same issue, I write a script to clear cell outputs before committing for security, but the command

jupyter nbconvert --ClearOutputPreprocessor.enabled=True --inplace

changes the line ending

If we change this behavior we should leave a flag to preserve system line-endings. I'd be ok with someone making that change so long as it could be opt-ed out of. There's sometimes system specific actions that complain when line-endings don't match the system but most modern windows applications and tools handle it smoothly now-a-days.

Maybe a option flag to let user select line ending among

• keep-as-it-is (the same to the source file)
• auto (depends on the OS running the command)
• LF
• CRLF

Agreed that the default value of such flag would preserve the current behavior (now is auto?)

Yes. I believe it is auto atm. I'd also be ok changing the default to LF in the 6.x branch if we wanted as conforming to other notebook executions might be nice and the major change is easy to justify default changes.

The default should be the same as the Jupyter notebook or lab editors. It's a mistake that it's not the same and I can't imagine how anyone would be happy that the line endings change. I'm happy for it to be an option, but the default should be the one the creates the least issues. If no-one has raised this as an issue before, then it's probably because they're not looking.

Sorry to be a bit :grrr: but LF line endings on Windows hasn't been an issue for years, and especially not for anything or anyone that would consume an ipynb file.

I added it to the 6.0 milestone -- could use some extra contributors helping get small issues like these moved across the finish line before the last major issues get resolved.

Hi, any update on this? I'm running nbconvert 6.0.0 and still experiencing the issue. Is the fix included in a more recent version?