The "registry" script in those instructions starts Verdaccio.
We'd want to have something similar, with the idea that you should be able to build jupyterlab without too much fuss. We can't escape fuss, though.
We need to surface what Verdaccio changes need to be made to publish and unpublish packages. Additionally, we should improve Verdaccio's reliability. On my local machine I've often needed to manually verify health after pushing packages, which kills cycles. Being able to trust the script to do it will improve developer burden and eliminate variables in local build failures.
There's some prior art for this in the Data Registry extension:
https://github.com/jupyterlab/jupyterlab-data-explorer/blob/master/docs/development.md#installation
The "registry" script in those instructions starts Verdaccio.
We'd want to have something similar, with the idea that you should be able to build jupyterlab without too much fuss. We can't escape fuss, though.
We need to surface what Verdaccio changes need to be made to publish and unpublish packages. Additionally, we should improve Verdaccio's reliability. On my local machine I've often needed to manually verify health after pushing packages, which kills cycles. Being able to trust the script to do it will improve developer burden and eliminate variables in local build failures.