
> 
> I don't particularly. But then again, I'm not a fan of lots of
> chaining that tries to make things sound like English narrative. It
> never really does, and I just lose track of it in the Ruby part of my
> brain.  That's why, as much as I respect the talent and effort behind
> it, I can't use RSpec.
> 
> 
> David
> 

I took the chaining examples over the top here on purpose.  I doubt 
they'd actually get used like that very often.  But it is worth pointing 
out that chained invocations are implicit in this style of looping with 
enumerators.  As in each.with_index and each.with_memo, for example. I 
was just taking this a step further.  And it raises again the question 
of whether there should be an each_with_memo method directly on 
Enumerable instead of or in addition to with_memo on Enumerator.

I believe that Matz has plans for Enumerators in Ruby 2.0 and I'm really 
beginning to feel that with_memo ought to be postponed and reconsidered 
in the context of Ruby 2.0

	David Flanagan