
#!/usr/bin/perl

# Shows how to mangle all plain  text in an HTML document, using an arbitrary
# Perl expression. Plain text is all text not within a tag declaration, i.e.
# not in <p ...>, but possibly between <p> and </p>
# Example (Reverse 'Debian' in all text) :
#       lynx -dump -source -raw http://www/debian.org > /tmp/a.txt
#       ./htextsub '$_ =~ s/Debian/Naibed/gi' /tmp/a.txt

use strict;
use warnings;
use HTML::Parser ();

my $code = shift || usage();
$code = 'sub edit_print { local $_ = shift; ' . $code . '; print }';

#print $code;
eval $code;
die $@ if $@;

my $p = HTML::Parser->new(
    unbroken_text => 1,
    default_h     => [sub { print @_; }, "text"],
    text_h        => [\&edit_print, "text"],
);

my $file = shift || usage();
$p->parse_file($file) || die "Can't open file $file: $!\n";

sub usage {
    my $progname = $0;
    $progname =~ s,^.*/,,;
    die "Usage: $progname <perlexpr> <filename>\n";
}