datasets : lint with brittany about dh-core HOT 4 CLOSED

This is a good suggestion!

https://hackage.haskell.org/package/brittany

Note: brittany is pinned to a given GHC version, since it must parse and emit code for it. This means that it constrains the range of GHC versions we build dh-core with.

For example, to try it out now I used :

stack install brittany --resolver lts-12.0

from dh-core.

I don't need datasets but

1. Requiring installed and configured brittany to make contribution will certainly raise the bar for contributors.

2. There's no reason to constrain supported GHC version just because autoformatter doesn't. It's possible to drop formatting check from CI for unsupported version

P.S. I'm not big fan of autoformatters for haskell. In my (limited) experience they tend to produce weird layout.

from dh-core.

After sleeping on this, I agree with you, @Shimuuar . The main problem is that datasets is terribly formatted right now; it would be nice to have someone do a cleanup pass for future diffs.

I don't think I've ever had an autoformatting tool that everyone was happy with. Brittany is pretty good for me, but I'm also unsatisfied with some of the defaults. In the aforementioned cleanup pass, if an autoformatter is used we should just save the rules so that they can be reused (or commented on) in case we need to do it again (or periodically). Maybe if dh-core gets a windfall of success adding linting, formatting, and weeding rules to CI should be revisited. No need to do this anytime soon.

from dh-core.

Scanning through this repo -- I think we can close this until further notice.

from dh-core.