- Haskell-Cafe - Haskell.org

A GHC error message puzzle
by Yitzchak Gale 16 Aug '10

16 Aug '10

The file "error_puzzle.hs" begins like this: main = do inp <- readFile "input" writeFile "output" $ process inp process :: String -> String When compiled with GHC 6.12.3 and run, it gives the following result: $ ./error_puzzle error_puzzle: output: hClose: illegal operation (handle is finalized) What is the cause of the error? Give a definition for process and an input file that reproduce this result. No "unsafe" functions like unsafePerformIO are allowed. Have fun! -Yitz

9 24

xemacs newbie question - haskell mode
by rgowka1 16 Aug '10

16 Aug '10

HI - I have been struggling to get the Xemacs recognize hs file. I have installed the Haskell-mode.. However I keep getting the message "File mode specification error: (wrong-number-of-arguments require 3)".. How do I go about fixing this?? Many thanks

2 1

Assorted AT fun and games
by Andrew Coppin 16 Aug '10

16 Aug '10

It all began with class (Vector (Point x)) => HasSpace x where type Point x :: * So far, so good. I was rather surprised that you _can_ write class (Complete (Completed x)) => Incomplete x where type Completed x :: * complete :: x -> Completed x I was almost as surprised to discover that you _cannot_ write class (HasSpace x, Complete (Completed x), HasSpace (Completed x), Point x ~ Point (Completed x)) => Incomplete where... It just means that every time you write an Incomplete instance, you might have to add the Point constraint manually. Which is mildly irritating. More worrying, adding Point foo ~ Point bar to an instance declaration causes GHC to demand that you turn on Undecidable Instances, for reasons beyond my comprehension. It's also interesting that when you write a class instance that has some constraint on it, and then you try to write a subclass instance, you still have to repeat the exact same constraint, even though the superclass instance declaration implies it. The only reason I can think of is that theoretically somebody could add a new superclass instance without the constraint. (Wouldn't that be an overlapping instance though?) And now things get *really* interesting. Consider this: data Foo x = Foo !x !(Point x) Surprisingly, GHC accepts this. This despite the rather obvious fact that "Point x" can exist if and only if "x" has a HasSpace instance. And yet, the type definition appears to say that "x" is simply an unconstrained type variable. Intriguing... Next, you can't derive Eq or Show for this type, but you *can* declare them manually: instance (Show x, Show (Point x)) => Show (Foo x) where show (Foo x px) = "Foo (" ++ show x ++ ") (" ++ show px ++ ")" Again, no hint of the fact that this will only work if we have HasSpace x. And yet GHC happily accepts this. I'm starting to think maybe I'm pushing the type system further than it can cope, and I should just completely redesign the whole thing...

4 4

couchdb - newView command
by Andrew U. Frank 15 Aug '10

15 Aug '10

i use haskell couchdb 0.10 with more luck than others: it works for me quite well. I have been careful to convert to json strings separately from the calls to runCouchDB - so i could check that the values were correct. i sense there is a difficulty with 'newView' - it does not allow to update a viewset, but one has to select for each change a new viewset (design) name; which is sort of annoying. the newView operation should ptobably be called 'newViewset', because it sets multiple views in one design document. it would be nice, if that could not only be set, but also updated. (the full set at once) i would also recommend that the first parameter is of type DB and not String - or are there any particular reasons for this? i appreciate the package and i like it! andrew

2 1

hgettext and cabal
by abau 15 Aug '10

15 Aug '10

Hi, [1] describes how to use hgettext with cabal. When running "runhaskell Setup build" this works fine. But "cabal build" says: Not in scope: `__MESSAGE_CATALOG_DOMAIN__' Not in scope: `__MESSAGE_CATALOG_DIR__' [1] says: "__MESSAGE_CATALOG_DOMAIN__ and __MESSAGE_CATALOG_DIR__ are macro definitions, whose hold configured strings from the Cabal." Why "cabal build" does not expand these macros? Thank you. [1] http://www.haskell.org/haskellwiki/Internationalization_of_Haskell_programs

1 0

Re: philosophy of Haskell
by Patai Gergely 15 Aug '10

15 Aug '10

> No. As you say the world1 value is immutable, but that's not > contradictory. If between 'getLine' and 'print' something was done by a > concurrent thread, then that change to the world is captured by 'print'. Well, that's technically true, but it basically amounts to saying that the 'model' of IO is itself. If I see 'print x', I don't really think of an action that performs arbitrary side effects and prints x in the process (unless some error prevents even that). Some notion of compositionality is required, which would allow me to state how 'print x' contributes to the changes of the world in a way that doesn't depend on any context. Actions in the state monad are composable in that sense, while the same doesn't apply to the IO monad when explained in terms of state passing. Gergely -- http://www.fastmail.fm - The professional email service

2 1

Is bumping the version number evil, if it's not mandated by the PVP?
by Sebastian Fischer 15 Aug '10

15 Aug '10

Hello, I wonder whether (and how) I should increase the version number of a library when the API does not change but the implementation gets more efficient. Should I bump a.b.C or even a.B to signal that it's worth using the new version or should I bump only a.b.c.D such that packages that depend on a.b get installed with the new version automatically? When bumping only a.b.c.D, the new version is not installed as a dependency if the old version already is installed (unless the new version is explicitly demanded.) It seems bumping a.b.c.D has advantages for some users and disadvantages for others. Hence, I guess I should make a major version bump. Is it bad habit to make a major version bump if the API does not change? Maybe I should simply change the API too ;) Cheers, Sebastian -- Underestimating the novelty of the future is a time-honored tradition. (D.G.)

4 6

bug in Network.Browser
by Alexander Kotelnikov 15 Aug '10

15 Aug '10

Hello. Yesterday I hit a bug in Network.Browser: connection in the connection pool is not reused if you are connecting to a destination with qualified port (not 80 for HTTP or 80 but explicitly provided like http://www.google.com:80) The reason for the failure is quite trivial, but I had not found an easy fix which won't break anything. Problem is that in TCP.hs function isTCPConnectedTo :: HandleStream ty -> String -> IO Bool isTCPConnectedTo conn name = do v <- readMVar (getRef conn) case v of ConnClosed -> return False _ | map toLower (connHost v) == map toLower name -> catch (getPeerName (connSock v) >> return True) (const $ return False) | otherwise -> return False socket's hostname, (connHost v) is compared to hostname or hostname:port which is stored in name. Hope this message will reach this code maintainers Alexander

1 0

Help us test gio APIs!
by Andy Stewart 15 Aug '10

15 Aug '10

Hi all, I'm working on merge gio-branch (https://patch-tag.com/r/AndyStewart/gio-branch/home) to gtk2hs-0.12.0. GIO (http://library.gnome.org/devel/gio/stable/) is cross-platform APIs for file operation, we can use gio APIs develop file manager or similar application. If anyone want to use those APIs, help us test those APIs at https://patch-tag.com/r/AndyStewart/gio-branch/home, then we can release gio-0.12.0 quicker. A GIO file-manager screenshot at : http://www.flickr.com/photos/48809572@N02/4793031888/ Thanks! -- Andy

1 0

Removing polymorphism from type classes (viz. Functor) (Again)
by Alexey Karakulov 15 Aug '10

15 Aug '10

I was inspired by George Pollard's post<http://www.haskell.org/pipermail/haskell-cafe/2009-July/063981.html>at haskell-cafe and tried to implement the non-polymorphic Functor class ( I named it Functor' ). I changed some names and added reasonable constraints. type family NewPt f a class Functor' f where type Point f map ∷ (a ~ Point f, b ~ Point g, g ~ NewPt f b, Functor' g) ⇒ (a → b) → f → g I would like to be able to write: type OldPt f = NewPt f (Point f) class (f ~ OldPt f) ⇒ Functor' f ... but ghc says it's not implemented yet (version 6.12.1). However, it's not the main problem. Now I can write some instances: type instance NewPt [a] b = [b] instance Functor' [a] where type Point [a] = a map = fmap type instance NewPt ByteString a = ByteString instance Functor' ByteString where type Point ByteString = Word8 map = BS.map But I can't write instance for Set: type instance NewPt (Set a) b = Set b instance Ord a ⇒ Functor' (Set a) where type Point (Set a) = a map = Set.map ghci complains: Could not deduce (Ord a1) from the context (g ~ NewPt (Set a) a1, a1 ~ Point g, Functor' g) arising from a use of `Set.map' at ... The type of Set.map is Set.map :: (Ord a, Ord b) => (a -> b) -> Set a -> Set b (Ord a) is in the instance context, and what about b? Type of map for Set instance would be: original: map ∷ (a ~ Point f, b ~ Point g, g ~ NewPt f b, Functor' g) ⇒ (a → b) → f → g substitute: f → Set a, g → Set b map :: Functor' (Set b) ⇒ (a →b) →Set a →Set b (Ord b) must be deduced from (Functor (Set b)) but it doesn't. I don't know whether it's my mistake somewhere or ghc problem. (Sorry for my English, it's not perfect). -- All the best, Alexey

3 7