Hello, The discussion on the libraries list is archived here: http://www.haskell.org/pipermail/libraries/2011-September/016699.html There hasn't been a corresponding discussion for Haskell Prime so, technically, GHC deviates from the standard. -Iavor On Fri, Dec 23, 2011 at 9:41 AM, Simon Peyton-Jones wrote:

I'm confused too.  I'd welcome clarification from the Haskell Prime folk.

S

-----Original Message----- From: Serge D. Mechveliani [mailto:mechvel@botik.ru] Sent: 23 December 2011 17:36 To: Simon Peyton-Jones Subject: Re: 7.4.1-pre: Show & Integral

On Thu, Dec 22, 2011 at 08:14:54PM +0000, Simon Peyton-Jones wrote:

...

|  2011/12/22 Edward Kmett : |  > The change, however, was a deliberate _break_ with the standard that |  > passed through the library review process a few months ago, and is now |  > making its way out into the wild. | |  Is it reasonable to enquire how many standard-compliant implementations |  of Haskell there are?

Just to be clear, the change IS the standard.  GHC has to change to be compliant. At least that's how I understand it.

I am confused. I am looking now at the on-line specification of  Haskell-2010, 6.3 Standard Haskell Classes. It shows that  Integral  includes  Show:

                          Eq     Show                             \   /                              Num                              |                       Enum  Real                          \   |                           Integral

This is also visible in the further standard class declarations in this chapter.

Hence, for  `x :: Integral a => a'  it is correct to write  (shows x ""). And  ghc-7.4.0.20111219  does not allow this. So,  ghc-7.4.0.20111219  breaks the 2010 standard. Now, Edward Kmett writes that this break is done deliberately.

Am I missing something?

I witness this for the first time: that GHC deliberately breaks the current Haskell standard. Probably, many people (as myself) dislike this point of the standard. Well, they can write a dummy Show implementation for their type T:      showsPrec _ _ = showString "(<t> :: T)",

and wait for an improved standard, say, Haskell-II -- ?

Regards,

------ Sergei mechvel@botik.ru

_______________________________________________ Haskell-prime mailing list Haskell-prime@haskell.org http://www.haskell.org/mailman/listinfo/haskell-prime