Closed schmittst closed 3 years ago
Thanks for reporting it. I will check on the weekend, but I guess it might be a compatibility issue for the newer version. It works great for me with 0.3.4 (packageVersion('languageserver')
); if you need it ASAP you could try downgrading the server (or maybe the kernel?).
I quickly tried to downgrade the languageserver package to 0.3.4 like you suggested - did not help. So I guess I would need to downgrade the kernel. Not that urgent - I will wait until it is compatible! Thanks for the great work you have done with jupyterlab-lsp!
If you use dockerfile, try this.. you should install JupyterLab 2.1.2 that following this script. make sure follow the steps below.
RUN pip install jupyter-lsp && \ conda install -c conda-forge nodejs && \ jupyter labextension install @krassowski/jupyterlab-lsp --no-build && \ conda install -c conda-forge python-language-server
Thanks. I am using the jupyter/datascience-notebook (3b1f4f5) docker image. And I already used those commands in the dockerfile to setup the lsp (w/o the --no-build flag) However, I realized that this is running on JupyterLab 2.1.3 So it might be just incompatibility with 2.1.3?
Unlikely - it works for me with:
but then, I do not use docker...
Does the R language server return any suspicious messages if run from an R console with languageserver::run(debug = TRUE)
?
@schmittst we had a major release recently aimed at JupyterLab 2.2. Would you like to try it out and check if this is still an issue for you?
I will close it for now as it seems to be working fine with R for other users (and myself), but please feel welcome to re-open or create a new issue if you need any assistance.
Description
I have an issue with jupyterlab-lsp. I am new to JupyterLab and use it within a Docker (jupyter/datascience-notebook:latest) and within a Galaxy installation (but I have the same issue without Galaxy). I installed jupyterlab-lsp as described in the manual (incl. R and Python LSP) and the language server for Python (pyls) initializes correctly and and jupyterlab-lsp works as expected. However, when I switch to the R kernel the jupyterlab-lsp stops as "connecting..." for a while and then just says "connected" for the r-languageserver (instead of "initialized" as it does for pyls). As one would expect, jupyterlab-lsp then does not work for the R notebook (i.e. does not display any suggestions).
Reproduce
Expected behavior
For 4. and 5. from above I would expect the language server switching to "initialized" for both notebooks. This works for 4. (and e.g. autocomplete works) but not for 5. (here jupyterlab-lsp only shows "connected"). Expecially the output of the Browser JS concsole (see below) i find strange. Hee I get an "TypeError: can't convert undefined to object manager.js:" when opeing an R notebook. When opening an Python Notebook this does not happen.
Any ideas on how to solve this?
Context
Required: installed server extensions
Required: installed lab extensions
Troubleshoot Output
Command Line Output
Browser Output