[Haskell-cafe] Which is the most "hackable" compiler?

[Sergey Bushnyak]

You should go for a long run with GHC, it has most advanced extension 
system and you'll be able to integrate your extensions with existing 
codebase if they come out successful.

If you still want to work with hackablel compiler look at jhc 
https://github.com/jimcrayne/jhc and it's fork ajhc 
https://github.com/ajhc/ajhc they are very low-level but codebase very 
small and you can grasp whole system very quickly to start hacking


On 08/12/2016 10:58 PM, MarLinn via Haskell-Cafe wrote:
> So I have some crazy ideas for language extensions/reinterpretations 
> and ways I might try to implement them. But ghc is a complex beast and 
> my crazy ideas have too many interrelationships for a noob like me to 
> try to bold them onto it. After all, I have never worked on ghc - or 
> any compiler for that matter. I did look at the code, but I reckon 
> maybe I should gather some experience first before diving into that 
> one. It would be even worse to offload my ideas onto someone else 
> before I at least tried them and got some feedback. No fun for anyone.
>
> But of course I don't want to start from scratch. As fun as that could 
> be, there's no need to reinvent all the wheels just for a different 
> axle. And I hear rumors that there are other Haskell compilers out 
> there, even if most of them live in the shadow/slipstream of their 
> brother. Alas, I don't have any idea what their statuses and 
> philosophies are.
>
> So maybe you can help me here: Have you experienced any of the 
> alternative compilers as especially easy for a newcomer to pick up and 
> play around with? If it helps, I would be satisfied with plain 
> Haskell2010 or even Haskell 98, although some GADT and/or TypeFamilies 
> code to butcher would be nice, too. The ideas are mostly about larger 
> scale structures like whole functions. One representative example idea 
> is "Could it help the implementation - and does it even make sense - 
> to view a module as just a weirdly written zero parameter type class?" 
> As I said: crazy ideas.
>
> I'm thankful for any thoughts and ideas.
>
>
> MarLinn
>
> _______________________________________________
> Haskell-Cafe mailing list
> To (un)subscribe, modify options or view archives go to:
> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
> Only members subscribed via the mailman list are allowed to post.