From: Ian Macdonald <ian@...>
Date: 2007-02-16T18:30:46+09:00
Subject: Re: Puzzling regex behaviour

On Fri 16 Feb 2007 at 00:40:08 +0900, Robert Klemme wrote:

> On 15.02.2007 16:19, Ian Macdonald wrote:
> 
> >Look at this:
> >
> >$ irb
> >irb(main):001:0> foo = "pr�f�r�es"
> >=> "pr\351f\351r\351es"
> >irb(main):002:0> foo =~ /[^[:alnum:]]/
> >=> nil
> >
> >$ irb --noreadline
> >irb(main):001:0> foo = "pr�f�r�es"
> >=> "pr\351f\351r\351es"
> >irb(main):002:0> foo =~ /[^[:alnum:]]/
> >=> 2
> >
> >This is _very_ unexpected and undesirable behaviour and, as such,
> >probably qualifies as a bug.
> 
> Yeah, seems so.  Unless it's documented behavior. :-)
> 
> >Interestingly, adding "require 'readline'" to the stand-alone script
> >does _not_ introduce this behaviour, so it must be something to do with
> >the initialisation that irb does.
> 
> It's really strange as both print the same output.

You mean that both of them show foo to contain the same string of bytes?

> How about doing this 
> - just to be sure that both strings contain the same sequence of bytes:
> 
> require 'enumerator'
> foo.to_enum(:each_byte).to_a.join(", ")

In both cases:

=> "112, 114, 233, 102, 233, 114, 233, 101, 115"

Somehow, it is the regex that is being handled differently, not the
string.

Ian
-- 
Ian Macdonald               | Reality does not exist -- yet. 
ian@caliban.org             | 
http://www.caliban.org/     | 
                            | 
                            |