simplygreatwork
simplygreatwork on Hacker News
It's pretty easy to create @mentions by creating and inserting a custom "atom" using the function "insert_atom" to insert the atom's HTML. And then you can do further initialization by listening to "atom-will-enter".
I need to go back and turn on the ability for sections inside atoms and cards to be contentEditable=true to be able to type text directly inside the element. During development, I had disabled all mutation observation for cards and atoms so as to not corrupt undo history. But I will work to turn editability back on soon within atoms and cards - but with precision.
I had dabbled with creating a basic editor setup in ProseMirror but overall even ProseMirror itself felt a bit too opinionated for my taste. Another usability problem was that the ProseMirror example was implemented as a module instead of as a top level project. So I had started a buildless ProseMirror example but ultimately I decided to move on anyway: https://github.com/simplygreatwork/prosemirror-buildless
Show HN: Godsend: modular separation of concerns for streamable micro-services
Yes, you could log every message by calling request.accept() with no criteria and applying a weight of -100, for example.
By default, only one instance of a processor (transformer) id is permitted - unless you mount additional versions. In "godsend-examples", there's an example named "versioning-processors". Selection of processor versions is determined in the profile of the user sending the request. In that example, in "users.json", the user "sender" is configured with any processor versions which ought to be substituted for that user. If there are multiple versions of a processor and none of those is indicated by the user sending the message, the processor version marked as the default, when it was mounted, will be run.