Repo status: has this been abandoned? Are there active forks?

[derekperkins]

We're about 2.5 years out from the last commit. There are seemingly viable PRs to be merged, but no activity from maintainers. Is there an active fork somewhere?

[williamstein]

Hi @derekperkins for what it is worth I care about this library and use it in production for all realtime sync for https://cocalc.com. I maintain a slightly modified dmp.js here

https://github.com/sagemathinc/cocalc/blob/master/src/packages/util/dmp.js

with some minor fixes for edge cases I've hit. I may end up creating a fork at some point to revive this project. I've also been thinking about writing a new implementation of the algorithm in zig (+compiled to wasm), but we'll see if that makes sense, as Javascript is pretty fast.

[burnit66]

Hi @williamstein I'm curious what edge cases you've run into? I am encountering an issue when comparing superscripts. Interested to see if this case is covered in your modifications.

Text 1
(I(X))u
Text 2
(I(X))<sub>u</sub>
Compare output
(I(X))<ins><s</ins>u<ins>b>u</sub></ins>

As you can see, the html tag for subscript has been flagged as a diff the return is broken html

[williamstein]

I'm curious what edge cases you've run into?

They were issues involving the timeout implementation, and I think also some really bad performance in edge cases, e.g., shifting a 1MB string over by 1. When using this algorithm in production, especially on a server, these become important.

Thanks for sharing your example. It would be easier for me to test my code on it if you could give a snippet of javascript using dmp that I can just run, importing my slightly modified dmp instead of the official one. If it really is a bug, I'll be very interested in that and hopefully maybe have an idea how to fix it.