Frédéric Collonval

Thanks @itsmevichu I'm more and more convinced that we should stop increasing our imperfect parsing rules to use a more robust approach based on the tokenizer of the markdown renderer....

> Of note we are more and more parsing Markdown again in the table of contents but only what we assumed to be headings.

> @fcollonval incredibly minor typo, but it's `starlette.requests`, not `starlette.request`. Thanks for this fix. Good catch I updated the snippet.

I'll work on a fix for this this coming week.

The pattern we used for the "copy shareable link" command in the file browser's context menu was to create a small targeted plugin that adds just that one command (this...

> What if the `parseHeading` function followed the same principle as `getHeadingId`, this is rendering the heading using a proper markdown parser and sanitizer to a temporary HTML element (instead...

In JupyterLab 4, it will be possible to modify the positioning via settings:  > Keeping this PR opened as adding user interactions will make it easier to update the...

i don't know if this was the reason. But I saw that version 0.50.1 failed to be published on conda-forge. It is done now. Does that solve your issue?

If I set a password (the same but it does not matter) for notebook and jupyter-server, here is the server log: ``` [I 12:23:07.559 NotebookApp] 302 POST /login?next=%2Flab (127.0.0.1) 21.400000ms...

As far as I know, yes. The easiest setup to make binder run with Jupyter Server rather than the old Notebook server is to add a `start` file in the...