June 2018 - Haskell-Cafe

Call for Presentations and Demonstrations: Workshop on Functional High-Performance Computing 2018 at ICFP
by Kei Davis 26 Jun '18

26 Jun '18

7th ACM SIGPLAN Workshop on Functional High-Performance Computing September 29, 2018, St. Louis, Missouri, USA https://icfp18.sigplan.org/track/FHPC-2018-papers The 7th ACM SIGPLAN Workshop on Functional High-Performance Computing (FHPC 2018) is being held as in previous years in conjunction with the International Conference on Functional Programming (ICFP 2018) together with numerous other workshops/symposia, and as a first, colocated with Strange Loop, in St. Louis, MO, USA. Workshop Objectives The FHPC 2018 workshop seeks to bring together researchers and practitioners exploring uses of functional (or more generally, declarative or high-level) programming systems or concepts in application domains where high performance is essential. The aim of the meeting is to enable sharing of results, experiences, and novel ideas about how high-level, declarative specifications of computationally challenging problems can serve as maintainable and portable code that approaches (or even exceeds) the performance of machine-oriented (low-level) imperative implementations. All aspects of performance-critical programming and parallel programming are in scope for the workshop, irrespective of hardware target. This includes both traditional large-scale distributed-memory scientific computing (HPC), as well as work targeting single node systems with SMPs, GPUs, FPGAs, or embedded processors. FHPC 2018 seeks to encourage a range of submissions, focusing on work in progress and facilitating early exchange of ideas and open discussion on innovative and/or emerging results. Original research, experience reports, case studies, and evaluations of programming systems are all welcome. Work on incorporation of functional programming concepts into more traditional (imperative) HPC applications is explicitly solicited. Papers, Presentations, and Demonstrations The refereed paper deadline has passed. In the workshop spirit, the refereed paper presentation program is being augmented to include less formally reviewed presentations in the form of talks not based on accepted papers, and software demonstrations. Topic areas of interest include research or development in progress, experience reports, and position/white paper statements. Proposals will be subject to a mild reviewing process to insure relevance and general interest to FHPC. We expect time slots to be between one-half and one hour. Prospective presenters are requested to submit an abstract describing their proposed contribution via the paper submission site. The deadline is July 29, 2018, midnight anywhere on Earth, but the PC reserves the right to accept proposals at any time after submission. As such, early submission is strongly encouraged. Important Dates - Submission Deadline: Midnight July 29, 2018 Anywhere on Earth. - Notification: early-mid August (TBD) - FHPC 2018: Saturday Sept. 29, 2018. Official website, submissions, registration details https://icfp18.sigplan.org/track/FHPC-2018-papers Previous FHPC websites https://icfp17.sigplan.org/track/FHPC-2017-papers https://sites.google.com/site/fhpcworkshops/ ICFP and related workshops https://conf.researchr.org/home/icfp-2018 https://icfp18.sigplan.org Strange Loop https://www.thestrangeloop.com/ Questions or comments? fhpc18(a)gmail.com

1 0

ANN: network-2.7.0.1
by evan＠evan-borden.com 24 Jun '18

24 Jun '18

Announcing network-2.7.0.1 * A new API: socketPortSafe. [#319](https://github.com/haskell/network/pull/319) * Fixing a drain bug of sendAll. [#320](https://github.com/haskell/network/pull/320) * Porting the new CALLCONV convention from master. [#313](https://github.com/haskell/network/pull/313) * Withdrawing the deprecations of packFamily and unpackFamily. [#324](https://github.com/haskell/network/pull/324) Thank you to network contributors for bug reports in the push to 3.0.0.0. As well as the continued effort to increase windows reliability and support. -- Evan Borden

1 0

Munich Haskell Meeting, 2018-06-26 @ 19:30
by Leah Neukirchen 24 Jun '18

24 Jun '18

Dear all, Next week, our monthly Munich Haskell Meeting will take place again on Tuesday, June 26 at Cafe Puck at 19h30. 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-jun-2018/ Everybody is welcome! cu, -- Leah Neukirchen <leah(a)vuxu.org> http://leah.zone

1 0

Colored Haskell Listings in LaTeX
by Artem Pelenitsyn 23 Jun '18

23 Jun '18

Dear Cafe, In his recent Stitch manuscript, https://cs.brynmawr.edu/~rae/papers/2018/stitch/stitch.pdf Richard Eisenberg mentions that he uses lhs2TeX to typeset Haskell listings, but I'm not aware of the support for colors in lhs2TeX. Can anyone suggest how to get such a nice code highlighting (presumably with lhs2TeX). -- Best wishes, Artem

3 4

Type level seq
by Benjamin Franksen 23 Jun '18

23 Jun '18

I have observed that type functions seem to be lazy. A type function that is partial does not give me a type error when it is called with an argument for which it is not defined. Is there a 'seq' at the type level? Here is my use case, simplified of course. Suppose we want to statically track the state of a switch with a phantom type parameter: > {-# LANGUAGE DataKinds, TypeFamilies #-} > > data State (s :: Bool) = State Bool deriving Show > > off :: State False > off = State False We want user code to be able to call the function 'turnOff' only if the switch is currently on: > turnOff :: State True -> State False > turnOff (State True) = State False This works fine: *Main> turnOff off <interactive>:1:9: error: ? Couldn't match type ?'False? with ?'True? Expected type: State 'True Actual type: State 'False But now I want to abstract this pattern and write a (closed) type function > type family TurnOff s where > TurnOff True = False > turnOff' :: State x -> State (TurnOff x) > turnOff' (State True) = State False > bad = turnOff' off *Main> :t bad bad :: State (TurnOff 'False) *Main> bad *** Exception: TypeLevelSeq.lhs:37:3-37: Non-exhaustive patterns in function turnOff'

4 4

Call for a maintainer of protocol-buffers
by Kostiantyn Rybnikov 22 Jun '18

22 Jun '18

Hey everyone, I'm looking for a person who would become the major maintainer of the protocol-buffers package. Details here: https://github.com/k-bx/protocol-buffers/issues/64 Please comment on the issue or send me an email if you'd like to apply. Thank you

1 0

dfinity-radix-tree: A Merkleized key–value data store.
by Enzo 22 Jun '18

22 Jun '18

I am pleased to announce the first release of dfinity-radix-tree: A Merkleized key–value data store. The library provides a simple and easy to use data integrity layer for LevelDB. For more information, please see: [1] https://github.com/dfinity-lab/hs-radix-tree [2] https://hackage.haskell.org/package/dfinity-radix-tree Warm regards, Enzo Haussecker DFINITY Stiftung *We're hiring Haskell developers in Palo Alto and Zürich. Apply now:* [3] https://dfinity.org/openings

1 0

Timeouts inside a ConduitParser
by Luke Lau 21 Jun '18

21 Jun '18

I have a ConduitParser <https://hackage.haskell.org/package/conduit-parse-0.2.1.0/docs/Data-Conduit…> (a Sink with some parsing state) with a version of satisfy that can time out: satisfy :: MonadIO m => (a -> Bool) -> ConduitParser a m a satisfy pred = do tId <- liftIO myThreadId timeoutThread <- liftIO $ forkIO $ do threadDelay 1000000 throwTo tId TimeoutException x <- await liftIO $ killThread timeoutThread if pred x then return x else empty However I would rather not deal with the risks involved with handling concurrency myself and use a system library like System.Timeout: satisfy :: MonadIO m => (a -> Bool) -> ConduitParser a m a satisfy pred = do x <- timeout 1000000 await if pred x then return x else empty This doesn’t work though since I need to be able to both lift and unlift await from IO, and ConduitParser lies on top of ConduitT, which is one of the types of monads that UnliftIO cannot be an instance of <https://github.com/fpco/unliftio#limitations>. Are there any better approaches to this? <https://github.com/fpco/unliftio#limitations> <https://github.com/fpco/unliftio#limitations> <https://github.com/fpco/unliftio#limitations>

2 1

Re: [Haskell-cafe] How to use Reader?
by Hilco Wijbenga 21 Jun '18

21 Jun '18

Ah, quite simple. Thanks! On Tue, Jun 19, 2018 at 9:39 PM, Matt <parsonsmatt(a)gmail.com> wrote: > showText text = do > a <- showA > return ("[" `append` text `append` a `append` "]") > > > Matt Parsons > > On Tue, Jun 19, 2018 at 10:35 PM, Hilco Wijbenga <hilco.wijbenga(a)gmail.com> > wrote: >> >> Hi all, >> >> I think I'm pretty close but not quite there yet. Let's say I have the >> following: >> >> {-# LANGUAGE OverloadedStrings #-} >> >> module Main where >> >> import Prelude (IO) >> import Data.Text (Text, append) >> import Data.Text.IO (putStrLn) >> >> main :: IO () >> main = >> putStrLn (showText "Hello World!" config) >> where >> config :: Config >> config = Config (A "Boo") >> >> data A = A Text >> >> data Config = Config { a :: A } >> >> showText :: Text -> Config -> Text >> showText text config = >> "[" `append` text `append` (showA config) `append` "]" >> >> showA :: Config -> Text >> showA config = >> "'" `append` text `append` "'" >> where >> A text = a config >> >> Now I want to use Reader. This is as close as I can get: >> >> {-# LANGUAGE OverloadedStrings #-} >> >> module Main where >> >> import Prelude (IO, return, ($)) >> import Data.Text (Text, append) >> import Data.Text.IO (putStrLn) >> import Control.Monad.Reader (Reader, asks, runReader) >> >> main :: IO () >> main = >> putStrLn (runReader (showText "Hello World!") config) >> where >> config :: Config >> config = Config (A "Boo") >> >> data A = A Text >> >> data Config = Config { a :: A } >> >> showText :: Text -> Reader Config Text >> showText text = >> return $ >> "[" `append` text `append` showA `append` "]" >> >> showA :: Reader Config Text >> showA = do >> A text <- asks a >> return $ >> "'" `append` text `append` "'" >> >> This is almost correct, except for the showA invocation in showText. >> It wants a Text but it's getting a Reader Config Text. What is the >> magic to make this work? >> >> Cheers, >> Hilco >> _______________________________________________ >> Haskell-Cafe mailing list >> To (un)subscribe, modify options or view archives go to: >> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe >> Only members subscribed via the mailman list are allowed to post. > >

1 1

Looking for simple C statements parser for Parsec, MegaParsec or Happy
by Bulat Ziganshin 21 Jun '18

21 Jun '18

Hello all, I develop a preprocessor transpiling subset of C statements into gcc _asm statements. For this project, I will be happy to reuse existing C statements parser written with any popular Haskell technology, or just start with some simple C subset parser in order to avoid redoing existing work. Unfortunately, so far I found grammars just for everything but C. And while I can start with Parsec grammars for Java/Go, it seems that MegaParsec now is better choice? I can quickly develop grammar for small C subset, but ready-to-use grammar for larger C subset will allow me to completely skip the development of C statement parser and focus on the meat of project - asm code generation. Eventually, it may turn into LLVM pass transpiling parts of C++ code, but for quick prototype, I prefer Haskell, especially if I can find ready-to-use parser. PS: It was the first time I ever asked at SO, but it turned out to be offtopic there: https://stackoverflow.com/questions/50949118/simple-c-grammar-for-haskell-p… -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com

2 2