[Haskell-cafe] Haskell I/O inside: down the rabbit's hole

[Bulat Ziganshin]

Hello Krasimir,

Friday, June 30, 2006, 2:26:53 PM, you wrote:

> In the last section you said that IO has the same definition in GHC
> and Hugs. As far as I can remember in Hugs IO is defined as:

> newtype IO a = IO ((a -> IOResult) -> (IOError -> IOResult) -> IOResult)

> Of course this definition is builtin and cann't be seen in the Haskell
> source code.

i'm not sure about this - can this be considered as an example of this
RealWorld strategy or not. Feel free to edit text if you know the
right answer


-- 
Best regards,
 Bulat                            mailto:Bulat.Ziganshin at gmail.com