Luke Palmer > Regexp-Subst-Parallel-0.11 > Regexp::Subst::Parallel

Download:
Regexp-Subst-Parallel-0.11.tar.gz

Dependencies

Annotate this POD

View/Report Bugs
Module Version: 0.11   Source  

NAME ^

Regexp::Subst::Parallel - Safely perform multiple substitutions on a string in parallel.

VERSION ^

Regexp::Subst::Parallel version 0.11, Feb 9, 2003.

SYNOPSIS ^

    # Rephrase $str into the form of a question.
    my $qstr = subst($str,
                   qr/I|me/  => 'you',
                   qr/my/    => 'your',
                   qr/mine/  => 'yours',
                   qr/you/   => 'me',
                   qr/your/  => 'my',
                   qr/yours/ => 'mine',
                   ...);
    
    # Apply implicit html highlighting
    my $html = subst($text,
                   qr/\{(.*?)\}/ => '$1',  # Protect things in braces
                   qr/_(\w+)_/   => '<u>$1</u>',
                   qr/<(\w+)>/   => '<i>$1</i>',
               );

    # Toggle the case of every character
    my $vAR = subst($Var,
                  qr/([a-z]+)/ => sub { uc $_[1] },
                  qr/([A-Z]+)/ => sub { lc $_[1] },
              );

DESCRIPTION ^

Regexp::Subst::Parallel is a module that allows you to make multiple simultaneous substitutions safely. Using the sole exported subst function has a rather different effect from doing each substitution sequentially. For example:

    $text = '{process_the_data} was _called_ without <data>!';
    $text =~ s/\{(.*?)\}/$1/g;
    # $text eq 'process_the_data was _called_ without <data>!'
    $text =~ s/_(\w+)_/<u>$1</u>/g;
    # $text eq 'process<u>the</u>data was <u>called</u> without <data>!'
    $text =~ s/<(\w+)>/<i>$1</i>/g;
    # $text eq 'process<i>u</i>the</u>data was <i>u</i>called</u> without <i>data</i>!'

Which is clearly the wrong result. On the other hand, Regexp::Subst::Parallel does them all in parallel, so:

    $text = '{process_the_data} was _called_ without <data>!';
    $text = subst($text,
                qr/\{(.*?)\}/ => '$1',  # Protect things in braces
                qr/_(\w+)_/   => '<u>$1</u>',
                qr/<(\w+)>/   => '<i>$1</i>',
            );
    # $text eq 'process_the_data was <u>called</u> without <i>data</i>'

Which seems to be right.

The algorithm moves from left to right, and the longest match is substituted in case of conflict. The substitution side of the pairs can either be a string, in which non-backslashed $n's are substituted, or a coderef, in which the sub is called and passed the list of captures in @_. $_[0] is analogous to $& : it refers to the entire match.

Gotchas

Make sure when you're using the string method to have the $'s included in the string. That means if you're using an interpolating quote ("", qq{}, etc.) that you backslash $1, $2, etc. Otherwise you will get the $n's from the current lexical scope, which is not what you want.

Caveats

To include a single backslash followed by an interpolated capture, subst needs to see '\\$1', which means that you have to type '\\\\$1' when you just want a single backslash. That's sick.

AUTHORS ^

Luke Palmer <fibonaci@babylonia.flatirons.org>

COPYRIGHT ^

Copyright (C) 2003 Luke Palmer. This module is distributed under the same terms as Perl itself.

    http://www.perl.com/perl/misc/Artistic.html
syntax highlighting: