Bug in Curry type-checker

From: Wolfgang Lux <wlux_at_uni-muenster.de>
Date: Mon, 18 Dec 2006 12:04:27 +0100

I wrote:

> The Curry type checker *is* correct (I'm sure, since I wrote it :-).

It turns out that my claim is wrong -- at least as far as MCC 0.9.10 and
PAKCS 1.7.2 are concerned. Both implementations accept the following
program even though it is not type correct

   strangeList = x=:=1 &> f True "Hello"
     where (x:xs) = f True (repeat unknown)
           f _ [] = []
           f b (y:ys) = (if b then y else x) : f (not b) ys

I've fixed this bug recently in MCC's source tree and maybe someone can
apply the fix to PAKCS' source tree.


curry mailing list
Received on Mo Dez 18 2006 - 12:39:18 CET

This archive was generated by hypermail 2.3.0 : So Dez 03 2023 - 07:15:09 CET