Markdown differences

[liamwhite]

GitHub supports a "difference mode" in its Markdown support. This can be demonstrated e.g. on this repository here by selecting the "Rich diff":

Capture

The difference algorithm seems to be at least vaguely syntax-aware, so this isn't necessarily handled on its own by just a Myers diff. Is it possible for comrak to support such a difference mode?

[gjtorikian]

I actually worked on this feature when I was at GitHub. It's not a complicated process; I vaguely remember it being a process where the before and after nodes of the HTML trees were walked, with a diff on children that were added or removed.

I don't know whether comrak should support this natively per se, but one can already walk through the AST on their own. Would a code sample showing that be helpful? How would you ultimately use this feature?

[liamwhite]

A code sample would be very helpful, yes. :)

My ultimate plan is to use it in a similar way to how GitHub uses it, for highlighting differences between different versions of forum posts/comments,

[liamwhite]

@gjtorikian any update on this?

[gjtorikian]

Sorry, I haven't been able to give this much focus to generate a full example.

First, I'm going to close the issue, since I think this is out of scope.

Second, here's an example of how you can walk a parsed AST: https://github.com/kivikakk/comrak/blob/b27a3dd0a16c2960aad706eefa80f4c1977fca5e/examples/traverse_demo.rs

extract_text_traverse shows how you can pass the root node to a function and walk down its descendants. In your case, you'd walk down two trees. Keep track of whichever nodes you encounter; you can store them as a flat list, like: [:document, :list, :item, ...]. Then, just compare the two arrays: if they've changed, you can represent that information as HTML or text or however necessary. Let me know how that works!