- Beginners - Haskell.org

Re: [Haskell-beginners] Is it possible such Monad? (Baa)
by 鲍凯文 26 Oct '17

26 Oct '17

Hi, 1. I think your data type is isomorphic to data Allpass w m a = Pass Bool (m a) w 2. Maybe first try to write a Functor instance? Then you can find out what constraints are necessary for 'w' and 'm' to write it. 3. What does this data type intend to represent? Best, toz On Wed, Oct 25, 2017 at 5:32 AM <beginners-request(a)haskell.org> wrote: > Send Beginners mailing list submissions to > beginners(a)haskell.org > > To subscribe or unsubscribe via the World Wide Web, visit > http://mail.haskell.org/cgi-bin/mailman/listinfo/beginners > or, via email, send a message with subject or body 'help' to > beginners-request(a)haskell.org > > You can reach the person managing the list at > beginners-owner(a)haskell.org > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Beginners digest..." > > > Today's Topics: > > 1. Is it possible such Monad? (Baa) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Wed, 25 Oct 2017 13:56:31 +0300 > From: Baa <aquagnu(a)gmail.com> > To: The Haskell-Beginners Mailing List - Discussion of primarily > beginner-level topics related to Haskell <beginners(a)haskell.org> > Subject: [Haskell-beginners] Is it possible such Monad? > Message-ID: <20171025135631.1f9913d2@Pavel> > Content-Type: text/plain; charset=US-ASCII > > Hello All! > > Is it possible to write Monad for such type: > > data Allpass w m a = Nopass (m a) w | Allpass (m a) w > > I can write (>>=), IMHO such type can not be Monad due to `w` is not > under `m` monad, right? > > > === > Best regards, Paul > > > ------------------------------ > > Subject: Digest Footer > > _______________________________________________ > Beginners mailing list > Beginners(a)haskell.org > http://mail.haskell.org/cgi-bin/mailman/listinfo/beginners > > > ------------------------------ > > End of Beginners Digest, Vol 112, Issue 22 > ****************************************** >

2 1

Is it possible such Monad?
by Baa 25 Oct '17

25 Oct '17

Hello All! Is it possible to write Monad for such type: data Allpass w m a = Nopass (m a) w | Allpass (m a) w I can write (>>=), IMHO such type can not be Monad due to `w` is not under `m` monad, right? === Best regards, Paul

1 0

Manipulate list, one element at a time
by martin 24 Oct '17

24 Oct '17

Hello all, How can I do something to each element of a list, but only modify one element at at time, so the result is a list of lists, where each sublist is the original list with one element altered. Something like type Each a = (a->a) -> [a] -> [[a]] I came up with: each :: Each a each f [] = [] each f (x:xs) = (f x : xs) : (map (x:) $ each f xs) λ> each (*10) [1..3] [[10,2,3],[1,20,3],[1,2,30]] but I wonder if there is a more standard way of doing this

2 1

runState (liftM (+100) pop) [1, 2, 3, 4] (Example from LYH)
by Olumide 24 Oct '17

24 Oct '17

Dear List, Apologies for the awkward question title -- its the best I can do ATM. I'm still working my way through chapter 13 of LYH (title for a few monads more) and I came across the following two beasts ghci> runState (liftM (+100) pop) [1,2,3,4] (101,[2,3,4]) ghci> runState (fmap (+100) pop) [1,2,3,4] (101,[2,3,4]) See http://learnyouahaskell.com/for-a-few-monads-more#useful-monadic-functions Even though I'm still struggling to wrap my mind around monads, I sort of understand what's going on here. The problem is that I can't explain why the function (+100) is applied to _only_ the value 1 in (1,[2,3,4]). Regards, - Olumide

2 1

Where clause indentation
by Wink Saville 20 Oct '17

20 Oct '17

I created a file with the dividedBy example from Chapter 8.5 of "Haskell Programming from first principles" : dividedBy :: Integral a => a -> a -> (a, a) dividedBy num denom = go num denom 0 where go n d count | n < d = (count, n) | otherwise = go (n - d) d (count + 1) I get the following error when I load into ghci: $ ghci chapter8_5-IntegralDivision.hs GHCi, version 8.2.1: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/wink/.ghci [1 of 1] Compiling Main ( chapter8_5-IntegralDivision.hs, interpreted ) chapter8_5-IntegralDivision.hs:4:7: error: parse error (possibly incorrect indentation or mismatched brackets) | 4 | | n < d = (count, n) | ^ Failed, 0 modules loaded. λ> But if I put the "go" function on its own line: dividedBy :: Integral a => a -> a -> (a, a) dividedBy num denom = go num denom 0 where go n d count | n < d = (count, n) | otherwise = go (n - d) d (count + 1) It does compile: $ ghci chapter8_5-IntegralDivision.hs GHCi, version 8.2.1: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/wink/.ghci [1 of 1] Compiling IntegralDivision ( chapter8_5-IntegralDivision.hs, interpreted ) Ok, 1 module loaded. Or I can put the "where" on the previous line: dividedBy :: Integral a => a -> a -> (a, a) dividedBy num denom = go num denom 0 where go n d count | n < d = (count, n) | otherwise = go (n - d) d (count + 1) it also compiles: $ ghci chapter8_5-IntegralDivision.hs GHCi, version 8.2.1: http://www.haskell.org/ghc/ :? for help Loaded GHCi configuration from /home/wink/.ghci [1 of 1] Compiling Main ( chapter8_5-IntegralDivision.hs, interpreted ) Ok, 1 module loaded. λ> Can someone shed light on what I've done wrong? -- Wink

3 3

Hack “Importify” by Serokell During #Hacktoberfest!
by Gints Dreimanis 18 Oct '17

18 Oct '17

All of us love to read explicit import lists (especially in an unknown module), but we all hate to write and maintain them. >From our point of view, it is obligatory for any code that goes into production to have only explicit imports. That way, the code is easier to comprehend – no one is confused about the location of a stray type or identifier, even when reading from GitHub or a simple editor. There are, however, usability and maintainability issues with explicit imports, which make enforcing explicit imports in style guides a topic for discussion. To enjoy the benefits of explicit imports while keeping the drawbacks in check, we have implemented *Importify*. *Importify* is a tool developed by Serokell tooling team led by Dmitry Kovalnikov. It is a refactoring tool that helps import management in Haskell code. We’re getting close to releasing it on Hackage, but it requires some polishing. You can contribute to Importify <https://github.com/serokell/importify/issues?q=is%3Aissue+is%3Aopen+label%3…> and Serokell’s own custom Prelude Universum <https://github.com/serokell/universum/issues?q=is%3Aissue+is%3Aopen+label%3…> during Hacktoberfest <https://github.com/blog/2433-celebrate-open-source-this-october-with-hackto…> . *Importify* is all about decluttering the import blocks. It can remove: - unused symbols in explicit import lists - unused qualified imports - unused implicit imports - unnecessarily hidden symbols Highlights of the functionality of *Importify* are that it works deeply with explicit import lists (it can, for instance, remove unused constructors), and it works with hiding imports. The latter feature isn’t present in any open-source refactoring tool that is known to us. It is also compatible with autoexporter <https://hackage.haskell.org/package/autoexporter>. *Importify *works extremely well with large Haskell projects — we use it to great success to clean up Cardano SL <https://github.com/input-output-hk/cardano-sl/>. Importify only parses source code instead of building projects. For a project of the size of Cardano SL it takes only two minutes to build Importify cache while building it would take forty minutes on the same machine. In the future, we plan for *Importify *to be able to add used symbols to imports automatically. To speed up *Importify *runs, we will implement a cache server that caches each Hackage package. You will be able to upload caches for your projects there as well. Next on the roadmap is functionality to convert imports between *implicit* and *explicit*, and between *qualified* and *unqualified* forms. Finally, we will be adding functionality to resolve merge conflicts in import section automatically. If you notice that managing imports becomes a slog, feel free to try this tool out yourself. Spotted a bug? Contact us. <bugs(a)serokell.io> To those, who are in the mood for some hacking this Autumn, we’re looking forward to your #Hacktoberfest contributions!

1 0

Dispatch on implemented instances
by PY 18 Oct '17

18 Oct '17

Hello, everyone! I want to call one function (or the same but with a special argument, no matter) if a type instantiates some type-class or another - if not. I know that it's easy in languages like D, Java, possible C++ - we can specialize templates more generic or less generic and to restrict its parameters with class/interface implementation/extending (there are special keywords/type operators for it, used in generics/templates signatures). So, I suppose Haskell can do the same. But as it turned out, this is not so easy for Haskell :( I found these useful links: * https://stackoverflow.com/questions/26303353/can-multiple-dispatch-be-achie… * https://wiki.haskell.org/GHC/AdvancedOverlap So, I wrote something like this (this example is compiling and works as I expect, you need `mustache` and `network-uri` packages). Goal is to substitute type in special way for HTML template and for text template if type implements a special interface (for deterministic substitution). Otherwise default substitution will be used. Actually, this code allows usage of not only HTML/text templates... OK, here it is: {-# LANGUAGE EmptyDataDecls #-} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE FunctionalDependencies #-} {-# LANGUAGE MultiParamTypeClasses #-} {-# LANGUAGE OverloadedStrings #-} {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE TypeFamilies #-} {-# LANGUAGE UndecidableInstances #-} import Data.Semigroup ((<>)) import qualified Data.Text as T import Network.URI import qualified Text.Mustache.Types as MUT type Link = (T.Text, URI) -- caption and URL, for example -- |Converts a type to substitutable to Mustache template entity based on 'SubTo' template selector. -- @sub@ must be ignored in implementation class ToMustaches sub a where toMustaches :: sub -> a -> SubTo -> MUT.Value -- |Substituted to HTML either to text data SubTo = ToHtml | ToText -- |Deterministic substitution data DetSub -- |Universal substitution data UniSub -- |Substitution predicate: it's determined uniquely for type class SubPred a sub | a -> sub -- TODO rename ToMustaches' (ToMustaches' ?) class ToMustaches' a where toMustaches' :: a -> SubTo -> MUT.Value instance (SubPred a sub, ToMustaches sub a) => ToMustaches' a where toMustaches' a to = toMustaches (undefined::sub) a to -- See: -- * https://stackoverflow.com/questions/36913922/how-to-resolve-overlapping-ins… -- * https://stackoverflow.com/questions/26303353/can-multiple-dispatch-be-achie… instance {-# OVERLAPS #-} (sub ~ UniSub) => SubPred a sub -- 'Link' can be substituted deterministically instance SubPred Link DetSub instance ToMustaches DetSub Link where toMustaches _ (_, url) ToText = MUT.String $ T.pack $ show url toMustaches _ (cap, url) ToHtml = MUT.String ("<a href=\"" <> (T.pack $ show url) <> "\">" <> cap <> "</a>") -- |All other types which does not implements 'ToMustaches' default instance transforms them to string w/ 'Show' instance Show a => ToMustaches UniSub a where toMustaches _ a _ = MUT.String $ T.pack $ show a ------------------------------------------------------------------- main :: IO () main = do let uri = URI "http:" (Just $ URIAuth "" "wikipedia.org" "") "" "" "" lnk = ("wikipedia", uri)::Link txt = "Hello world"::T.Text n = 444::Int print $ toMustaches' lnk ToText print $ toMustaches' lnk ToHtml print $ toMustaches' txt ToHtml print $ toMustaches' n ToText I don't know am I understand correctly the Haskell solution but this looks close to explained in the Haskell Wiki (case 1). But in all cases I see problem: always I need not only to implement this special interface but to "say" that a type implements special (deterministic) substitution, see: instance SubPred Link DetSub This means that clients of this code must "implements" 2 instances: 1) of ToMustaches and 2) of SubPred (with DetSub as last type param). This is terrible and no such thing in other languages! So, my question is: is a simple way to avoid it? === Best regards, Paul

1 0

Multiple parameters vs anonymous syntax
by Wink Saville 17 Oct '17

17 Oct '17

I'm going through "Haskell Programming from first principles" and in section 7.3 Anonymous Functions there is an exercise on converting multiple parameters to anonymous functions, and it asks: 1. Which (two or more) of the following are equivalent? mTh1 x y z = x * y * z mTh2 x y = \z -> x * y * z mTh3 x = \y -> \z -> x * y * z mTh4 = \x -> \y -> \z -> x * y * z So I created a file, anon.hs (attached): module Anon where mTh1 x y z = x * y * z mTh2 x y = \z -> x * y * z mTh3 x = \y -> \z -> x * y * z mTh4 = \x -> \y -> \z -> x * y * z I load that into ghci and check the function types: $ ghci anon.hs GHCi, version 8.2.1: http://www.haskell.org/ghc/ :? for help [1 of 1] Compiling Anon ( anon.hs, interpreted ) Ok, 1 module loaded. *Anon> :t mTh1 mTh1 :: Num a => a -> a -> a -> a *Anon> :t mTh2 mTh2 :: Num a => a -> a -> a -> a *Anon> :t mTh3 mTh3 :: Num a => a -> a -> a -> a *Anon> :t mTh4 mTh4 :: Integer -> Integer -> Integer -> Integer Why is mTh4 different from the rest? On the flip side If I enter "mTh4 = \x -> \y -> \z -> x * y * z" directly in ghci command line then it has same type as the others: $ ghci GHCi, version 8.2.1: http://www.haskell.org/ghc/ :? for help Prelude> mTh4 = \x -> \y -> \z -> x * y * z Prelude> :t mTh4 mTh4 :: Num a => a -> a -> a -> a -- Wink

2 2

Safe TCP accept loop
by Sumit Raja 16 Oct '17

16 Oct '17

Hello, 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? Thanks Sumit

1 0

monad question
by mike h 14 Oct '17

14 Oct '17

I have cap :: String -> String cap = toUpper rev :: String -> String rev = reverse then I make tupled :: String -> (String, String) tupled = do r <- rev c <- cap return (r, c) and to be honest, yes it’s been a long day at work, and this is coding at home rather than coding (java) at work but I’m not sure how tupled works!!! My first shot was supplying a param s like this tupled :: String -> (String, String) tupled s = do r <- rev s c <- cap s return (r, c) which doesn’t compile. But how does the first version work? How does the string to be processed get into the rev and cap functions?? Thanks Mike

2 4