Re: Crytpo for nhc
"Dominic Steinitz" <[email protected]> Sun, 18 May 2003 12:02:44 +0100
| Newsgroups | gmane.comp.lang.haskell.libraries,gmane.comp.lang.haskell.nhc.user |
|---|---|
| Message-ID | <[email protected]> |
Malcolm, Thanks for this. I'm restructuring the Crypto libraries so that it will contain Crypto.Modes, Crypto.DES, Crypto.Blowfish etc. The current Blowfish should work with nhc. When I get the chance, I'll see if I can restructure DES but it's not top of my priorities at the moment. [Dominic Steinitz ----- Original Message ----- From: "Malcolm Wallace" <[email protected]> To: "Dominic Steinitz" <[email protected]> Cc: <[email protected]>; <[email protected]> Sent: Tuesday, April 29, 2003 2:33 PM Subject: Re: Crytpo for nhc > "Dominic Steinitz" <[email protected]> writes: > > > I have been trying to write a Crypto library and so far it works with ghc > > and hugs. > > > > Am I right in thinking I am going to struggle with nhc? It uses Word8 and > > Word64 heavily. For example, cipher block chaining relies on xoring Word64. > > The things you mention are not a particular problem for nhc98. > The only problem I encountered in the Crypto.hs code is that its use > of the type class machinery is not Haskell'98 compliant. > > instance Bits [Bool] where ... > > is illegal in Haskell'98, because an instance head must be either a > simple type name, or a type constructor applied to only type variables. > (See the Report section 4.3.2) If ghc and Hugs permit this, then it is > a non-standard extension to the language. > > Unfortunately, I don't immediately see an easy way to convert this > construction into something more amenable. > > Regards, > Malcolm