March 2006 - Hugs-Users

formatting and guarded expressions
by Wittie 23 Mar '06

23 Mar '06

Hi, Are there special formatting constraints for using guards in Hugs? My students are learning Haskell in lab and the following simple function sometimes fails >fact n > | n == 0 = 1 > | otherwise = n * (fact(n-1)) Hugs complains about an unexpected ; Sometimes code after a guarded expression fixes the problem. We've tried putting the | up against the > or underneath the parameter. The same code seems to work for some students but not for others. Any ideas? - Lea

2 1

one more problem with Hugs
by Bulat Ziganshin 08 Mar '06

08 Mar '06

Hello , The following code compiles with GHC but not with Hugs. moreover, if you replace STURef with STRef in instance definition, all will be fine. bug? {-# OPTIONS_GHC -cpp -fglasgow-exts #-} import Control.Monad.ST import Data.STRef -- ----------------------------------------------------------------------------- -- Unboxed references in ST monad type STURef = STRef -- | Create new unboxed reference in ST monad newSTURef :: e -> ST s (STURef s e) newSTURef = newSTRef -- ----------------------------------------------------------------------------- -- Monad-neutral interface for fast unboxed references class (Monad m) => URef m r | m->r, r->m where newURef :: a -> m (r a) instance URef (ST s) (STURef s) where newURef = newSTURef -- ----------------------------------------------------------------------------- -- Main test = runST ( do newURef (0::Int) return '1' ) -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com

1 1

the bug or not the bug? :)
by Bulat Ziganshin 03 Mar '06

03 Mar '06

Hello one more example that works in GHC but not Hugs. why i'm so unlucky? :) import Foreign.Storable import Foreign.ForeignPtr class (Monad m) => STorIO m s | m->s where mLift :: IO a -> m a newtype MBA s a = MBA (ForeignPtr a) class (Storable value) => Unboxed value where readUnboxed :: (STorIO m s) => MBA s value -> m value readUnboxed (MBA arr) = mLift $ withForeignPtr arr $ \a -> peek a -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com

1 0

Fwd: Re[2]: [Hugs-users] Status of Hugs? (last major release for Windows was 2003..)
by Bulat Ziganshin 01 Mar '06

01 Mar '06

it seems that this message was lost somewhere... This is a forwarded message From: Bulat.Ziganshin(a)gmail.com <Bulat.Ziganshin(a)gmail.com> To: "Neil Mitchell" <ndmitchell(a)gmail.com> Date: Monday, February 20, 2006, 1:03:13 PM Subject: [Hugs-users] Status of Hugs? (last major release for Windows was 2003..) ===8<==============Original message text=============== Hello Neil, Monday, February 20, 2006, 6:31:44 AM, you wrote: >> > http://www-users.cs.york.ac.uk/~ndm/projects/winhugs.php >> >> Yes, please try that out. When WinHugs is ready, there will be a Hugs >> release. NM> It's pretty much ready now, I'll package up some final binaries NM> tomorrow - whenever Hugs is in a ready state I'm happy to release. i still have some suggestions before the final version can be rolled. it's from docs on my Streams library: The library fully supports Hugs 2003 and Hugs 2005, but 1) support for FD and MMFile is temporarily disabled because I don't know how to build DLLs 2) Hugs 2003 doesn't include support for "instance Bits Word" and vGetBuf/vPutBuf, so you need to add these implementations manually or delete the lines that use it (look for "2003" in the sources) 3) WinHugs doesn't support preprocessing, so I included the MakeHugs.cmd script to preprocess source files using cpphs so, list of my suggestions for Hugs include: * add preprocessing symbol HUGS_VERSION to distinguish Hugs06 from previous builds and to allow distinguish future Hugs modifications * (optional) add option to disable monomorphism restriction. as far as i understand this restriction is just additional part of code and it should be easy to skip this code when requested * (optional) add hSetFileSize support. it should be easy using ftruncate() and it will increase compatibility between Hugs ans GHC * afair, at least one error i reported was not fixed immmediately. i don't read hugs-bugs but if this still unfixed, it seems strange to me to release Hugs with known errors for WinHugs such list is bigger: * ffihugs * preprocessor. i'm displeased by the requirement to ship myrriad versions of my library, one for each Hugs version+platform combination. in addition, lack of preprocessor makes my edit/test cycle more complicated and less intuitive. On the other side, adding preprocessor to WInHugs should be very easy: instead of readig the source file as is just run "cpphs.exe --noline -D__HUGS__ -D__HUGS_VERSION__=2005 -DSIZEOF_HSINT=4 -DSIZEOF_HSWORD=4" or something like this on the file and read the resulting temporary file. i do't see _any_ problem in implementing this. CPPHS is easily build by ghc/mingw, its output with "--noline" don't contain any directives and all lines is just at their right places. this will finally allow to include in Hugs distribution original (non preprocessed) sources of standard libs i also have suggestions about WinHugs that should not delay its releasing: * reduce size of installer. it can be done in many ways but at least i'm not glad to download 16 megs that can be really packed in 2 megs * use text file to store global commands history thank you for patience :) -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com ===8<===========End of original message text=========== -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com

1 0