Marcin Wrochna

Thanks, I now realize it might be more complicated than just copying JUPYTERHUB_API_TOKEN. It's true that at the notebook level [the issue](https://github.com/jupyter/notebook/issues/1000) of getting a notebook's path is still open...

Sorry, this requires some understanding of all the requirements in lightning projects and of python packaging in general – I have neither, so I'll just say what I found by...

I believe `pytest_collection_modifyitems` should me marked with `@pytest.hookimpl(trylast=True)`, otherwise it may happen before items are deselected with the default built-in pytest key/mark selections (it happened to me).

I was very surprised than selecting a test doesn't also select its dependencies and that there is no option or at least documentation on how to change that. I had...

For the download of public databases to work, I needed to login to ghcr.io with `docker login ghcr.io -u $MYUSERNAME --password-stdin` and a [Personal Access Token](https://github.com/settings/tokens) with just `public_repo` access....

You're right, `docker logout ghcr.io` works, and is described in FAQ, sorry about that. (I actually probably had expired PATs for ghcr.io stored in configuration; can't verify now, but I...

@MatoFD Can you try if the following workaround works? Take CUDA packages from the nvidia channel instead of conda-forge, that is, replace `cudatoolkit` with `nvidia::cudatoolkit`, and replace `nvcc` with `nvidia::cuda-nvcc`...

@aklink Could you show your `conda list` and `conda info` in the failing container (and if possible in the working container image as well)?

The issue is that arguments are not properly [url-encoded](https://en.wikipedia.org/wiki/Percent-encoding) when sent via HTTP, so '%' followed by digits will be interpreted as a random character. One place where this needs...

Thanks!