New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
cell execution asterisk does not change to cell execution number #13383
Comments
The JavaScript error
Was also happening in JupyterLab v3.0.16 which doesn’t have this bug, so I don’t think that’s the problem. |
Possibly related? #5366 |
I checked more images for the presence of this bug. Bug is not present: Bug is present: |
Unfortunately subject to bug jupyterlab/jupyterlab#13383
So there it is, this bug was introduced somewhere in JupyterLab v3.2.5 https://github.com/jupyterlab/jupyterlab/blob/master/CHANGELOG.md#325 |
Here are some versions of packages installed in the JupyterLab v3.2.4 image:
And here's what's installed in the JupyterLab v3.2.5 image:
|
I've published some IHaskell Docker images for anyone who wants to try to debug this issue. These two images are the same except that the base image has a different version of JupyterLab. JupyterLab v3.2.4 image:
JupyterLab v3.2.5 image:
|
Triage notes: @krassowski reports the same issue in the R kernel, particularly if there is an error. |
@jamesdbrock could you install jupyterlab-kernelspy and display the kernel reply when executing a cell? |
Thanks for the advice @fcollonval , I'll try that. |
First ExecutionSecond ExecutionSixth ExecutionWhen I clicked Execute for the sixth time, the kernelspy reported receipt of a |
@jamesdbrock thanks for all the information - I was looking at this a bit. I tried your docker and can confirm your issue (thanks for that). I tried also your binder at https://mybinder.org/v2/gh/gibiansky/IHaskell/mybinder and the issue does not appear for me (JupyterLab is 3.2.9 in that binder). Did you implement a workaround that is deployed on the Binder or not? |
Trying to dive into this one a bit more, I added jupyterlab-kernelspy in the docker image (JLab 3.2.5) and compare what we get for the haskell kernel (above) and the Python kernel (below): ... no error this time - you can actually see that I get I uninstalled jupyterlab-kernelspy and the error was back. Debugging the Javascript, I see that the |
Thanks for the investigation @fcollonval . I didn't know that https://github.com/IHaskell/IHaskell/blob/master/Dockerfile is working with JupyterLab 3.2.9 (and it does work, I confirmed.) I've been using https://github.com/IHaskell/ihaskell-notebook/blob/master/Dockerfile for this issue. |
I did not try to update the JupyterLab version within the local docker I used. But my guess is that as Binder is a distant server the network messaging is slower and we fall on the good side of the race condition. I'm confident the error is coming because that promise is not resolved at the expected time:
But I didn't figure out yet which flag combination is not properly set (there is something about the need to have got a reply and being in idle state). |
|
Another note: when I build and run JuptyerLab with the IHaskell kernel using the latest https://github.com/tweag/jupyterWith , I get JupyterLab v3.4.3 and the bug occurs. So the bug isn't caused only by the https://github.com/IHaskell/ihaskell-notebook/blob/master/Dockerfile . This bug is a puzzle. |
This bug is still present in JupyterLab v3.6.1 |
This bug is still present in JupyterLab v4.0.2 |
Description
When executing cells with the IHaskell kernel, the cell execution asterisk does not change to a cell execution number after execution has finished.
However, it will change after some number of other cells have been executed.
In this demo GIF, the cell execution asterisk will change after two more cells have completed execution. (I have also observed sometimes that the cells change after five more cells have completed execution, or some other number).
Note that the execution result is displayed immediately — it’s only the execution asterisk which doesn’t update until two more cells have executed.
Reproduce
You can observe the bug in this docker image from ihaskell-notebook, which is based on https://hub.docker.com/layers/jupyter/base-notebook/lab-3.5.0/images/sha256-cc9e4b7e76519ef2c294b1b15188a012da4aad3b8b8a4b09c366e0c83a90b45f
Context
This bug was not present in JupyterLab v3.0.16, but the bug has appeared in every version which I’ve checked after that version:
Maybe this bug caused by some strange behavior of the IHaskell kernel? If so, then that strange behavior did not manifest as this bug until something changed in JupyterLab after JupyterLab v3.0.16. And the execution cells for IHaskell work fine in VS Code notebooks.
I would welcome any hints about how I should investigate this bug further.
The text was updated successfully, but these errors were encountered: