Hi Kartik, I think I'd be confident in saying that Yhc should now be able to run anything that nhc could. And if it can't then we'd be interested in seeing the program so we can fix it. Of course things which aren't Haskell 98 won't work (sadly we need someone to rewrite our type system to go beyond Haskell 98 ...). I would also say that the Yhc runtime is both a lot cleaner and a lot better documented that the nhc one. However, this is of course my own and very biased opinion, since I wrote the runtime :-) Yhc is also being much more actively developed than nhc, which means it should be easier to get help with problems. Your project sounds interesting, what is it that you'd like to do with yhc more exactly? Haskell is a notoriously hard language to write a useful debugger for (due to laziness), also a large part of my research is writing a tracing/debugging tool for Haskell using yhc, so I'd be interested to hear your plans :-) Anyway, hope that's informative :-) Cheers Tom On Wednesday 13 September 2006 12:16, Kartik Vaddadi wrote:
Hello, I'm building a debugger for Haskell that works by modifying the runtime, and I need to choose between yhc & nhc. Everything I've seen points towards yhc, but I wonder how stable it is. I know it's not stable for production use as a compiler, but as a platform for a debugger? For my purposes, it does not have to work with every program. I find that yhc works fine with almost all of the half-dozen small programs (<1 page) I threw at it, but chokes on many of the hugs demos.
So would you suggest I work with yhc rather than nhc? Any comments you can give me are greatly appreciated. Thanks.