- Haskell-Cafe - Haskell.org

Interruptible exception wormholes kill modularity
by Edward Z. Yang 04 Jul '16

04 Jul '16

In 2010, in the thread "Asynchronous exception wormholes kill modularity" [1], Bas van Dijk observed that 'unblock :: IO a -> IO a' broke modularity, as the sequence of calls 'block . block . unblock $ io' would result in 'io' being run with asynchronous exceptions unblocked, despite the outer 'block' "expecting" that asynchronous exceptions cannot be thrown. I would like to make two claims: 1. The new mask/restore interface is insufficient to "solve" this modularity problem, as *interruptible* operations can still be used to catch asynchronous exceptions. 2. Thus, we should provide an unblock combinator which can be used to catch asynchronous exceptions from a 'mask' (though not an 'uninterruptibleMask')--though it is doubtful if anyone should ever use 'mask' in the first place. Claim 1: Here is some code which reimplements 'unblock': import Control.Exception import Control.Concurrent import Control.Concurrent.MVar unblock :: IO a -> IO a unblock io = do m <- newEmptyMVar _ <- forkIO (io >>= putMVar m) takeMVar m The main idea is that 'takeMVar' is an interruptible operation: when it blocks, the thread can now receive asynchronous exceptions. In general, a thread can unmask exceptions by blocking. Here is a simple test-case: main = do let x = 10000000 -- Just do a bit of work tid <- myThreadId forkIO $ (threadDelay 10000 >> killThread tid) r <- mask $ \restore -> do -- restore $ do -- unblock $ do -- do something non-blocking evaluate (f x []) -- If the exception is delivered in a timely manner, -- shouldn't get here. print r f 0 r = r f n r = f (n-1) (n:r) With both restore and unblock commented, the ThreadKilled exception is delayed; uncommenting either restore or unblock causes the exception to be delivered. This admonition does not apply to uninterruptibleMask, for which there are no interruptible exceptions. Claim 2: Thus, I come to the conclusion that we were wrong to remove 'unblock', and that it is no worse than the ability for interruptible actions to catch asynchronous exceptions. You could very well argue that interruptible actions are a design flaw. Then you should use 'uninterruptibleMask' instead, which effectively removes the concept of interruptibility--and is thus modular. Indeed, Eyal Lotem proposed [2] that 'bracket' should instead use 'uninterruptibleMask', for precisely the reason that it is too easy to reenable asynchronous exceptions in 'mask'. But assuming that interruptible masks are a good idea (Simon Marlow has defended them as "a way avoid reasoning about asynchronous exceptions except at specific points, i.e., where you might block"), there should be an 'unblock' for this type of mask. It should be said that the absence of 'unblock' for 'uninterruptibleMask' only implies that a passed in IO action (e.g., the cleanup action in bracket) does not have access to the exceptions thrown to the current thread; it doesn't actually guarantee uninterruptibility, since the passed in IO action could always raise a normal exception. Haskell's type system is not up to the task of enforcing such invariants. Cheers, Edward [1] https://mail.haskell.org/pipermail/libraries/2010-March/013310.html https://mail.haskell.org/pipermail/libraries/2010-April/013420.html [2] https://mail.haskell.org/pipermail/libraries/2014-September/023675.html P.S. You were CC'ed to this mail because you participated in the original "Asynchronous exception wormholes kill modularity" discussion. P.P.S. I have some speculations about using uninterruptibleMask more frequently: it seems to me that there ought to be a variant of uninterruptibleMask that immediately raises an exception if the "uninterruptible" action blocks. This would probably of great assistance of noticing and eliminating blocking in uninterruptible code.

3 6

How to make Repa do stencils unboxed?
by William Yager 04 Jul '16

04 Jul '16

Hello all, TL;DR: Repa appears to be calculating my stencil operation using boxed arithmetic, which is very slow. How do I make it do unboxed arithmetic? I am currently working on a project where I need to simulate a 2D environment that exhibits wave behavior. I'm accomplishing this by directly discretizing the wave equation and turning it into a cellular automaton. The math here is relatively simple. Let's specify the example to the physical system of a rubber sheet, which exhibits wave behavior. To simulate a rubber sheet, you divide it up into grids. Then, you keep track of the velocity and height of each grid. At each iteration of the simulation, you update the velocity of each grid based on the force exerted upon it by its neighboring grids. This is simply a linear multiple of the distance between the grid and its neighbors (i.e. how stretched the sheet is). For example, if cell x is at height X, and its neighbors a,b,c,d are at height A,B,C,D, the tension on X is, for some constant L, T = L*((A-X)+(B-X)+...) = L*(A+B+C+D-4X). You can calculate this efficiently via a matrix convolution, in particular by a convolution with the matrix [[ 0 1 0 1 -4 1 0 1 0 ]] You can be more physically accurate by including the diagonals (weighted down a bit). A popular stencil for such purposes is [[ 1 2 1 2 -12 2 1 2 1 ]] Technically the off-diagonal elements should be something like sqrt(2), but this is sufficient for most purposes. After calculating the tension matrix, you divide it by the mass of each grid. This is the acceleration on each grid. Multiply it by the length of your time step, and you get the change in velocity of each grid. Add this to the velocity matrix. Then, multiply the velocity matrix by the time step, and that's the change in height. Add this to the height matrix. Boom, you've simulated a single time step of a rubber sheet. Keep doing this over and over to get the time-evolution of the rubber sheet. My code to do this is as follows: https://gist.github.com/wyager/8c468c9d18ad62aff8bc9738aa947ea4 The imports are messy; sorry about that. You'll notice I'm using the Dimensional library, which provides statically typed physical dimensions over any numeric value using a newtype wrapper with typelits. Hopefully this isn't slowing things down somehow. According to profiling via "ghc -Odph -rtsopts -threaded -fno-liberate-case -funfolding-use-threshold1000 -funfolding-keeness-factor1000 -optlo-O3 -fforce-recomp -prof -fprof-auto -fexternal-interpreter RepaSim.hs" (whew), fully 80% of time and 84% of allocation comes from the stencil operation. The simulation starts with a rubber sheet with a smooth bump in the middle, simulates 100ms of time steps, saves a .png file of the sheet, and then repeats n times (n is typed in via standard input). How can I make Repa do this stencil unboxed, and are there other things I should be doing to get better performance? Thanks, Will

2 4

Octree and Bounding Boxes
by Michael Litchard 04 Jul '16

04 Jul '16

I've forked the octree library and added some bounding box functionality. The problem is, it turns out kd-trees are better for my use case. I'd rather not junk the code, but I can't see a use case for the octree library as it stands. Below is the link to the branch where the code is, is there a use case for point octrees needing bounding boxes? https://github.com/mlitchard/octree/tree/test-reorg/Data/Octree/BoundingBox

1 0

Re: [Haskell-cafe] Haskell-Cafe Digest, Vol 154, Issue 27
by Vlad Lopatin 03 Jul '16

03 Jul '16

Re: up-to-date glfw examples (briand(a)aracnet.com) Hey, Brian Here's a bare-bones example of modern OpenGL(4.5) (VBO's, vertex/fragment shaders), using GLFW-b to handle windows/input and Yampa for animation. Shader-loading is handled by Sven Panne's code: https://github.com/madjestic/Haskell-OpenGL-Tutorial/tree/master/Mandelbrot… Cheers, Vlad

2 1

haskell-browser-mods - v0.1
by Erik Rantapaa 03 Jul '16

03 Jul '16

Hi everyone, I've created a browser extension to aid in navigating around Haskell-related websites. The extension adds information to visited pages, create links to other relevent pages and adds keyboard shortcuts and conveniences. The sites the extension knows about include: - haskellnews.org - hdiff.luite.com - hackage.haskell.org - packdeps.haskellers.com - hayoo.fh-wedel.de The extension is currently available for Safari and Chrome. For more information and downloads visit: https://github.com/erantapaa/haskell-browser-mods Feedback, suggestions, bug reports and PRs are very welcome. Erik

1 0

type level Min function for GHC.TypeLits?
by Baojun Wang 02 Jul '16

02 Jul '16

Hi Cafe, I'm trying to do some basic type level tricks, I wonder whether it is possible to get ``vZipWith`` in below code snippet working: {-# LANGUAGE DataKinds #-} {-# LANGUAGE PolyKinds #-} {-# LANGUAGE KindSignatures #-} {-# LANGUAGE TypeOperators #-} {-# LANGUAGE TypeFamilies #-} {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE RankNTypes #-} {-# LANGUAGE GADTs #-} {-# LANGUAGE StandaloneDeriving #-} {-# LANGUAGE UndecidableInstances #-} {-# OPTIONS_GHC -fplugin GHC.TypeLits.Normalise #-} import GHC.TypeLits import Data.Proxy data Vect :: (Nat -> * -> *) where VNil :: Vect 0 a VCons :: a -> Vect n a -> Vect (1+n) a deriving instance (Show a) => Show (Vect n a) vlength :: forall n a . KnownNat n => Vect n a -> Integer vlength _ = natVal (Proxy :: Proxy n) vconcat :: Vect n a -> Vect m a -> Vect (n+m) a -- Need type checker plugin vconcat VNil ys = ys vconcat xs VNil = xs vconcat (VCons x xs) ys = VCons x (vconcat xs ys) type family Min3 (t :: Bool) (m :: Nat) (n :: Nat) :: Nat type instance Min3 'True m n = m type instance Min3 'False m n = n type Min m n = Min3 (m <=? n) m n vZipWith :: (a -> b -> c) -> Vect m a -> Vect n b -> Vect (Min m n) c vZipWith f VNil _ = VNil vZipWith f _ VNil = VNil vZipWith f (VCons x xs) (VCons y ys) = f x y `VCons` vZipWith f xs ys As I got errors: /home/wangbj/downloads/stack/vect/src/Vect.hs:51:21: Could not deduce (Min3 (m <=? 0) m 0 ~ 0) from the context (n ~ 0) bound by a pattern with constructor VNil :: forall a. Vect 0 a, in an equation for ‘vZipWith’ at /home/wangbj/downloads/stack/vect/src/Vect.hs:51:14-17 Expected type: Vect (Min m n) c Actual type: Vect 0 c Relevant bindings include vZipWith :: (a -> b -> c) -> Vect m a -> Vect n b -> Vect (Min m n) c (bound at /home/wangbj/downloads/stack/vect/src/Vect.hs:50:1) In the expression: VNil In an equation for ‘vZipWith’: vZipWith f _ VNil = VNil /home/wangbj/downloads/stack/vect/src/Vect.hs:52:40: Could not deduce (Min3 (m <=? n) m n ~ (1 + Min3 (n1 <=? n2) n1 n2)) from the context (m ~ (1 + n1)) bound by a pattern with constructor VCons :: forall a (n :: Nat). a -> Vect n a -> Vect (1 + n) a, in an equation for ‘vZipWith’ at /home/wangbj/downloads/stack/vect/src/Vect.hs:52:13-22 or from (n ~ (1 + n2)) bound by a pattern with constructor VCons :: forall a (n :: Nat). a -> Vect n a -> Vect (1 + n) a, in an equation for ‘vZipWith’ at /home/wangbj/downloads/stack/vect/src/Vect.hs:52:26-35 Expected type: Vect (Min m n) c Actual type: Vect (1 + Min n1 n2) c Relevant bindings include ys :: Vect n2 b (bound at /home/wangbj/downloads/stack/vect/src/Vect.hs:52:34) xs :: Vect n1 a (bound at /home/wangbj/downloads/stack/vect/src/Vect.hs:52:21) vZipWith :: (a -> b -> c) -> Vect m a -> Vect n b -> Vect (Min m n) c (bound at /home/wangbj/downloads/stack/vect/src/Vect.hs:50:1) In the expression: f x y `VCons` vZipWith f xs ys In an equation for ‘vZipWith’: vZipWith f (VCons x xs) (VCons y ys) = f x y `VCons` vZipWith f xs ys Failed, modules loaded: none. Also I tried https://github.com/yav/type-nat-solver but got error like: *** Exception: user error (Unexpected result from the SMT solver: Expected: success Result: (error "line 57 column 16: unknown constant tn_0" ) Which I have no idea what it really means. Is there any way to get ``vZipWith`` working without too much hassle? Thanks baojun

1 0

Type-class predicates head-normal forms
by sheng chen 01 Jul '16

01 Jul '16

Hello, I got a question about what the form of types in Haskell are allowed. Specifically, what are the requirements of the context of a type? The paper Typing Haskell in Haskell required the type to be in the form P => t, where P to be in head-normal form or in other words shouldn't be something like Num Bool. Now what about multi-parameter classes? The paper Type classes: an exploration of the design space listed several potential choices but didn't say exactly what. The Haskell 2010 report covered single-parameter type classes only. Given the following definition. class Collects ce e | ce -> e where empty :: ce insert :: e -> ce -> ce insert2 c = insert True (insert False c) The function insert2 has the following type insert2 :: Collects ce Bool => ce -> ce This shows that the definition of head-normal forms for multi-parameter class differs from that for single-parameter classes. Is there any definite specification about the context of Haskell types? Thank you, Sheng

2 1

ICFP 2016 Student Research Competition: Call for Submissions
by Lindsey Kuper 01 Jul '16

01 Jul '16

====================================================================== CALL FOR SUBMISSIONS SRC at ICFP 2016 Nara, Japan 18-24 September 2016 http://conf.researchr.org/info/icfp-2016/student-research-competition Co-located with the International Conference on Functional Programming (ICFP 2016) ====================================================================== Student Research Competition ---------------------------- This year ICFP will host a Student Research Competition where undergraduate and postgraduate students can present posters. The SRC at ICFP 2016 consists of three rounds: * Extended abstract round. All students are encouraged to submit an extended abstract of up to 800 words outlining their research. * Poster session. Based on the abstracts, a panel of judges will select the most promising entrants to participate in the poster session which will take place at ICFP. Students who make it to this round will be eligible for some travel support to attend the conference. In the poster session, students will have the opportunity to present their work to the judges, who will select three finalists in each category (graduate/undergraduate) to advance to the next round. * ICFP presentation. The last round will consist of an oral presentation at ICFP to compete for the final award. Prizes ------ * The top three graduate and the top three undergraduate winners will receive prizes of $500, $300, and $200, respectively. * All six winners will receive award medals and a two-year complimentary ACM student membership, including a subscription to ACM's Digital Library. * The names of the winners will be posted on the ACM SRC web site. * The first-place winners will be invited to participate in the ACM SRC Grand Finals, an on-line round of competition among the winners of conference-hosted SRCs. * Grand Finalists and their advisors will be invited to the Annual ACM Awards Banquet for an all-expenses-paid trip, where they will be recognized for their accomplishments along with other prestigious ACM award winners, including the winner of the Turing Award (also known as the Nobel Prize of Computing). * The top three graduate Grand Finalists will receive an additional $500, $300, and $200. Likewise, the top three undergraduate Grand Finalists will receive an additional $500, $300, and $200. All six Grand Finalists will receive Grand Finalist certificates. * The ACM, Microsoft Research, and our industrial partners provide financial support for students attending the SRC. You can find more information about this on the ACM website. Eligibility ----------- The SRC is open to both undergraduate (not in a PhD programme) and graduate students (in a PhD programme). Upon submission, entrants must be enrolled as a student at their universities, and are ACM student members. Furthermore, there are some constraints on what kind of work may be submitted. Previously published work: Submissions should consist of original work (not yet accepted for publication). If the work is a continuation of previously published work, the submission should focus on the contribution over what has already been published. We encourage students to see this as an opportunity to get early feedback and exposure for the work they plan to submit to the next ICFP or POPL. Collaborative work: Students are encouraged to submit work they have been conducting in collaboration with others, including advisors, internship mentors, or other students. However, submissions are individual, so they must focus on the contributions of the student. Submission Details ------------------ Each submission should include the student author's name, institutional affiliation, e-mail address, and postal address; research advisor's name; ACM student member number; category (undergraduate or graduate); research title; and an extended abstract addressing the following: * Problem and Motivation: Clearly state the problem being addressed and explain the reasons for seeking a solution to this problem. * Background and Related Work: Describe the specialized (but pertinent) background necessary to appreciate the work. Include references to the literature where appropriate, and briefly explain where your work departs from that done by others. * Approach and Uniqueness: Describe your approach in attacking the problem and clearly state how your approach is novel. * Results and Contributions: Clearly show how the results of your work contribute to computer science and explain the significance of those results. The abstract must describe the student's individual research and must be authored solely by the student. If the work is collaborative with others and/or part of a larger group project, the abstract should make clear what the student's role was and should focus on that portion of the work. The extended abstract must not exceed 800 words and must not be longer than 2 pages. The reference list does not count towards these limits. To submit an abstract, please register through the submission page and follow the instructions. Abstracts submitted after the deadline may be considered at the committee's discretion, but only after decisions have been made on all abstracts submitted before the deadline. If you have any problems, don't hesitate to contact the competition chair. Please submit your abstract at the EasyChair submission page: https://easychair.org/conferences/?conf=icfp2016src Important Dates --------------- * Deadline for submission: 3 August * Notification of acceptance: 10 August Selection Committee ------------------- To include: Chair: David Van Horn, University of Maryland Ronald Garcia, University of British Columbia Ilya Sergey, University College London

1 0

Deadline extended: Commercial Users of Functional Programming Call For Tutorials
by Lindsey Kuper 01 Jul '16

01 Jul '16

The 2016 Commercial Users of Functional Programming (CUFP) call for tutorials have passed the initial deadlines, but we have more seat for tutors. CUFP tutorial is a good opportunity if you have some idea you want to spread to the real world, say, your favorite programming language (Clojure, Erlang, F#, F*, Haskell, ML, OCaml, Scala, Scheme...), concepts (e.g. Lens, Liquid-Haskell, Proof Assistance, ...), applications (Web, high-performance computing, finance...,) tools and techniques (QuickCheck, Elm, ...), or theories. If you are interested, we kindly invite you to submit a proposal for the Call For Tutorials from here: http://cufp.org/2016/call-for-tutorials.html , before July 3rd. If you have any questions, please do not hesitate to contact us: Roman Gonzalez: roman(a)unbounce.com Takayuki Muranushi: muranushi(a)gmail.com best regards, Takayuki

1 0

Compose :: Conference, Melbourne, 2016, Call for Presentations!
by Noon van der Silk 01 Jul '16

01 Jul '16

(Web version: http://www.composeconference.org/2016-melbourne/cfp/) # Call for Presentations Compose Melbourne - http://www.composeconference.org/ - is pleased to announce its call for presentations. Compose Melbourne is a new functional programming conference focused on developing the community and bringing typed functional programming to a wider audience. It is a 2-day event being held in Melbourne, Australia on the 29th and 30th of August 2016. The first day features a single track of presentations followed by a second day of workshops and an unconference. It is the new sister-conference of the NY based Compose Conference. Submit your presentation proposal here: https://easychair.org/conferences/?conf=cmc2016 # Important Dates - CFP Launch Party - 19th May - CFP Opens - 19th May - CFP Closes - June 30th - Notify Presenters - July 15th - Conference Day 1: Presentations - Monday, 29-Aug-2016 - Conference Day 2: Unconference - Tuesday, 30-Aug-2016 # Talk and Workshop Submission ## Day 1 - Presentations Submit a Presentation Proposal via Easy Chair: https://easychair.org/conferences/?conf=cmc2016. ## Day 2 - Unconference and Workshops Contact us (composemel-admin(a)googlegroups.com) with your workshop proposals. # Submission Guidelines Talk slots will be 30 minutes: 25 minute talk and 5 minutes for questions. Provide sufficient detail in your submission to enable the reviewers to understand your proposal and clearly identify what an attendee will gain from attending your session. You should include the following information in the submission system: - Authors - The details of each of the presenters - Title - Title of the presentation (please keep it brief and specific) - Abstract - A 300-500 word description of your presentation, ideally including... - Description of the content - Goal - Keywords - List any keywords that will help the program committee and attendees categorise your presentation. Also indicate the functional programming languages that you will be targeting. Nothing upsets an audience more than a speaker that stands on stage blatantly promoting his or her company, product or achievements. Please keep the content of your talk on-topic and do not use this speaking opportunity as a sales pitch. # Feel We want Compose Melbourne to be all about the community. We're aiming to help foster the growth of Functional Programming in Melbourne and unite all interested parties to spark a unified presence and a feeling of camaraderie amongst FP and Theory proponents in this wonderful city! # Audience Functional-Programming and Programming-Language-Theory professionals and enthusiasts. Newcomers, experts, anyone from other disciplines of fields who is interested in what FP is or how it could help them with their work, or simply make life more enjoyable! # Diversity Just like Compose-Conference, we would also like to put an emphasis on soliciting a diverse set of speakers - anything you can do to distribute information about this CFP and encourage submissions from under-represented groups would be greatly appreciated. We welcome *all* (new and established) contributions and encourage you to apply! # Day One All topics for proposals for presentations are invited, but not limited to explore the following topics: ## New Languages The development of new and emerging Functional Languages, and the toolsets around them. ## Libraries and Tools Exploring the use and development of new and underrepresented tools and libraries. ## Production Systems 2016 has seen many previously undeployed languages and systems break through into production. We're looking for war-stories, success-stories and sob-stories! ## Theory Theory is the cutting edge of new functional programming. Show the world what is rising over the horizon. ## Art and Music Exciting and innovative usage of functional programming language in the arts! # Day Two The second day will be dedicated to workshops and unconferences. ## Workshops We will be running a beginners Haskell workshop. If you wish to run a workshop on the second day of the conference, then please let us know. Alternatively, if there is a workshop that would interest you as an attendee, then please also let us know. ## Unconference Got something random you want to talk about? The unconference will run the second day. There will be a whiteboard with spaces where anyone can write their name down to speak. There's no need to bring anything but yourself and your ideas and speaking voice. # Sponsors We are seeking sponsors to support us in putting on this conference. If you or your company would like to sponsor us, please get in contact via composemel-admin(a)googlegroups.com. You can find our prospectus here: https://github.com/composeconference/Compose-Melbourne/blob/master/ComposeM… . -- Compose Conference Melbourne Organising Committee http://www.composeconference.org/

1 1