Our MDAST pipeline essentially translates a Markdown structure to a Lexical structure and vice versa, maintaining a 1:1 correspondence. Although the structure remains lossless in both directions, we cannot say the same for syntax when converting back to Markdown.
examples
links:
`https://stacker.news(/image.jpeg)` -> [https://stacker.news](https://stacker.news)
-> 
inline syntax:
`*` or `_` -> *
lists:
`*` or `-` -> *
These are not huge problems, actually in some cases our pipeline fixes Markdown indentation and style (can't beat the serialization of a tree), and maybe we just need to retain the user syntax choices when converting.
Our MDAST pipeline essentially translates a Markdown structure to a Lexical structure and vice versa, maintaining a 1:1 correspondence. Although the structure remains lossless in both directions, we cannot say the same for syntax when converting back to Markdown.
examples
These are not huge problems, actually in some cases our pipeline fixes Markdown indentation and style (can't beat the serialization of a tree), and maybe we just need to retain the user syntax choices when converting.