feat: interactive rebase should include current commit

[star-szr]

As discussed in #862

I really tried to write a test for this, but couldn't figure out how to wait until the git-rebase-todo was present to check its contents. Maybe later :)

Feedback very welcomed, as always.

[CKolkey]

Is something like this useful? https://stackoverflow.com/questions/2221658/whats-the-difference-between-head-and-head-in-git

I was imagining (though maybe I'm wrong) that using <commit>^ would work here.

[star-szr]

Thanks for taking a look.

If we just change the commit parameter to commit^, it will produce an error if the user selects the initial commit in a repository, since that commit has no parent. But it is possible to rebase the initial commit by passing --root.

Since we need to figure out the parent to handle the case of the initial commit, my thinking was that it may be worthwhile to just extract it from that output via vim.split rather than use something like commit^.

[CKolkey]

Ah, ok, that makes sense. Thanks for the explanation :)

[star-szr]

Ah, ok, that makes sense. Thanks for the explanation :)

Of course! It could be worth adding a comment in the code, let me know if you think so and I can submit something.

[CKolkey]

Maybe as a commit message, or message here, but inline can get... overwhelming.

[star-szr]

Good point, well I think there's enough detail here so I will leave it. Next time I work on something like this I can add more details to the commit message. Thanks!

By the way, it looks like tests failed on the merge commit, but as far as I can tell it might just be a random fail, possibly a race condition? I found similar failures within the last few weeks though I didn't dig further back than that.

[star-szr]

I can't reproduce any failure locally so far.

I ran the status_spec test that failed in the merge commit over 140 times with no failure and I ran the whole suite 70 times with no fail.

[star-szr]

Breaking news: I let it run (only the status_spec test) while I was in a meeting and it failed on run 895 in the same way.

[star-szr]

To not keep cluttering up this PR, I created #887 to track the random test fail issue now that it has been confirmed.

[CKolkey]

Breaking news: I let it run (only the status_spec test) while I was in a meeting and it failed on run 895 in the same way.

As they say, fool me 894 times, shame on you. Fool me 895 times, shame on me.

The state-with-side-effects nature of testing a local git repo is... rough sometimes.