I think this is a mistake, yes. They should not raise such exceptions, but rather just wrap around—minBound `quot` (-1) should be -minBound=minBound. That would justify the behavior of rem and mod, and makes much more sense than the current behavior for Int as a ring.
According to the documentation, rem and mod must satisfy the following
laws:
-- > (x `quot` y)*y + (x `rem` y) == x
rem
-- > (x `div` y)*y + (x `mod` y) == x
mod
https://hackage.haskell.org/package/base-4.8.0.0/docs/src/GHC-Real.html
Note, however, that there is a case when quot and div result in an
arithmetic overflow:
Prelude> (minBound :: Int) `quot` (-1)
*** Exception: arithmetic overflow
Prelude> (minBound :: Int) `div` (-1)
*** Exception: arithmetic overflow
while rem and mod don't:
Prelude> (minBound :: Int) `rem` (-1)
0
Prelude> (minBound :: Int) `mod` (-1)
0
Is this a mistake?
For the record, I'm aware of the safeint package, which raises the error
for rem and mod, and this ticket:
https://ghc.haskell.org/trac/ghc/ticket/8695
_______________________________________________
Glasgow-haskell-users mailing list
Glasgow-haskell-users@haskell.org
http://mail.haskell.org/cgi-bin/mailman/listinfo/glasgow-haskell-users