Jupyter4Science / jupyterlab-ide

JupyterLab + tab completion + debugger + linter = Fully-fleged IDE
MIT License
3 stars 0 forks source link

Difficulties installing JupyterLab-Vim #55

Closed Dcosthephalump closed 8 months ago

Dcosthephalump commented 1 year ago

I was going to start the next notebook with a blank slate environment to make sure that I had everything working. I removed the jupyterlab-ide environment. When I attempted this, I started to get this error repeatedly:

InvalidArchiveError('Error with archive C:\\Users\\David\\anaconda3\\pkgs\\jupyterlab-vim-v0.8.0-0.tar.bz2.  You probably need to delete and re-download or re-create this file.  Message was:\n\nfailed with error: seeking backwards is not allowed')

I had a similar issue here.

I tried uninstalling and reinstalling Anaconda. This worked last time I had a similar issue. It did not work this time.

I have temporarily solved this problem by locally removing the jupyterlab-vim package from the environment.yml and installing it after building the environment. The problem does not exist for the Binder, although Binder does not seem to allow for jupyterlab-vim to be used when it creates an environment.

Open to solutions. Maybe it's a Windows thing?

nicole-brewer commented 1 year ago

1) Maybe we look up what version is installed when we install "after-the-fact". Then specify that version that worked.

2) Can you add

dependencies:
- conda
  - jupyterlab-vim

to force conda to run that command separately and after the fact? Hopefully?

3) If that doesn't work, put the "conda install... " line in a postBuild script. IF this is the route we take, we need to update REAME.md with further instructions

Dcosthephalump commented 1 year ago
  1. This did not work. I got a package not found error when I tried to install with the specific package number that worked after the fact.

  2. This returned the same error as solution 1.

  3. For code formatter and for vim, I followed the conda-forge instructions after installation and got the desired results.

  4. I did try making jupyterlab-vim and jupyterlab-code-formatter pip installed dependencies. This worked without problems.

I see two options:

  1. Use pip for installing code formatter and vim
  2. Go with the postBuild script.

I prefer pip as it involves only altering environment.yml instead of doing more complicated things.

nicole-brewer commented 1 year ago

Sounds good!