Draft: revive Wingman

[santiweight]

It's looking pretty good. Still working on it. I expect to have a final version working this weekend, depending on whether I feel like working on this or the extract method code action...

Tests are at 14/77. ~10 of these failures are non-serious (such as fresh variables being slightly different). 4 of these failures are real, and I haven't investigated. I suspect I will be ignoring them for the sake of this MR

[michaelpj]

Thanks for taking this on! Was there anyone else keen on maintaining wingman who we could pull in for review?

[michaelpj]

I suggest we don't do this - we do publish these to hackage, so renaming them is just a bit unnecessarily annoying.

[santiweight]

So I think it's mostly up to Sandy whether he wants to review. I was hoping the fact that I marked this draft wouldn't notify anyone, but I'm getting the feeling marking draft doesn't do much...

So we had some discussion here: #2971

I think the next steps are:

1. take Sandy's simplification changes (the big red diff), but undo the renaming, which we don't want
2. merge (1)
3. rebase my changes atop
4. merge my changes in

For the transition I'm going to take the approach outlined by Michael here: #2971 (comment)

The old codebase will be frozen for <9.2, and >9.2 will have a vastly simplified codebase that lives in parallel

[michaelpj]

I was hoping the fact that I marked this draft wouldn't notify anyone

I'm just nosy :D and wondering if we should get anyone else involved.

[ChrisPenner]

Very excited this is being added, thanks for taking it on.

Mostly I use the pattern match expansion action 🙌🏼

[michaelpj]

I'm closing this, but we should remember that it exists for reference if someone wants to come back to this.

[bzm3r]

@michaelpj I'd like to take up maintenance of this.

[michaelpj]

@bzm3r do you mean you want to pick up the work? I think there's quite a lot to do, not just maintenance 😅 You might be able to discuss with @santiweight what needs doing

[bzm3r]

@michaelpj Yeah, I did read the following two messages in particular:

1. Simpler Tier 1 tactics for case splits #3525 (comment)

2. Simpler Tier 1 tactics for case splits #3525 (comment)

("Maintenance" is no longer the right word; I initially thought it was just a matter of putting-in-the-hours, but did not appreciate the full scope of the problem.)

I have some thoughts on how to handle this, but better than thoughts, will be code. So, that's why I have been somewhat silent since I wrote "I'd like to take up maintanence of this": putting together code takes a lot of time for me.

[artimath]

@bzm3r are you still working on this?

I am rather keen to dive in face-first, and rather sure that I have very little understanding of the scope of work entailed.

[bzm3r]

@artimath No, I have stopped working entirely with Haskell. 'Needs must' and all that.