On Wed, Feb 01, 2006 at 09:40:20AM -0500, Robert Dockins wrote:
Additionally, a standard for core would allow a new level of tool interoperability. Haskell front ends and backends could be cleanly separated along a well-defined border. DrIFT and Haddock and others as well could benefit. Happy could generate core directly.... anyway you get the idea.
Yeah, i have always wanted ghc to allow one to inject a preprocessor (midprocessor?) after desugaring but while it is still haskell. it would allow preprocessors to not worry about a lot of the front-end extensions and let the compiler have first crack at reporting type errors to the user before they are obfcusiated by any preproccesing. I think a formal definition of core-haskell would be an excellent thing. John -- John Meacham - ⑆repetae.net⑆john⑈