Add support for custom markdown parsers

[adriansalamon]

Use case: I want to use a custom markdown parser (md) instead of Earmark, in order to use my own markdown format with support for HEEx inside of markdown.

Example usage:

  use NimblePublisher,
    ...
    markdown_parser: MarkdownParser,

defmodule MarkdownParser do
  def parse(body) do
     "<p>This is a custom markdown parser :)</p>\n"
  end
end

[josevalim]

What if we call it html_converter? This could put us in a path to support other formats beyond markdown in the future. Also, should we pass the filename as argument? I can see matching on the extension being useful?

[josevalim]

Actually, we should make it the same API as the convert_body function. Receive the extension, body, and opts.

[adriansalamon]

Actually, we should make it the same API as the convert_body function. Receive the extension, body, and opts.

👍 This makes sense, since then users would be free to define what files to process without having to change the lib. Eg. maybe they want to use .mdx files or similar.

[josevalim]

This is tricky. This is private API we would need to expose it but I am not sure if we should. In particular, I am not sure if the regex it uses would match other engines.

[adriansalamon]

I agree, I did a basic test with md and the regex worked, but as you said there is no guarantee that it should work. Because this is a "custom" option, if using a engine which it is not compatible with, you can always implement your own highlighting?

[josevalim]

Yeah, but how do you feel if you have to implement that module yourself? Maybe we make the Regex an argument to the highlight function and make it public?

[josevalim]

It is looking great. I have suggested some nitpicks on the option name and API. :)

[josevalim]

Are you sure we need this code? We already start all highlighters before, on __extract__, no?

[adriansalamon]

What if someone calls this function without having called use NimblePublisher ...?

[josevalim]

💚 💙 💜 💛 ❤️