[Haskell-cafe] Replace data constructors via meta programming

[Vilem-Benjamin Liepelt]

Oh, indeed I wasn't aware of this. That's great news—thank you, Frerich!

This actually leads me to another question: what are the tradeoffs between implementing this via Template Haskell as in the case of this package vs the `deriving` mechanism as for the foldable instance?

V

> On 2018-02-13, at 00:03, Theodore Lief Gannon <tanuki at gmail.com> wrote:
> 
> Vilem, you may have missed the post from the catamorphisms author where he updated the library in response to this thread? :)
> 
> On Feb 12, 2018 3:58 PM, "Vilem-Benjamin Liepelt" <vl81 at kent.ac.uk <mailto:vl81 at kent.ac.uk>> wrote:
> Thank you for your great suggestions.
> 
> The type-foo looks very cool and I will have to dig more into the different options.
> 
> Generating a catamorphism using the library of the same name works like a charm and integrates well with my existing code—once I managed to get it to install (thank you Stack LTS!) it just worked out of the box and let me write a one-line evaluator and a one-line pretty-printer. Woohoo!
> 
> Unfortunately It's not really viable to use this "for real" at the moment because it requires such an old version of GHC.
> 
> Something I miss is the clear correspondence between data constructors and "substitutions" (like in my `magic` example), since they become entirely positional, leading to potentially brittle code (imagine reordering the data constructors). I think some of the other solutions might be better in this respect.
> 
> It's a shame that the catamorphism package doesn't work for a more up-to-date version of GHC, because I think I would use this quite often.
> 
> I suppose the fold that Sergey proposed is essentially what the catamorphism package generates. Although I want to avoid having to write functions by hand when really the computer should be doing them for me, I think I will use this for now as it integrates nicely with my existing code and leads to quite idiomatic Haskell.
> 
> I will definitely check out the other suggestions as well though, thank you again.
> 
> Best,
> 
> Vilem
> 
> > On 2018-02-12, at 09:08, Frerich Raabe <raabe at froglogic.com <mailto:raabe at froglogic.com>> wrote:
> >
> > On 2018-02-12 03:30, Vilem-Benjamin Liepelt wrote:
> >> I am looking for a solution to get rid of this silly boilerplate:
> >> eval :: Ord var => Map var Bool -> Proposition var -> Bool
> >> eval ctx prop = evalP $ fmap (ctx Map.!) prop
> >>  where
> >>    evalP = \case
> >>        Var b -> b
> >>        Not q -> not $ evalP q
> >>        And p q -> evalP p && evalP q
> >>        Or p q -> evalP p || evalP q
> >>        If p q -> evalP p ==> evalP q
> >>        Iff p q -> evalP p == evalP q
> >
> > [..]
> >
> > You might benefit from the 'catamorphism' package:
> >
> >  https://hackage.haskell.org/package/catamorphism-0.5.1.0/docs/Data-Morphism-Cata.html <https://hackage.haskell.org/package/catamorphism-0.5.1.0/docs/Data-Morphism-Cata.html>
> >
> > It provides a template Haskell function which, given a data type, produces a function which reduces (folds) that data type.
> >
> > --
> > Frerich Raabe - raabe at froglogic.com <mailto:raabe at froglogic.com>
> > www.froglogic.com <http://www.froglogic.com/> - Multi-Platform GUI Testing
> 
> _______________________________________________
> Haskell-Cafe mailing list
> To (un)subscribe, modify options or view archives go to:
> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe <http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe>
> Only members subscribed via the mailman list are allowed to post.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/haskell-cafe/attachments/20180213/4fa76543/attachment.html>