use strict;
use warnings;
use HTML::Entities qw( encode_entities );
use HTML::Tagset qw( );
use HTML::TreeBuilder qw( );
use Object::Destroyer qw( );
sub extract_html {
# It would be better if we had access to the unparsed text,
# but this will do as long as the parser doesn't change.
my $html = '';
local *helper = sub {
my ($node) = @_;
if (!ref($node)) {
$html .= encode_entities($node);
return;
}
my $tag = $node->tag();
$html .= $node->starttag();
helper($_) for $node->content_list();
$html .= $node->endtag()
if !$HTML::Tagset::emptyElement{$tag}
&& !$HTML::Tagset::optionalEndTag{$tag};
};
my $node = @_ ? $_[0] : $_;
helper($_) for $node->content_list();
return $html;
}
{
my $tree = HTML::TreeBuilder->new();
$tree = Object::Destroyer->new($tree, 'delete');
$tree->parse_content(<<'__EOI__');
Foo
Foo
Not bar
__EOI__
print extract_html(
$tree->look_down( '_tag' , 'body' )
);
}
####
Foo
Not bar