Upgrade extension for JLab 4 #263
Conversation
mahendrapaipuri
changed the title
There was a problem hiding this comment.
Thanks you so much for raising this.
Did you follow the migration guide to generate this? I've run the tool myself and it seems to have made many more changes than are included in this PR. However I also can't get things to install correctly after running the tool.
|
@jacobtomlinson Not really. I just bumped the node deps manually to If you want, I can extend this PR to hatch migration. We have already done it for our custom extensions and we have a very positive experience. |
|
If you're happy to take that on that would be fantastic. |
|
@jacobtomlinson Please check the updated PR description! |
|
@mahendrapaipuri thank you so much for this, and apologies for the slow review! I'm prioritizing this to get it over the line, and will have some more detailed comments this evening. We just published a patch release targeting JupyterLab 3, and there are a few conflicts to resolve. I'm happy to do that unless you get to it faster. I'm very excited to see versioneer and the |
|
@ian-r-rose You are welcome. I dont see any huge problems due to conflicts that are raised due to #267. Most of the conflicting files ( I will wait for your comments before I resolve the conflicts. |
There was a problem hiding this comment.
Thanks @mahendrapaipuri. I'm still digesting the hatch-based changes, but had a few questions/comments
There was a problem hiding this comment.
Overall, this looks really great, my comments and questions are all pretty minor. Thank you @mahendrapaipuri!
Update editor model in injectCode function Upgrade to jlab 4, lumino 2 and yarn 3 Use jl>3 as build dep to support jl 3 and 4 Include .yarnrc in VCS and add .yarn in gitignore
Test on jl 3 and 4 and py 3.8 and 3.11 Should cover most of the cases
Update classifiers and min pyver to 3.8
Remove deprecated serverextension subcommand from build tests
Migrate from versioneer to nodejs versioning. The package will be versioned based on version in package.json. More details https://hatch.pypa.io/latest/version/ Removed obselete versioneer files and references Removed contents of setup.py and unnecessary file MANIFEST.in setup.cfg is retained as pyproject does not support flake8 config Added black and pytest config to pyproject.toml
Seems like flake8 does not support inline comments anymore
Get rid of setup.cfg which is obsolete for the package
Fix minor bugs in pyproject config
mahendrapaipuri
force-pushed
the
bump_node_deps
branch
from
a5bd7e6 to
cda85b0
Compare
Add instructions to JLab 4 compatible installation Update instructions for release using hatch
|
@ian-r-rose Addressed your PR comments, rebased onto new main and updated README too. |
There was a problem hiding this comment.
Thanks @mahendrapaipuri! One small suggestion, which I could do in a follow-up if you feel like wrapping this one up.
Seems like this is JLab 4 effect as well.
@jupyterlabto4.0.0and@luminoto2.0.0.index.tsare made to work with JLab 4versioneertonodejsversioning. Python package will take the version string frompackage.json. Bumping of versions can be done usinghatch versioncommand.blackandpytestconfig topyproject.toml3.8inline with JLab 43.8and3.11in CIMaintainers - If you prefer to have
versioneer, let me know. I will add it back. The thing isversioneeris not very well integrated withpyproject.toml. There is hatch-vcs plugin that is made for hatch build system that does exactly same thing as versioneer but with less config burden.