From: Robert Klemme <bob.news@...>
Date: 2005-04-02T16:39:42+09:00
Subject: Re: Synchronous but different iterators


"Belorion" <belorion@gmail.com> schrieb im Newsbeitrag 
news:a48d774d050401150863a424fe@mail.gmail.com...
> (somewhat contrived example:) Say I have 2 different data structures
> which I want to iterate over, and Tree, and an Array.  I had to write
> specific iterators for my Tree structure, such as depth_first,
> breadth_first, etc.  Is there a good way to iterate over each
> structure in a synchronous manner, but with a different iteration
> style?
>
> For example, I want to step through my Array normally using
> each_index, but want to step through my Tree structure using, say,
> depth_first, but I want to do this concurrently because of what I want
> to do to/with the data.
>
> "pseudo" ruby code:
> (myTree.depth_first.each, myArray.each_index){ |tree_node, array_element|
>    # do stuff with each in tandem
> }
>
> Is there a trick to doing this?  Or is my best bet to iterate over my
> Tree object and create a temporary array, and then just iterate in
> step using with_index?
>
> tmpArr = []
> myTree.depth_first{ |node| tmpArr.push node }
>
> myArray.each_index{ |ii|
>    tree_node, array_element = tmpArr[ii], myArray[ii]
> }
>
> The first approach would raise some questions such as what to do if
> one iterator reaches the end before the other ... and one would have
> to decide wether the "next" object in that iteration is nil, or if the
> synchronous iterator terminates whenever the shortest iteration
> reaches the end.
>
> Obviously, the (2nd) working ruby solution is not that difficult, but
> I am more curious from an academic standpoint if the first approach is
> possible through some fancy yield tricks.

You should look at Generator of the standard lib.  That does exactly what 
you want:
http://www.ruby-doc.org/stdlib/libdoc/generator/rdoc/index.html

Kind regards

    robert