Skip to content
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

In windowed notebook cells are sometimes rendered out of order #15610

Closed
krassowski opened this issue Jan 6, 2024 · 9 comments
Closed

In windowed notebook cells are sometimes rendered out of order #15610

krassowski opened this issue Jan 6, 2024 · 9 comments
Labels
bug tag:Regression Behavior that had been broken, was fixed, and is broken again
Milestone

Comments

@krassowski
Copy link
Member

Description

In windowed notebook cells are sometimes rendered out of order. The order of navigation is thus broken in cells, affects both edge navigation and command mode navigation with arrows. Dragging the affected cells does not change their position.

It looks like a problem with the soft hiding of cells; it looks like a soft-hidden cell stuck around, but maybe this is just a symptom of another problem. Cutting and pasting the affected cell solves the problem.

It is likely related to #15286

Reproduce

I don't know yet but it only happens in long notebooks with full windowing mode enabled. It likely involves scrolling the cell in/out view and possibly some edge cases on modifications

Expected behavior

Cells are rendered in correct order.

Context

  • Browser and version: Chrome
  • JupyterLab version: 4.1.0b0
@krassowski krassowski added bug status:Needs Triage Applied to new issues that need triage tag:Regression Behavior that had been broken, was fixed, and is broken again labels Jan 6, 2024
@krassowski krassowski added this to the 4.1.0 milestone Jan 6, 2024
@krassowski krassowski added the tag:Release Blocker A must-have bug for the milestone to which it is tagged label Jan 8, 2024
@JasonWeill JasonWeill removed the status:Needs Triage Applied to new issues that need triage label Jan 9, 2024
@JasonWeill
Copy link
Contributor

@krassowski Just to be sure, do you see this behavior in 4.0.x?

@krassowski
Copy link
Member Author

No, it is 4.1.0b0 specific. As mentioned, likely related to #15286.

@DenisaCG
Copy link
Member

I'm currently looking into this issue and will keep you posted on the progress. Feel free to jump in if you have any insights or suggestions. Thanks!

@DenisaCG
Copy link
Member

I'm having some trouble reproducing the issue. So far I've tired running notebooks of around 200-300 cells and navigating the cells in both edge or command mode using the arrows, but the rendering order of the cells doesn't seem to change. I've also tried selecting cells or their outputs and moving them in and out of view, same results. In all cases, I have change the settings to full windowing mode, and I am using version 4.1.0b0.

Is there anything more you have done or noticed when encountering this issue? @krassowski

@krassowski
Copy link
Member Author

I will try to reproduce again tomorrow morning. I suspect a combination of cutting/pasting cells or maybe changing their types when the active cell is out of view could be the key.

@krassowski
Copy link
Member Author

I put quite some effort into reproducing it but I could not find any sequence to reproduce it in a clean environment (but I had seen it a few times since reporting it initially in my work environment which has a number of extensions and often complex workspaces making my computer slow down significantly).

@krassowski
Copy link
Member Author

This is how the DOM looks like when it happens:

Screenshot from 2023-12-25 21-30-43

Notice two code cells with 262 index with a markdown cell with index 261 in between

@krassowski krassowski self-assigned this Jan 17, 2024
@krassowski krassowski removed the tag:Release Blocker A must-have bug for the milestone to which it is tagged label Jan 17, 2024
@krassowski
Copy link
Member Author

We discussed this on the weekly JuypterLab call today: we will still attempt to fix it before the final release, but it is not deemed blocking because it happens in the windowing mode which is not used by default; instead we will add a recommendation in the release notes for users who use the full windowing mode to either postpone upgrading to 4.1 or help with testing that mode in the new version (thanks @jtpio for the suggestion!)

@JasonWeill
Copy link
Contributor

Closing per #16013.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug tag:Regression Behavior that had been broken, was fixed, and is broken again
Projects
None yet
Development

No branches or pull requests

3 participants