- Haskell-Cafe - Haskell.org

Need a backup GSOC to apply for? Last minute new proposal -- memory reuse analysis for GHC
by Ryan Newton 27 Mar '12

27 Mar '12

Hi potential GSOC'ers, If there are multiple students interested in the project you're applying to it's a good idea to put in more than one application. This is a project proposal that would focus on performance analysis -- in particular in reusing some of the tools for analyzing memory access patterns in the context of Haskell: http://hackage.haskell.org/trac/summer-of-code/ticket/1619 I think this project would yield pretty good bang for the buck because almost all the pieces are already there. They just need to be combined, polished, documented, and evangelized. Cheers, -Ryan

1 0

ANN: Latest package versions on Hackage as JSON/JSONP
by Simon Hengel 26 Mar '12

26 Mar '12

Hi, I provide a relation between /package name/ and /latest version/ on Hackage as both JSON and JSONP. http://www.typeful.net/~tbot/hackage/ This are static files, and they are regenerated whenever new packages are uploaded (with a delay of about one minute). Two simple usage examples are given in the README[1]. Feel free to access this information for whatever purpose! Cheers, Simon [1] https://github.com/sol/hackage-jsonp#readme

2 1

difficulty building (some) TH packages with ghc-7.4.1
by John Lato 26 Mar '12

26 Mar '12

Hello, I've run into an odd problem when building certain packages that use Template Haskell with GHC-7.4.1. For example, RepLib: $ cabal --version cabal-install version 0.13.3 using version 1.14.0 of the Cabal library $ cabal install -O2 -w ~/.ghc-7.4.1/bin/ghc RepLib --reinstall Resolving dependencies... Configuring RepLib-0.5.2... Building RepLib-0.5.2... Preprocessing library RepLib-0.5.2... [ 1 of 12] Compiling Generics.RepLib.R ( Generics/RepLib/R.hs, dist/build/Generics/RepLib/R.o ) [ 2 of 12] Compiling Generics.RepLib.R1 ( Generics/RepLib/R1.hs, dist/build/Generics/RepLib/R1.o ) [ 3 of 12] Compiling Generics.RepLib.Derive ( Generics/RepLib/Derive.hs, dist/build/Generics/RepLib/Derive.o ) [ 4 of 12] Compiling Generics.RepLib.PreludeReps ( Generics/RepLib/PreludeReps.hs, dist/build/Generics/RepLib/PreludeReps.o ) Loading package ghc-prim ... linking ... done. Loading package integer-gmp ... linking ... done. Loading package base ... linking ... done. Loading package type-equality-0.1.0.2 ... linking ... done. Loading package array-0.4.0.0 ... linking ... done. Loading package deepseq-1.3.0.0 ... linking ... done. Loading package containers-0.4.2.1 ... linking ... done. Loading package pretty-1.1.1.0 ... linking ... done. Loading package template-haskell ... linking ... done. Loading package transformers-0.2.2.0 ... linking ... done. Loading package mtl-2.0.1.0 ... linking ... done. cabal: Error: some packages failed to install: RepLib-0.5.2 failed during the building phase. The exception was: ExitFailure 11 This is without library profiling enabled. If I enable library profiling, the unprofiled build is fine but the profiled build exits with the same error, at the same location. If I pass the "-v" argument to cabal-install, the package builds fine with profiling disabled. The "-v" flag makes no difference with profiling enabled, although some packages (but not RepLib) work with "-v2" I haven't seen this with ghc-7.2.*, but I'm guessing this is some sort of cabal-install bug. Has anyone else witnessed anything like this? Thanks, John L.

1 0

Using HaXml
by Yves Parès 26 Mar '12

26 Mar '12

Hello café, Provided what I read, HaXml seems to be the recommended library for parsing XML files (I'm trying to browse and alter spreadsheets (ODS) and possibly release a package when I'm done), is there somewhere tutorials on how to use it? I used 'xml' package by the past, but HaXml is more substantial. Thanks.

2 2

Execute only one step in Hood-Melville Real time queue.
by Xinyu LIU 26 Mar '12

26 Mar '12

Hi, I read the Hood-Melville real time queue realization in [1]. There are 2 lists maintained in queue, front and rear. When the queue gets unbalanced due to push/pop, it amortized the f ++ reverse r incrementally to build the new front list based on the below mechanism reverse r = reverse' r [] where reverse' [] acc = acc reverse' (x:xs) = reverse' xs (x:acc) and f ++ reverse r == reverse (reverse f) ++ reverse r == reverse' (reverse f) [] ++ reverse r == reverse' (reverse' f []) (reverse' r []) And the re-balance happens once | f | + 1 = | r |. Let's denote m = |f|. incremental f ++ reverse r takes total 2*m + 2 steps. In the the realization mentioned in [1], it execute 2 steps every push and pop to make sure the incremental computation finish before next queue re-balancing. However, I found it's possible to execute only one step per push/pop. because: 1. Next re-balance happens at earliest | f' | + 1 = |r|+|f|+1+1 = 2*m + 2 times by continuously push operation; 2. If we keep a copy of f, and a counter of how many elements left in f, which need to be incrementally 'appended', by continuously m times popping operation. we can finish the f ++ reverse r (actually, only reverse r is needed, as all elements in f are popped) Based on this fact, I rewrite the program as the following: data State a = Empty | Reverse Int [a] [a] [a] [a] -- n, f', acc_f' r, acc_r | Append Int [a] [a] -- n, rev_f', acc | Done [a] -- result: f ++ reverse r deriving (Show, Eq) -- front, length of front, on-goint reverse state, rear, length of reverse data RealtimeQueue a = RTQ [a] Int (State a) [a] Int deriving (Show, Eq) -- we skip the empty error for pop and front empty = RTQ [] 0 Empty [] 0 isEmpty (RTQ _ lenf _ _ _) = lenf == 0 -- O(1) time push push (RTQ f lenf s r lenr) x = balance f lenf s (x:r) (lenr + 1) -- O(1) time pop pop (RTQ (_:f) lenf s r lenr) = balance f (lenf - 1) (abort s) r lenr front (RTQ (x:_) _ _ _ _) = x balance f lenf s r lenr | lenr <= lenf = step f lenf s r lenr | otherwise = step f (lenf + lenr) (Reverse 0 f [] r []) [] 0 -- execute f ++ reverse r step by step step f lenf s r lenr = case s' of Done f' -> RTQ f' lenf Empty r lenr s' -> RTQ f lenf s' r lenr where s' = if null f then next $ next s else next s next (Reverse n (x:f) f' (y:r) r') = Reverse (n+1) f (x:f') r (y:r') next (Reverse n [] f' [y] r') = Append n f' (y:r') next (Append 0 _ acc) = Done acc next (Append n (x:f') acc) = Append (n-1) f' (x:acc) next s = s -- Abort unnecessary appending as the element is popped abort (Append 0 _ (_:acc)) = Done acc -- Note! we rollback 1 elem abort (Append n f' acc) = Append (n-1) f' acc abort (Reverse n f f' r r') = Reverse (n-1) f f' r r' abort s = s Note the 'where' clause in step function. This is because we need an extra step to change the state from (Append 0 _ xs) to (Done xs). I tested this program with invariant testing with QuickCheck. The behavior is correct as other queue implementation. The program can be found here: https://github.com/liuxinyu95/AlgoXY/blob/algoxy/datastruct/elementary/queu… Reference: [1]. Chris Okasaki. ``Purely functional data structures''. P102. Section 8.2.1. Cambridge University Press. ISBN 0521663504 -- Larry, LIU Xinyu https://sites.google.com/site/algoxy/ https://github.com/liuxinyu95/AlgoXY

1 0

What happened with goa package repo?
by Dmitry Malikov 26 Mar '12

26 Mar '12

Cloning into bare repository '/usr/portage/distfiles/egit-src/goa.git'... fatal: The remote end hung up unexpectedly * ERROR: dev-haskell/goa-9999 failed (unpack phase): * git-2_initial_clone: can't fetch from git://github.com/chrisdone/goa.git https://github.com/chrisdone/goa returns 404. So it seems like Chris dropped that repo. Did someone have clone of it? It will be nice to put it back on github.

2 1

Google Summer of Code: a NUMA wishlist!
by Sajith T S 26 Mar '12

26 Mar '12

Dear Cafe, It's last minute-ish to bring this up (in my part of the world it's still March 25), but graduate students are famously a busy and lazy lot. :) I study at Indiana University Bloomington, and I wish to propose^W rush in this proposal and solicit feedback, mentors, etc while I can. Since student application deadline is April 6, I figure we can beat this into a real proposal's shape by then. This probably also falls on the naive and ambitious side of things, and I might not even know what I'm talking about, but let's see! That's the idea of proposal, yes? Broadly, idea is to improve support for NUMA systems. Specifically: -- Real physical processor affinity with forkOn [1]. Can we fire all CPUs if we want to? (Currently, the number passed to forkOn is interpreted as number modulo the value returned by getNumCapabilities [2]). -- Also kind of associated with the above: when launching processes, we might want to specify a list of CPUs rather than the number of CPUs. Say, a -N [0,1,3] flag rather than -N 3 flag. This shall enable us to gawk at real pretty htop [3] output. -- From a very recent discussion on parallel-haskell [4], we learn that RTS' NUMA support could be improved. The hypothesis is that allocating nurseries per Capability might be a better plan than using global pool. We might borrow/steal ideas from hwloc [5] for this. -- Finally, a logging/monitoring infrastructure to verify assumptions and determine if/how local work stays. (I would like to acknowledge my fellow conspirators and leave them unnamed, lest they shall be embarrassed by my... naivete.) Thanks, Sajith. [1] http://www.haskell.org/ghc/docs/latest/html/libraries/base/Control-Concurre… [2] http://www.haskell.org/ghc/docs/latest/html/libraries/base/Control-Concurre… [3] http://htop.sourceforge.net/ [4] http://groups.google.com/group/parallel-haskell/browse_thread/thread/7ec1eb… [5] http://www.open-mpi.org/projects/hwloc/ -- "the lyf so short, the craft so long to lerne." -- Chaucer.

1 0

How to handle C library dependencies on Hackage?
by . 25 Mar '12

25 Mar '12

Hello Cafe, I recently did my first upload to Hackage: <http://hackage.haskell.org/package/jalla>. The library does not build (and therefore unfortunately the documentation is not built either), because the C library that it needs to wrap is usually not installed with normal Linux distributions. It's LAPACKE (note the E at the end). I /could/ upload the source code of that library alongside my Haskell library and build it using Setup.hs (this works on my computer). However, LAPACKE is large compated to my own library (about 16MB sources) and I don't want to flood Hackage with it if it's not necessary. Can anyone here give me some advice on how to solve this on the system that builds the hackage packages? Thanks very much, Christian

2 1

Re: [Haskell-cafe] Open-source projects for beginning Haskell students?
by John Lato 25 Mar '12

25 Mar '12

> From: Heinrich Apfelmus <apfelmus(a)quantentunnel.de> > > Tom Murphy wrote: >> If you want to do Haskell audio synthesis, you could also use >> hsc3 (good start here: http://slavepianos.org/rd/ut/hsc3-texts/) With >> hsc3 you can start on serious audio synthesis with only a few lines of >> Haskell. In my opinion it could use a much larger community. > > While Rohan's bindings to SuperCollider are great, I have found that > SuperCollider itself is quite difficult to understand for a new user. > (My tomata-rubato project aims to be much easier to learn.) > > Also, as far as I am aware, you can't do low-level audio programming in > SuperCollider, i.e. play a list of samples that you've calculated > yourself. That's cool if you're only interested in sound design, but bad > for learning how audio programming works. I think this charge is a bit unfair. If you really want to do low-level stuff, it's possible within SC. You just have to work in SuperCollider, not Haskell (AFAIK). However, it is possible to transfer audio data between Haskell and Csound, in several ways. The hCsound package comes with some examples of transferring the audio input and output streams between csound and haskell. Named channels provide for even more complicated routing if you like. John L.

3 2

GSoC project ideas for web technology
by Jeremy Shaw 25 Mar '12

25 Mar '12

Hello, If you are looking for some ideas for a GSoC project, I have written down some web technology related projects I would like to see. So far I mostly have ideas for improvements to HSX (a templating solution) and acid-state (a pure, haskell persistent datastore). Both these technologies can be used with any Haskell web framework (though they are embraced most fully by Happstack). http://code.google.com/p/happstack/wiki/GoogleSummerOfCode I am an officially registered Haskell GSoC mentor. I am happy to help expand any of these ideas into a genuine GSoC proposal and to act as a mentor on the project. If you have other ideas you wish to pitch, I am happy to hear those as well. - jeremy

1 0