Markdown decoder and encoder improvements

[ambrose-gladwell-sonocent]

This change adds support for creating well-formed delta and attributes when decoding markdown. This means encoding a document to markdown then decoding again will produce the same document. This change also makes the encoder and decoder handle empty markdown documents and markdown documents with only line breaks in it correctly.

This PR builds on the current open PR to add a Markdown decoder: https://github.com/memspace/zefyr/pull/191

[hyouuu]

Thanks for this @ambrose-gladwell-sonocent ! Hopefully the owner of the repo would get back to the project soon, but nonetheless your code definitely would help on my project!