Integrity check seems buggy at least on classic notebook app

[falexwolf]

This is feedback from @paulgb, with a case that we should be able to reproduce:

[Koncopd]

Thx, i will check.

[Koncopd]

Now the rule is that the cells are considered run consequently when they start at 1. I can change that ofc, but not sure if i should. I agree that this specific example does't look like it has integrity problem, but the first cell run multiple times can have integrity problems in general.

[falexwolf]

You're right about this, and I also think that starting_at_1 should remain a hard requirement for integrity == 1 to be true.

The problem here is the logging output:

It calls out cell transitions as violations that don't seem to be present in the notebook. And it doesn't state that the violation is "Your notebook does not start with cell number 1".

So, there are two things

1. there might be a bug in computing violations nonetheless
2. we should have a more specific logging output pertaining to violations of starting_at_1

What do you think, @Koncopd?

[Koncopd]

I am pretty sure this happens because the notebook was not saved before publishing.

[falexwolf]

Could you update the documentation correspondingly, @Koncopd? To state that cells need to start with 1?