- Haskell-Cafe - Haskell.org

Config Files and monoids
by Olaf Klinke 30 May '18

30 May '18

Dear cafe, a recent post here [1] mentioned that configurations, such as the ones read from a config file, can be given Monoid instances, where mempty is the empty or default configuration and mappend merges two partial configurations, producing a more complete one. The vgrep package explicitly does this, for instance. Although the ConfigParser type from the ConfigFile package has a binary 'merge' operation, it does define neither a Monoid not a Semigroup instance. I'm struggling to make the concept of monoidal configuration work when there is no sensible default configuration. Suppose my configuration type is data Config = Config {foo :: Bool, bar :: Int} with no reasonable default, e.g. emptyConfig = Config { foo = error "you did not specify option foo", bar = error "you did not specify option bar" } Some configuration monoids seem to have the second operand override the first, or the other way around. However, I wish that when cfg1 = emptyConfig {foo = True} cfg2 = emptyConfig {bar = 4} then cfg1 <> cfg2 == Config {foo = True, bar = 4}. So it seems that for mappend to work as intended one needs a terminating function that tells me if a record field is already defined, e.g. when all fields are Maybes. Vgrep.Environment.Config.Monoid does it this way. My solution so far was to resort to the monoid of endofunctions (as the getflag package does), that is, define cfg1, cfg2 :: Config -> Config cfg1 = \cfg -> cfg {foo = True} cfg2 = \cfg -> cfg {bar = 4} And then build (cfg1.cfg2) emptyConfig. (Alternatively, one might structure these as lenses instead of endofunctions, see e.g. Data.Monoid.Endo.Fold in the endo package.) Thus I arrived at class Config cfg where emptyConfig :: cfg -- may contain some defaults configOptions :: [Parser (cfg -> cfg)] Do you think every other concept of configuration parsing can be cast into this typeclass? -- Olaf [1] https://mail.haskell.org/pipermail/haskell-cafe/2018-May/129063.html

5 5

Re: [Haskell-cafe] (names for) invariants for Eq and Ord?
by Olaf Klinke 30 May '18

30 May '18

Prelude> import qualified Data.Map as M Prelude M> let k = 0/0 in let v = 0 in M.lookup k (M.insert k v M.empty) == Just v False This one is not surprising because the presence of 0/0 destroys the total order of Double, so strictly speaking this should not have an Ord instance. Olaf

1 0

ANN: primitive 0.6.4.0
by Carter Schonwald 30 May '18

30 May '18

This is a pretty exciting (albeit minor version) I'd like to especially thank David Feuer (treeowl ) and Andrew Martin (andrewthad) for their contributions the change log has a huge number of bug fixes relative to 0.6.2-3 also we added some new modules of features PrimArray --- unboxed nonsliceable vectors MVar -- Mvars that run in any PrimMonad augmented the Data.Primitive exports for increased consistency (the new .Ptr and .Mvar modules aren't exported in this minor version so users can easily migrate) Changes in version 0.6.4.0 - Introduce Data.Primitive.PrimArray, which offers types and function for dealing with a ByteArray tagged with a phantom type variable for tracking the element type. - Implement isByteArrayPinned and isMutableByteArrayPinned. - Add Eq1, Ord1, Show1, and Read1 instances for Array and SmallArray. - Improve the test suite. This includes having property tests for typeclasses from base such as Eq, Ord, Functor, Applicative, Monad, IsList, Monoid, Foldable, and Traversable. - Fix the broken IsList instance for ByteArray. The old definition would allocate a byte array of the correct size and then leave the memory unitialized instead of writing the list elements to it. - Fix the broken Functor instance for Array. The old definition would allocate an array of the correct size with thunks for erroring installed at every index. It failed to replace these thunks with the result of the function applied to the elements of the argument array. - Fix the broken Applicative instances of Array and SmallArray. The old implementation of <*> for Arrayfailed to initialize some elements but correctly initialized others in the resulting Array. It is unclear what the old behavior of <*> was for SmallArray, but it was incorrect. - Fix the broken Monad instances for Array and SmallArray. - Fix the implementation of foldl1 in the Foldable instances for Array and SmallArray. In both cases, the old implementation simply returned the first element of the array and made no use of the other elements in the array. - Fix the implementation of mconcat in the Monoid instance for SmallArray. - Implement Data.Primitive.Ptr, implementations of Ptr functions that require a Prim constraint instead of a Storable constraint. - Add PrimUnlifted instances for TVar and MVar. - Use compareByteArrays# for the Eq and Ord instances of ByteArray when building with GHC 8.4 and newer. - Add Prim instances for lots of types in Foreign.C.Types and System.Posix.Types. - Reexport Data.Primitive.SmallArray and Data.Primitive.UnliftedArray from Data.Primitive. - Add fold functions and map function to Data.Primitive.UnliftedArray. Add typeclass instances for IsList, Ord, and Show. - Add defaultSetByteArray# and defaultSetOffAddr# to Data.Primitive.Types.

1 0

Selecting a transformer in a monad transformer stack
by Paul Brauner 30 May '18

30 May '18

Hello, I have a newtype for a monad transformer stack: newtype MyMonad a = MyMonad { runMyMonad :: ((ReaderT Env (ExceptT String (StateT Store (ListT Identity)))) a) } and I'd like to it derive MonadPlus. The default instance for MonadPlus I get via GeneralizedNewtypeDeriving is the ExceptT one because it is the outermost transformer implementing MonadPlus. I would like to use the ListT instance instead. The best solution I have come up with is to define: instance MonadPlus MyMonad where mzero = MyMonad $ ReaderT $ \e -> ExceptT $ StateT $ \s -> ListT $ Identity [] mplus a b = MyMonad $ ReaderT $ \e -> ExceptT $ StateT $ \s -> ListT $ Identity $ let run x = runIdentity (runListT (runStateT (runExceptT (runReaderT (runAbstract x) e)) s)) as = run a bs = run b in as ++ bs My questions are: - Is there a shorter way to define mplus? - Is there maybe some variation of monad transformer stacks where the layers are named and can be accessed by name? (I know of effects handlers but I'd like to stick to monad transformers for now.) Paul

2 3

8.4.2 FreeBSD bindist not working with gold & lld
by Sumit Raja 30 May '18

30 May '18

Hi, Just got 8.4.2 running in my FreeBSD 11 environment and am getting the following: cabal: Failed to build entropy-0.4.1.1 (which is required by exe:test-ghoa from oauth2-simple-0.1.0.0). The failure occurred during the configure step. The exception was: dieVerbatim: user error (cabal: '/usr/local/bin/ghc' exited with an error: /usr/local/bin/ld.gold: error: /usr/local/lib/ghc-8.4.2/rts/libHSrts_thr.a(RTS.thr_o): unexpected reloc 8 in object file LLD fails with configure step. The exception was: dieVerbatim: user error (cabal: '/usr/local/bin/ghc' exited with an error: /usr/bin/ld.lld: error: do not know how to handle relocation 'R_X86_64_RELATIVE' (8) Saw another thread that discussed this but there was no resolution. Any ideas on how to fix this? Thanks Sumit

2 1

Fwd: Safe TCP accept loop
by Sumit Raja 30 May '18

30 May '18

Hi, I sent this to beginners but got no responses so I'm hoping someone here can help. I’ve looked at https://wiki.haskell.org/Concurrency_demos/Graceful_exit and used(2) Using throwTo without the use of block and unblock. It runs on 8.2.1 with my limited testing. I can’t find out much about what the > 7.x GHC replacement for block/unblock is other than mask. What does the unblock do in acceptConnections'? Does this method of handling an accept loop still need masking of async exceptions ? If so where does this need to be done? I've got a version with async running for my specific application (https://bitbucket.org/sumitraja/hvrr/src/6927216597a35a9a0d7f2e55cca83fa5b7…) but I don't know if I need to unmask at any point so was hoping for some guidance. Thanks Sumit

2 4

test
by Gershom B 29 May '18

29 May '18

Testing this list, due to a report. -g

1 0

Munich Haskell Meeting, 2018-05-28 @ 19:30
by Leah Neukirchen 26 May '18

26 May '18

Dear all, Next week, our monthly Munich Haskell Meeting will take place again on Monday, May 28 at Hirschgarten (self-service area) at 19h30. (The weather should be fine for sitting outside, else can move inside.) For details see here: http://muenchen.haskell.bayern/dates.html If you plan to join, please add yourself to this dudle so we can reserve enough seats! It is OK to add yourself to the dudle anonymously or pseudonymously. https://dudle.inf.tu-dresden.de/haskell-munich-may-2018/ Everybody is welcome! cu, -- Leah Neukirchen <leah(a)vuxu.org> http://leah.zone

1 0

Optimizing nearest-k code
by Richard Evans 25 May '18

25 May '18

Dear Haskell Cafe, Given a set of sets, and a particular target set, I want to find the sets that are nearest (in terms of Hamming distance) to the target set. I am using the following code: import Data.List import qualified Data.Set as Set nearest_k :: Ord a => Int -> [(Set.Set a, v)] -> Set.Set a -> [(Set.Set a, v)] nearest_k k bs b = take k bs' where bs' = sortOn (hamming b) bs hamming :: Ord a => Set.Set a -> (Set.Set a, v) -> Int hamming x (y, _) = hamming_distance x y hamming_distance :: Ord a => Set.Set a -> Set.Set a -> Int hamming_distance xs ys = Set.size (Set.difference xs ys) + Set.size (Set.difference ys xs) subsets :: [a] -> [[a]] subsets [] = [[]] subsets (x:xs) = subsets xs ++ map (x:) (subsets xs) int_lists :: [[Int]] int_lists = subsets [1..20] values :: [(Set.Set Int, Int)] values = map f (zip [1..] int_lists) where f (i, x) = (Set.fromList x, i) test = nearest_k 8 values (Set.fromList [1,2,3]) ---- This works ok for the test above (with sets of ints), but is rather slow in my actual application (in which the sets are large sets of ground atoms of first-order logic). Is there some major optimization I should be doing here? thanks, Richard

3 5

Help evaluating TQueue variants
by David Feuer 25 May '18

25 May '18

I've written three related variations on TQueue (from the stm package). All three are designed to avoid the reader starvation problem described in GHC Trac 9539 [1] while limiting contention. Two are based on Okasaki's scheduled (i.e., real-time) banker's queues, while the third is based on his amortized-time banker's queue. I am hoping one or more people with programs that use TQueue heavily would be willing and able to try out these alternatives and see how they perform in their applications. I'm interested in how they compare to the current TQueue implementation, and also how they compare to each other. The queues are currently in a GitHub repository [2], but if it would be helpful I could put them up on Hackage. Thank you for your time. [1] https://ghc.haskell.org/trac/ghc/ticket/9539 [2] https://github.com/treeowl/stm-queues

1 0