Currently we set allow_errors: True in our jupyter book _config.yml. This means that if a cell in a notebook throws an error, the build continues and reports success.
So our "health checks" as shown in the nightly build badges aren't really measuring whether the notebooks run.
I wonder if we should toggle the allow_errors setting just for the nightly build, while keeping it for other builds. (At least during the development phase it's very useful to be able to see the rendered book with the errors included)
Currently we set
allow_errors: Truein our jupyter book_config.yml. This means that if a cell in a notebook throws an error, the build continues and reports success.So our "health checks" as shown in the nightly build badges aren't really measuring whether the notebooks run.
I wonder if we should toggle the
allow_errorssetting just for the nightly build, while keeping it for other builds. (At least during the development phase it's very useful to be able to see the rendered book with the errors included)