From: "A. S. Bradbury" <asbradbury@...>
Date: 2006-09-14T05:18:32+09:00
Subject: Re: How about Enumerable#find_pattern?

On Wednesday 13 September 2006 21:03, Rick DeNatale wrote:
> On 9/12/06, A. S. Bradbury <asbradbury@tekcentral.org> wrote:
> > On Tuesday 12 September 2006 19:50, Rick DeNatale wrote:
> > > Nice, but it can fall down if the pattern contains a regexp which
> > > matches more than one element:
> > >
> > > a = "aaaabbabbab".split(//)
> > > a.find_pattern /aab/,  'b'  #=> nil
> > >
> > > There's probably a way to fix that, but I'm not sure I can see how.
> >
> > I'd actually argue that's expected behaviour and you're just using it
> > wrong(!). /aab/ simply doesn't match any element, so it *should* fail.
>
> Well, I'd expect that if you allow regexps that you should allow
> regexps! But see below.

But why would it be expected behaviour for a regexp to try to match across 
multiple items? The regexp is applied to a single item of the Enumerable 
object, surely that's the only sensible way to deal with this?.

%w[This is an example].find_pattern /an example/    #=> false. 
Surely you wouldn't expect the above snippet to automagically match?

> > The Enumerable mixin provides #to_a, but relying on this seems like a
> > poor solution...
>
> Not sure why.

I'm just thinking that then you have to build the array in memory. It's never 
going to be a big deal unless you have *really large* arrays. It just seems 
cleaner to seek a solution that examines a single member of the Enumerable 
object at a time.

> > The string approach really doesn't seem like the right way to go about
> > this, if I understand you correctly. Not for the problem I'm trying to
> > solve at least. The idea of #find_pattern is that it will work for any
> > arbitrary object.
>
> But then regexps won't work except for a string.

Well no, #find_pattern isn't meant to be tied to regexp matching. You just 
could use it for that. I'm not sure I'm really following you that well on 
this point.

> Also, strings are a funny kind of enumerable here, since by default,
> they just yield themselves in each, unless they include newlines.  I'd
> think that the general use of find_pattern in a string would be to
> search for the pattern in the string, not as a pattern of the lines in
> the string.

Well, you'd use it in a way that is meaningful for your purposes. I'm 
basically using this to find patterns in an array of tokens. If my 
tokenization rules were simple, perhaps a regexp could be compiled, but it's 
really not possible in this situation.

Alex