
I can't answer all those questions, but I believe CoreSynType is the
AST representing Core.
On Tue, Jan 27, 2015 at 10:12 PM, Wojciech Danilo
Hello David, thank you for your response! :) I've got few questions regarding it: 1) Does it mean there are no datatypes dedicated to "a core language" ? 2) All passes are transforming AST -> AST? If so, what does the function `compileToCoreModule` does? 3) Additional - is there any subset of the AST that we can call "CORE" ? 4) Do all the documentation (like this one: https://ghc.haskell.org/trac/ghc/wiki/Commentary/Compiler/CoreSynType) should be considered obsolete?
All the best, Wojciech
Wed Jan 28 2015 at 3:55:12 AM użytkownik David Feuer
napisał: There used to be an external representation of core intended to be used for such things. It was recently removed because it had gotten stale and no one wanted to take responsibility for maintaining it. So at the moment, Core is really just an AST, not a proper language, and you'll have to hack GHC if you want to inject it.
On Tue, Jan 27, 2015 at 6:10 PM, Wojciech Danilo
wrote: Hello All! :) Recently I've came across a problem and hopefully you could help me with it. Basically I and some people here are running a startup and we have created a language for processing images. This language has 2 interchangeable representations - textual and visual, dataflow one.
Right now we are compiling it to haskell sources, but we want to move over GHC CORE. Getting into GHC has a high learning curve, as far as we see, and we would be very thankfull for any help with a minimal working example of compiling haskell sources to CORE and then CORE to binaries. I've posted couple days ago a detailed question on StackOverflow, here:
http://stackoverflow.com/questions/28059669/using-ghc-api-to-compile-haskell...
I would be very thankful for any help or hints! All the best, Wojciech
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe