a stateful bug
Frederik Eaton
frederik at a5.repetae.net
Sun Aug 21 05:33:57 EDT 2005
Hi all,
If you execute the following commands:
$ wget http://ofb.net/~frederik/futility/futility-0.1.0.tar.gz
$ tar -xvzf futility-0.1.0.tar.gz
$ cd futility-0.1.0/
$ ./do-build conf=1
Then if all goes well you should see an error like this:
./Fu/MonadComp.hs:46:0:
Duplicate instance declarations:
./Fu/MonadComp.hs:46:0: instance Monad (Either e)
Imported from Control.Monad.Error:
instance (Control.Monad.Error.Error e) => Monad (Either e)
Then, when you try to build again, after having changed nothing:
$ runghc Setup.hs build
the build suddenly succeeds!
Correct me if I'm wrong, but shouldn't a properly working compiler
either fail or succeed on a given package, and not fail one time, and
then succeed the next? I mean, this is the whole referential
transparency deal right? :)
I can reproduce this bug on ghc 6.4 and 6.4.1.20050812 on two i386
computers.
Frederik
--
http://ofb.net/~frederik/
More information about the Glasgow-haskell-bugs
mailing list