Simplify List.folds with Dict/Set/List insert to fromList

[caseyWebb]

Replaces instances of

List.foldl Set.insert Set.empty items
    -> Set.fromList items

List.foldl (\(k, v) -> Dict.insert) Dict.empty items
    -> Dict.fromList items

[jfmengels]

Hi @caseyWebb 👋
Thank you for making a PR ☺️

List.foldl Set.insert Set.empty items
--> Set.fromList items

This one sounds fine to me 👍

It would not replace a version with a lambda though which I guess would be nice, but that could be a future improvement.

List.foldl (\a b -> Set.insert a b) Set.empty items

Don't forget to add tests for this one.

---

List.foldl Dict.insert Dict.empty items
--> Dict.fromList items

This one doesn't work, or rather, the original line doesn't compile, so we should not try to simplify it
List.foldl (\(k, v) acc -> Dict.insert k v acc) Dict.empty [] would be what you were looking for. Which I guess we could simplify to Dict.fromList, but that's more work.

---

List.foldl (\x acc -> if condition x then Set.insert x acc else acc) Set.empty items
 --> items |> List.filterMap condition |> Set.fromList

List.foldl (\x acc -> if condition x then Dict.insert x acc else acc) Dict.empty items
 --> items |> List.filterMap condition |> Dict.fromList

These ones I don't think would be a good fit, because they're much slower. I tend to rewrite my code from the latter to the former when I need to crank out performance. Also, I think it should by |> List.filter condition?

---

In the future, don't hesitate to create an issue with the idea first to avoid you doing work on things that will not work out. But I sincerely hope you had fun building these anyway 😄

[caseyWebb]

Thanks for the speedy reply!

This is very much a quick and dirty proof of concept, still very incomplete (as noted by the bad Dict simplification 😅). Mostly wanted to get some eyes on it. Tbh I initially started working on this to run as a one-time janitor in our codebase (and it's waaaay easier to fork elm-review-simplify for this kind of stuff than start a new rule from scratch–seriously great work here), but figured a PR doesn't hurt anything.

[caseyWebb]

Yanked the conditionals, fixed the Dict case, added support for lambdas 🎉

[jfmengels]

Sorry that I took a bit longer to reply this time.

I think the code looks generally good, but I have noticed a few cases where the rule incorrectly applies a simplification. I've added those in this commit which you can cherry-pick. a98dc7c

Feel free to decide that you'd like to stick to a simpler set of simplifications. All is fine by me, and I wanted to say I highly appreciate the work you've put in, thank you ☺️