David Brochart

Since the event system is a new service that we are just starting to develop, how about making it as much as possible backend-agnostic? By that I mean that most...

You're right Zach, jupyverse also has implemented specific logic for configuration, and I guess depending on FastAPI makes it kind of specific to this framework too. I'm thinking about some...

I meant something more simple, like [this GET handler](https://github.com/jupyter-server/jupyter_server/blob/3e64fa5eef7fba9f8e17c30cec688254adf913bd/jupyter_server/services/contents/handlers.py#L91) calls [this get method](https://github.com/jupyter-server/jupyter_server/blob/3e64fa5eef7fba9f8e17c30cec688254adf913bd/jupyter_server/services/contents/manager.py#L387). If we can have the logic in the get method in a separate package, that's a great...

Thanks Kevin for the feedback. You're right, this is a bit aggressive, but I wanted to show that things still work when removing everything related to notebooks from the server...

> Removed from where? Sorry, "removed" is not the right word, but JupyterLab in RTC mode doesn't use the REST contents API to load and save documents. It uses the...

I think that in RTC, document validity is almost "built-in": a client which modifies e.g. a notebook does it in a consistent way. You might say that it was already...

I don't understand what you mean, but if you found something wrong, do you mind opening an issue in jupyverse?

Thanks for opening the issue. Jupyverse doesn't implement the current trust mechanism. As it is probably going to change with RTC, we preferred to wait for things to settle. >...

> including fixing `ImportError: cannot import name 'User' from 'fps_auth.models' (.../jupyverse/plugins/auth/fps_auth/models.py)` BTW, you didn't mention this in your issue, mind opening a new issue in jupyverse?

> that's just on the released version. latest git seem fine now. I could not reproduce, mind opening an issue?