bug fix: timestamp changed but content did not change

[ctappy]

References

Fixes #7743
Fixes #11352

Code changes

fixes timestamp changes but the content did not change. An issue with fuse filesystems, timing issue #7743, multiple users opening the same file #11352, and file touched but not edited issues.

User-facing changes

N/A

Backwards-incompatible changes

N/A

[jupyterlab-probot]

Thanks for making a pull request to jupyterlab!
To try out this branch on binder, follow this link:

[welcome]

Thanks for submitting your first pull request! You are awesome! 🤗

If you haven't done so already, check out Jupyter's Code of Conduct. Also, please make sure you followed the pull request template, as this will help us review your contribution more quickly.

You can meet the other Jovyans by joining our Discourse forum. There is also a intro thread there where you can stop by and say Hi! 👋

Welcome to the Jupyter community! 🎉

[ctappy]

I tested these changes with normal files and ipynb files

• editing the file in the filesystem and saving it in JupyterLab
• touching the file and saving it in JupyterLab

[krassowski]

To fix the failing CI jobs please run integrity script and commit:

Please commit the changes by running:
git commit -a -m "Package integrity updates"

[krassowski]

@jasongrout I see that you commented on both referenced issues, does this solution sound reasonable to you?

[ctappy]

updated code to check object recursively and sort, then convert to a string for 1:1 comparison

[fcollonval]

I definitely would like to avoid pulling the content as comparing large file will take time and increase the wire load.

Is there a way to set better this.contentsModel?.last_modified?

[ctappy]

• we could add a comparison to files under a specific size and modification time difference

[krassowski]

Would it be possible to compute hashsum (e.g. md5) of the content on the server and on the frontend and compare the hashsums instead of comparing the contents directly?

[ctappy]

@fcollonval The problem with last_modified with fuse filesystems is not all of them handle the change/modified time correctly. If someone has their home directory mounted on a fuse filesystem this issue will constantly pop up, even though the file didn't change

[ctappy]

when the message comes up "File Changed" it is not technically correct, it should be "Timestamp Changed" because that is what the check is looking at, this would run a file change check

[ctappy]

@krassowski I can't seem to find where the API call is going to where the file is opened. I'd love to add in a md5sum check if you can point me in the right direction. Thanks!

[krassowski]

That would be jupyter-server file contents API see https://github.com/jupyter-server/jupyter_server/blob/main/jupyter_server/files/ and https://jupyter-server.readthedocs.io/en/latest/developers/contents.html. It would probably be good to start by opening an issue on jupyter-server repository to discuss

FYI we ship a murmur2 function in the debugger package, so maybe we could reuse that for client-side check (if needed at all).

[Wh1isper]

Or we can compare the contents of the file on the server side and not manipulate the file if there is no change so that Timestamp doesn't change either.

[Wh1isper]

I'm working on adding the md5 to the jupyter-server file contents API: jupyter-server/jupyter_server#1363

[Wh1isper]

jupyter-server/jupyter_server#1363 has been merged, JupyterLab can now(next release actually) use md5 for file content verification!

cc @ctappy @fcollonval @krassowski

UPDATE:

md5 has been changed to hash and hash_algorithm in jupyter-server/jupyter_server#1367

Thank you all.

[krassowski]

@ctappy or @Wh1isper would you be interested to rebase this PR including the new jupyter-server feature?

[Wh1isper]

@krassowski Thank you for the invitation. I’m interested but currently occupied with other commitments. Also, I’m relatively new to TypeScript and frontend development.

I may revisit this when I have some spare time! I’ll discuss this with a colleague who initially informed me about this PR. They might be interested in contributing.

[Wh1isper]

Guys, I just created a PR that uses hash and hash_algorithm for file modification comparison: #15577

Forgive me for being a TypeScript and front-end novice, more advice would be appreciated, thanks!

cc @ctappy @krassowski @fcollonval

[krassowski]

Closing in favour of #15577. Thank you all!