The page is rST-parsed once in 'scan' and once in 'htmlize' (the first to generate...