Did you actually want my review? Because this surely could be missunderstood.

Sorry, I should have waited longer to submit this. If you have concerns we can always roll back or change it again. What do you think are the issues with "offset"?

Offsets usually are 0 based, while the movement (prev/next) is not. It also doesn't make sense if you look at the RevsetExpression::ancestors_at/descendants_at documentation.

The movement is zero-based relative something, isn't it? I think what it's relative to depends on where @ currently is and whether --edit is used.

I looked at the revset docs and I don't think "offset" is incompatible. Maybe I'm missing something?

    /// Ancestors of `self`, including `self` until `generation` back.
    pub fn ancestors_at(self: &Rc<RevsetExpression>, generation: u64) -> Rc<RevsetExpression> {
        self.ancestors_range(generation..(generation + 1))
    }

So an offset of 1 is one generation back from the current generation. And an offset of 0 is the current generation. Right?

While the name is fine, a offset of one implies the behavior you describe as a bug, so it does not align with the default argument.

I have been successfully convinced that the current behavior is not a bug and that I had a misunderstanding about what the movement is relative to (it appears to be relative to @-). Maybe the new name is ok then?

Maybe the new name is ok then?

It is ok.