I have a Blogger-based weblog, with special DIV tags around the individual posts, so I can parse the resulting file. The tags are like <div CLASS="post" ID="5736047" UNUSEDATTRIBUTE="9/17/2001 01:31:37 PM"> and can contain any HTML (although usually it's fairly basic formatting).

I have an HTML::Parser based search function for my site that can parse these files, but as a side effect, I lose all HTML tags between the DIVs. At the moment, that's OK because I can then grep more reliably. However, for a different project, I'd like the same thing but with the original HTML between the DIVs - how does this work in HTML::Parser 3? I have to confess I don't really understand how the parser is structured...

Here's the current guts of the search, basically adapted from one of the examples:

my $p = HTML::Parser->new(api_version => 3, start_h => [\&div_start_ha
+ndler, "self,tagname,attr"]);

$p->parse_file($file);
do_stuff();

sub div_start_handler
{
    my($self, $tag, $attr) = @_;
    my($blogdate);  

    return unless ($tag eq "div");
    return unless exists $attr->{class};
    return unless $attr->{class} eq 'post';

    #global, so the endhandler knows what the last ID was
    $blogid = $attr->{id};
    $BlogArticles{$blogid}{DateText} = $attr->{unusedattribute};

    $self->handler(text  => [], "dtext" );
    $self->handler(end   => \&div_end_handler, "self,tagname");
}

sub div_end_handler
{   
    my($self, $tag) = @_;

    return unless $tag eq "div";

    my $text = join("", map $_->[0], @{$self->handler("text")});
    $BlogArticles{$blogid}{BodyText} = $text;

    $self->handler("text", undef);
    $self->handler("start", \&div_start_handler);
    $self->handler("end", undef);
}
[download]