use warnings;
use strict;
use HTML::TreeBuilder;
my $Tree = HTML::TreeBuilder->new ();
$Tree->parse ("First para.
Second line.
Second para
");
my $Text;
my $empty_element_map = $Tree->_empty_element_map;
my(@C) = [$Tree]; # a stack containing lists of children
# I is a stack of indexes to current position in corresponding lists in @C
# In each of these, 0 is the active point
my(@I) = (-1); # initial value must be -1 for each list
my @Context = ""; # Contains stack of current nodes
# scratch:
my $this; # current node
my $content_r; # child list of $this
my $TagName;
# Loop over the tree
while (@C)
{
# Post processing
# Move to next item in this frame
if(!defined($I[0]) or ++$I[0] >= @{$C[0]})
{
$this = $Context [0];
if (defined $this and ref $this)
{# Close tag
my $StartTag = $this->starttag ();
$StartTag =~ s/[\r\n]*//gs;
if ($StartTag =~ /^/g)
{
$Text .= "";
}
}
shift @Context;
shift @I;
shift @C;
next;
}
$this = $C[0][$I[0]];
if (! ref $this)
{# Add the text
$Text .= $this;
}
else
{# Process this element
my $StartTag = $this->starttag ();
$StartTag =~ s/[\r\n]*//gs;
if ($StartTag =~ /^(<(?:p|br)\b.*?>)/g)
{
$Text .= "$1"
}
}
# Now queue up content list for the current element...
if(
ref $this and
not
( # ...except for those which
not($content_r = $this->{'_content'} and @$content_r) and
# ...have empty content lists
$this->{'_empty_element'} || $empty_element_map->{$this->{'_tag'} || ''}
# ...and that don't get post-order callbacks
)
)
{
unshift @Context, $this;
unshift @I, -1;
unshift @C, $content_r || [];
}
}
print $Text;
####
First para.
Second line.
Second para