[scaladoc] fix: Only trim one newline when preprocessing the content of a markdown code snippet#21519