The parser now works as expected with no known bugs. As expected from the spec each line of ordinary text is treated as a separate paragraph.
