1) if file not exists, dont parse it 2) more robust parsing with regex 3) include search_domains
searching domains is not used yet, I will supply another patch.
