> ghc-2.08-linux reports
>
> : Couldn't match the kind
>`ka10725 -> *' against `*'
> When unifying two kinds `* -> *' and `*'
> In an interface-file signature for `$gL'
>
>
> Many types have been modified in the module to include the additionl
> type
ghc-2.08-linux reports
: Couldn't match the kind
`ka10725 -> *' against `*'
When unifying two kinds `* -> *' and `*'
In an interface-file signature for `$gL'
Many types have been modified in the module to include the additionl
type parameter. So it st
I've been building GHC 2.08 on Win32 (having managed to get a 0.29 build
first). The basic thing seems to work, but I can't get the profiling
version up. The first problem was three missing symbols timeBeginPeriod,
timeSetEvent and timeKillEvent. They're Windows functions which gynus
provide an in
> I can't remember having seen anything like this before.
> The following program compiles fine with ghc2.0698
>
> > module Main( main ) where
>
> > main = return ()
> >where 0 = 1
Looks fine to me. The pattern "0" is matched lazily, that is
when any of its free variables
ghc-2.08-linux reports that `#-}' is wrongly placed in the program:
module M where
f :: Integral a => a
f = 1
{-# SPECIALIZE f :: Int
#-}
Is ghc right?
Does this pragma parenthesis conflicts with the cpp preprocessor key
symbol `#' - when in th
Hello there,
I can't remember having seen anything like this before.
The following program compiles fine with ghc2.0698
> module Main( main ) where
> main = return ()
>where 0 = 1
Hope this helps.
Regards,
Marc van Dongen
With ghc-2.08 under x86_Solaris2.5.1, the included program, "lc_1f.hs", failed to
execute due to the following troubles:
(1) A bug in cbits/getDirectoryContents.lc
# Sorry, if this bug has already been reported.
The fix is included.
(2) "doesDirectoryExist" failed due to EVERFLOW.
I'm not
