[Haskell-cafe] fixity declaration in GHCi

[Takayuki Muranushi]

Hello everyone,

while I'm playing with fixity declaration in GHCi, I found a strange
behavior. Can anyone explain this?
Is this actually an expected behavior, or is it a glitch?

----------------------------------------------------------------
$ ghci
GHCi, version 7.0.3: http://www.haskell.org/ghc/  :? for help
Loading package ghc-prim ... linking ... done.
Loading package integer-gmp ... linking ... done.
Loading package base ... linking ... done.
Loading package ffi-1.0 ... linking ... done.
Prelude> minus    --(1)

<interactive>:1:1: Not in scope: `minus'
Prelude> let minus x y = x-y; infixl 6 `minus`    --(2)
Prelude> 10^5 `minus` 2    --(3)
1000
Prelude> let infixl 6 `minus`    --(4)

<interactive>:1:14:
    The fixity declaration for `minus' lacks an accompanying binding
Prelude> let subtract x y = x-y    --(5)
Prelude> 10^5 `subtract` 2    --(6)
1000
Prelude> let f x y = x-y; infixl 6 `f` in 10^5 `f` 2    --(7)
99998
Prelude>
----------------------------------------------------------------

(1) suggests that `minus` is not defined here. in (2), I try to define
`minus` with precedence level 6.
However, (3) suggests that precedence level of `minus` is stronger than ^ .

(4) suggests that you cannot redefine fixity here.
(5)(6) is consistent with Haskell 98 report section 4.4.2; "Any
operator lacking a fixity declaration is assumed to be infixl 9."
in (7), precedence level is set as expected.

So, what is happening in (3)? Where did the fixity declaration in (2) went?

Best,

-- 
MURANUSHI Takayuki
The Hakubi Center, Kyoto University : http://www.hakubi.kyoto-u.ac.jp/