From: Jeff 'japhy' Pinyan <jeffp@...>
Date: 2002-02-07T07:02:57+09:00
Subject: Re: Regexp::Parser ported to Ruby...?

[posted & mailed]

On Feb 6, Tobias Reif said:

>Jeff 'japhy' Pinyan wrote:
>
>>   "japhy" =~ /((.)(?{ print "<$2>" }))+/;
>>   # <j><a><p><h><y>
>
>One way:
>
>irb(main):003:0> "japhy".gsub(/./)do print "<#{$&}>" end
><j><a><p><h><y>""

Ok, let me be a bit more specific.  A common request is to be able to have
NAMED captures.  In Perl regexes, "japhy" =~ /(..)(..)$/ stores "ap" to $1
and stores "hy" to $2.  Some people would like the ability to assign
DIRECTLY to some variable, like:

  "japhy" =~ /(?<$front>..)(?<$back>..)$/;

Well, this syntax isn't possible in Perl (yet).  However, with my regex
parser, it is.  My regex parser will parse the non-existent (?<VAR>PAT)
construct and turn it into working Perl regex code.  I present it in an
expanded format for legibility (?):

  "japhy" =~ m{

    (?{ local (@S, @E) })
      # creates two local arrays for use in the regex only

    (?{ local $S[0] = $+[0] })
      # our current position in the string
    ..
    (?{ local $E[0] = $+[0] })
      # our current position in the string

    (?{ local $S[1] = $+[0] })
      # our current position in the string
    ..
    (?{ local $E[1] = $+[0] })
      # our current position in the string

    $

    (?{ $front = substr($&, $S[0], $E[0] - $S[0]) })
    (?{ $back  = substr($&, $S[1], $E[1] - $S[1]) })

  }x;

The local() makes sure that when backtracking is done, values are restored
as needed, etc.  The end result is a regex that stores "ap" in $front, and
"hy" in $back.

-- 
Jeff "japhy" Pinyan      japhy@pobox.com      http://www.pobox.com/~japhy/
RPI Acacia brother #734   http://www.perlmonks.org/   http://www.cpan.org/
** Look for "Regular Expressions in Perl" published by Manning, in 2002 **
<stu> what does y/// stand for?  <tenderpuss> why, yansliterate of course.