Martin Kleppmann

I had a go at an alternative implementation, which seems to work fine with three peers all connected to each other: https://gist.github.com/ept/fe4f772bcc9294c953728b2407f46c77 My version does not deliver messages immediately, but...

Hi @salzhrani, I'm not up-to-date on current best practices for JS bundling/packaging, so I'm happy to take your advice. How would we make this change? Do we require an update...

Hi @camiboj, thank you very much for this contribution! As I mentioned on #438, we currently have a project underway to redesign the API between Automerge's frontend and backend, and...

No problem @zhang123cnn, feel free to make a fork!

This does seem useful, although so far our focus has been on representing history efficiently, so that you can keep history without too much storage or performance cost. See #253...

Thanks for taking the initiative for this change! I've updated the PR to allow to forms of `Text.insertAt()`: ```js // means Automerge should split the string into individual characters text.insertAt(index,...

@pvh If we require any non-string values to be wrapped in an array, should we also change the API for lists to be consistent? That API currently works like this:...

Yes, this code will split text into Unicode scalar values, which means some emoji will get broken up. I have previously argued to use grapheme cluster segmentation, which would keep...

Yes! And it will also need to explain how to convert between the individual changes and the whole-document compression, in both directions.

Hi @kpruden, I was going to suggest exactly the solution that you've already come up with (an array of rows, an array of columns, and a map from (rowId,columnId) to...