👈 Launch a Binder on branchblink1073/notebook/use-hatch

We're blocked bypypa/hatch#243.

Found a workaround usingforce-include

FYI just releasedhttps://github.com/pypa/hatch/releases/tag/hatchling-v1.0.0

Also I think you can copy what you did hereipython/ipyparallel@e3e896a

Thanks, and congratulations!

We can removesetup.py oncejupyter-server/jupyter_releaser#319 is released, then this should be good to go!

It's undocumented until I find an auto-generator for non-Click CLIs, but if build deps are met you can dohatchling version

if build deps are met you can do hatchling version

Thanks! Perhapshttps://sphinx-argparse.readthedocs.io/en/stable/?

Perhapssphinx-argparse.readthedocs.io/en/stable?

Ah, you are usingmkdocs...

@jtpio, any idea why the snapshot tests are failing?

It looks like the menu and toolbar customization files are missing (defined inschema) for one of the failures:

It looks like the menu and toolbar customization files are missing (defined in schema) for one of the failures

Thanks!

Hmm, after adding the schemas the tests don't appear to start at all.

Hmm, after adding the schemas the tests don't appear to start at all.

hmm strange. For the UI tests the package is first installed with the following:

python -m pip install -vv notebook*.whl

Will check locally.

Nice, CI is now passing withfcdb15f.

Looks like there are a few conflicts after the merge of#6429.

Ah I thought CI was already uploading the built assets as artifacts.

Maybe we could add the following step to the check release workflow so it's easier to inspect the sdist and wheel?

-name:Upload Distributionsuses:actions/upload-artifact@v2with:name:notebook-jupyter-releaser-dist-${{ github.run_number }}path:.jupyter_releaser_checkout/dist

Good call. I also played with my wip comparisonscript and got this output now:

python ../extension-cookiecutter-ts/compare.py ../notebook_orig . sdist

Removed_files:notebook-7.0.0a4/jupyter-config/jupyter_server_config.dnotebook-7.0.0a4/jupyter-config/jupyter_server_config.d/notebook.jsonnotebook-7.0.0a4/notebook/labextension/static/remoteEntry.4fb60df0e7df1388ab11.jsAdded files:notebook-7.0.0a4/notebook/labextension/static/remoteEntry.5ea5fe7dc5a5fbc91231.jsnotebook-7.0.0a4/jupyter_config.jsonnotebook-7.0.0a4/lerna.jsonnotebook-7.0.0a4/pyrightconfig.jsonnotebook-7.0.0a4/.eslintrc.json

Here's the latest output:

Successfully built notebook-7.0.0a4.tar.gzRemoved_files:notebook-7.0.0a4/jupyter-config/jupyter_server_config.dnotebook-7.0.0a4/notebook/labextension/static/remoteEntry.4fb60df0e7df1388ab11.jsAdded files:notebook-7.0.0a4/notebook/labextension/static/remoteEntry.5ea5fe7dc5a5fbc91231.js

Successfully built notebook-7.0.0a4-py3-none-any.whlRemoved_files:notebook/labextension/static/remoteEntry.4fb60df0e7df1388ab11.jsnotebook-7.0.0a4.data/data/share/jupyter/labextensions/@jupyter-notebook/lab-extension/static/remoteEntry.4fb60df0e7df1388ab11.jsnotebook-7.0.0a4.dist-info/top_level.txtAdded files:notebook-7.0.0a4.data/data/share/jupyter/labextensions/@jupyter-notebook/lab-extension/static/remoteEntry.5ea5fe7dc5a5fbc91231.jsnotebook/labextension/static/remoteEntry.5ea5fe7dc5a5fbc91231.js

Looking good!

The Binder for this PR fails to build because of the following error:

Which seems to be related to thejupyter labextension develop command expecting to find asetup.py:

https://github.com/jupyterlab/jupyterlab/blob/6099bcb4d32902bd61b82df818cdd2df3fe1685c/jupyterlab/federated_labextensions.py#L427-L438

Which seems to be related to thejupyter labextension develop command expecting to find asetup.py:

This is likely to be fixed byjupyterlab/jupyterlab#12606.

Yeah I think we need to use asetup.py shim for extensions that are targeting versions ofjupyterlab that require it. Once we have 3.x and 4.x versions that don't require it, the cookiecutter itself can be updated to not include it. For the migration script, we can include it if there was already one there and eventually remove it.

Drat, now we have the issue that symlinks can't be created over existing files. Withsetuptools the shared data was not written in editable mode. It feels like we need to fix a few bugs and backport them in JupyterLab before doing any migrations. We shouldn'tneed asetup.py and we shouldn't error out when overwriting symlinks.

Hatchling resolves symlinks fyi

The issue is the JupyterLab build command is trying to create symlinks over the files that are placed into/share and causing an error. This issue is purely of our own making. 😄

Ah, the symlink error was actually from a file in this repository, the binder is working now. I actually think this PR is good to go now.

Starting a new pre-release with this change.

Ah looks like something is wrong with the Python version:#6448