Why does `gh stack rebase --upstack` include the current branch?

Say I have the following stack

```
main <- branch1 <- branch2 <- branch3
```

I'm _on_ `branch1`, and just addressed some reviewer feedback by adding a few extra commits. Now I need to do the following:

- rebase `branch2` on `branch1`
- rebase `branch3` on `branch2`

I do _not_ want to rebase `branch1` on `main` right now.

My instinct is to call `gh stack rebase --upstack` while on `branch1`, where I just addressed the feedback and added commits. I was expecting that to do what I outlined above and not change `branch1` in any way.

Surprisingly (to me), it did this:
- rebase `branch1` on `main`
- rebase `branch2` on `branch1`
- rebase `branch3` on `branch2`

Can I ask about why `--upstack` includes a rebase of the current branch on the branch under it? That actually feels a bit "down stack"-ish to me.

It feels like this exact workflow is going to be pretty common for me, and the current workaround is for me to `gh stack switch` up to `branch2` and _then_ run `gh rebase --upstack`, but that feels pretty unnatural to me!

入口点是 `gh stack rebase --upstack`；首先在 issue 中描述的 stack 上重现该问题，并检查命令如何选择分支。将观察到的顺序与所要求的顺序进行比较，然后检查该命令现有的测试（如果有）。当 rebase `branch1` 后它保持不变，而 `branch2` 和 `branch3` 被 rebase 到各自更新后的前置分支上时，即表示完成。

由索引模型根据 Issue 内容生成。