David Brochart

Ah yes I forgot.

I just [bumped jupyter-ydoc v3.0.0](https://github.com/jupyterlab/jupyter-collaboration/pull/388), and I am going to release jupyter-collaboration v3.0.0 final.

I'm wondering if it is a display message issue. Could you try and see if `print(ds_to_reproject)` has the same issue (that would make it a stream message)?

Note that in Observable notebooks, creating a new notebook starts with an empty cell, and deleting it just clears its content. This is the behavior @krassowski described in https://github.com/jupyterlab/jupyterlab/pull/13141#issuecomment-1260785653, and...

Commenting out this line solves the issue: https://github.com/jupyterlab/jupyterlab/blob/eb678f70df0376080df2e80d47b74a989ac552c9/packages/rendermime/src/outputmodel.ts#L188 But I'm not sure about the consequences.

The problem is in the Mermaid MIME renderer: https://github.com/jupyterlab/jupyterlab/blob/eb678f70df0376080df2e80d47b74a989ac552c9/packages/mermaid/src/mime.ts#L92-L106 And more precisely the call to `model.setData` which leads to a circular update between two clients.

Actually this is not related to top-level await, the kernel dies even while interrupting this: ```py import time time.sleep(100) ```

> Do you have any benchmarks to give an idea when the benefits start to materialize and how much of a speed up one could expect? It's rather on the...

> If the change in behaviour for overwriting characters is a requirement rather than a work-in-progress state it would need to be a really substantial benefit to consider breaking the...

The way shared output changes are handled currently is not optimal: any change in the Y model results in removing all previous outputs from the modelDB and adding all new...