July 2010 - Haskell-Cafe

ANNOUNCE: container-classes-0.0.0.0
by Ivan Lazar Miljenovic 18 Jul '10

18 Jul '10

I'm pleased to announce the fruits of my labours during AusHac [1]: version 0.0.0.0 of the container-classes library [2] I was talking about in my last blog post [3]. [1]: http://www.haskell.org/haskellwiki/AusHac2010 [2]: http://hackage.haskell.org/package/container-classes [3]: http://ivanmiljenovic.wordpress.com/2010/07/14/data-oriented-hierarchies/ The purposes of this library is two-fold: * As a library writer, let your users choose which data structure they want to use rather than whichever one you decided upon when writing the library (e.g. you use lists predominantly, but your users may want to use Sets; by outputting a generic Container value yoru users don't need to keep using Set.fromList and Set.toList). * When writing your own code and you realise that you're using the wrong data structure for your purposes, if you used this library (correctly) then the amount of work you need to do to swap data structures is minimised. This current version is only a first draft. As such, I would appreciate feedback on the API. Currently, it provides its own definitions of all list-oriented functions in the Prelude plus a few more (e.g. partition), and list instances of all the classes. In terms of definitions, most class methods have default definitions in terms of fold. Furthermore, as much as possible I've defined them using a build/fold setup outlined in the paper "A Short Cut to Deforestation" by Andrew Gill, John Launchbury and Simon Peyton Jones (however I haven't yet defined the rules to get this used yet). However, the current list instances use the pre-defined versions in the Prelude and Data.List; I'm not sure if I should keep it this way or - where equivalent - use the pre-defined ones so that hopefully conversions from lists to other data types would also be removed without needing intermediary data structures. I also plan on adding more instances and a Lookup class for Maps, etc. However, we have to start packing up the room now and so my hacking is at an end for this weekend :( -- Ivan Lazar Miljenovic Ivan.Miljenovic(a)gmail.com IvanMiljenovic.wordpress.com

2 2

DpH/repa cache locality
by Gregory Crosswhite 18 Jul '10

18 Jul '10

Hey everyone, Just out of curiosity, what work is being done in the data parallel haskell / repa projects regarding cache locality? The reason I am asking is because, as I understand it, the biggest bottleneck on today's processors are cache misses, and the reason why optimized platform-specific linear algebra libraries perform well is because they divide the data into chunks that are optimally sized for the cache in order to maximize the number of operations performed per memory access. I wouldn't expect data parallel haskell/repa to automatically know what the perfect chunking strategy should be on each platform, but are there any plans being made at all to do something like this? (To be explicit, this isn't meant as a criticism; I'm just curious and am interested in seeing discussion on this topic by those more knowledgeable than I. :-) ) Thanks! Greg

3 2

Best links for Haskell Platform distro packages?
by Don Stewart 17 Jul '10

17 Jul '10

Can distro maintainers confirm these are the best links for each distro package? Debian http://packages.debian.org/squeeze/haskell-platform (or should it be sid?) Fedora: https://admin.fedoraproject.org/pkgdb/acls/name/haskell-platform Gentoo: http://www.haskell.org/haskellwiki/Gentoo/HaskellPlatform NixOS: http://hydra.nixos.org/job/nixpkgs/trunk/haskellPackages_ghc6121.haskellPla… OpenSUSE: https://build.opensuse.org/project/show?project=devel:languages:haskell ?? For Ubuntu, I find a few confusing things: Ubuntu: https://launchpad.net/ubuntu/+source/haskell-platform But nothing at: http://packages.ubuntu.com/lucid/haskell-platform But there's instructions for the src build: http://davidsiegel.org/haskell-platform-in-karmic-koala/ -- Don

5 4

RE: Design for 2010.2.x series Haskell Platform site (Don Stewart)
by Niemeijer, R.A. 17 Jul '10

17 Jul '10

Here's my take on the new design: Screenshot: http://imgur.com/9LHvk.jpg Live version: http://dl.dropbox.com/u/623671/haskell_platform_redesign/index.htm

11 12

unusual behavior from cabal
by Michael Litchard 17 Jul '10

17 Jul '10

Not sure what the correct list is for this observation. I was trying to install gitit, and here is what happened. michael@michael:~/haskell/blog-example$ cabal install gitit Resolving dependencies... cabal: dependencies conflict: happstack-server-0.5.1 requires time ==1.1.4 however time-1.1.4 was excluded because happstack-server-0.5.1 requires time ==1.2.0.3 huh? Cabal seems to think happstack-server-0.5.1 has two conflicting requirements. How do I suss this out?

3 2

ANN: hledger 0.11
by Simon Michael 17 Jul '10

17 Jul '10

hledger 0.11 is released! Thanks to all testers and to Michael Snoyman for much help keeping up with Yesod. Best, -Simon home: http://hledger.org Release notes: 2010/07/17 hledger 0.11 ........................ * split --help, adding --help-options and --help-all/-H, and make it the default command * use "journal" instead of "ledger file"; default suffix is .journal, default file is ~/.journal * auto-create missing journal files rather than giving an error * new format-detecting file reader (mixed journal transactions and timelog entries are no longer supported) * work around for first real-world rounding issue (test zero to 8 decimal places instead of 10) * when reporting a balancing error, convert the error amount to cost * parsing: support double-quoted commodity symbols, containing anything but a newline or double quote * parsing: allow minus sign before commodity symbol as well as after (also fixes a convert bug) * parsing: fix wrong parse error locations within postings * parsing: don't let trailing whitespace in a timelog description mess up account names * add: allow blank descriptions * balance: --flat provides a simple non-hierarchical format * balance: --drop removes leading account name components from a -- flat report * print, register, balance: fix layout issues with mixed-commodity amounts * print: display non-simple commodity names with double-quotes * stats: layout tweaks, add payee/description count * stats: don't break on an empty file * stats: -p/--period support; a reporting interval generates multiple reports * test: drop verbose test runner and testpack dependency * web: a new web ui based on yesod, requires ghc 6.12; old ghc 6.10- compatible version remains as -fweb610 * web: allow wiki-like journal editing * web: warn and keep running if reloading the journal gives an error * web: --port and --base-url options set the webserver's tcp port and base url * web: slightly better browser opening on microsoft windows, should find a standard firefox install now * web: in a web-enabled build on microsoft windows, run the web ui by default Stats: 55 days and 136 commits since last release. Now at 5552 lines of code with 132 tests and 54% unit test coverage.

1 0

Cabal lib (Setup.hs), --package-db, --enable-shared, haskell platform
by Albert Y.C.Lai 17 Jul '10

17 Jul '10

I accidentally found a rarely encountered omission in Cabal (the lib, because via Setup.hs) by building Haskell Platform as shared libs from source. It is rare because you have to use both --enabled-shared and --package-db=blah together to run into it. --package-db=blah is already rare enough (probably only Haskell Platform's build script uses it); --enable-shared is my addition to make it even rarer. (Also I have only tried this on linux x86 32-bit.) Following the essence of scripts/build.sh plus my own --enable-shared: Create an empty lookaside package database with: echo '[]' > /tmp/lookaside Install a package X that depends only on base (say), ask for shared libs, register in /tmp/lookaside only: ./Setup configure --package-db=/tmp/lookaside \ --ghc-pkg-option=--package-conf=/tmp/lookaside \ --enable-shared ./Setup build ./Setup register --inplace Install another package Y that depends on base and X only (say), ask for shared libs. There is a problem now related to X known to /tmp/lookaside only, but not known to the global or the user database: ./Setup configure --package-db=/tmp/lookaside \ --ghc-pkg-option=--package-conf=/tmp/lookaside \ --enable-shared ./Setup build -v2 Now it bombs. As the verbose build messages show, there is no problem creating the static *.a version; it bombs when creating the shared *.so version, and it is because it can't find X, and it is because -package-conf=/tmp/lookaside is not given to ghc at the *.so stage. (Though, the *.dyn_o stage works correctly.) My current fix is: ./Setup configure --package-db=/tmp/lookaside \ --ghc-pkg-option=--package-conf=/tmp/lookaside \ --ghc-option=-package-conf=/tmp/lookaside \ --enable-shared I surely hope we no longer need --ghc-pkg-option and --ghc-option in a week! :)

1 1

Re: [Haskell-cafe] is this a bug ?
by gate03＠landcroft.co.uk 17 Jul '10

17 Jul '10

On Sat 17/07/10 04:17 , Alexander Solla ajs(a)2piix.com sent: > Why are you performing unsafe IO actions? They don't play nice > with laziness. OK, fair cop, but without the unsafe IO action, it still misbehaves. http://hpaste.org/fastcgi/hpaste.fcgi/view?id=27650 Michael.

3 2

Re: in-equality type constraint?
by oleg＠okmij.org 17 Jul '10

17 Jul '10

Ryan Ingram wrote: > But it doesn't generalize; you need to create a witness of inequality > for every pair of types you care about. One can do better, avoiding the quadratic explosion. One merely needs to establish a map from a type to a suitable, comparable representation -- for example, to a type level list of type level numerals. Comparing types for equality, inequality and even order is a simple matter of comparing their representations. The fact that types become totally ordered lets us even implement type-level maps: Data.Map on types. (We may need a Type.* module hierarchy.) That idea was described in the HList paper http://homepages.cwi.nl/~ralf/HList/paper.pdf Section 9. The code is still available, http://code.haskell.org/HList/examples/TTypeable.hs see also TypeEqExplosive.hs, TypeEqTTypeable.hs. > But given the use of UndecidableInstances and OverlappingInstances, I > was hoping that type families could come a little cleaner. Normally if the use of functional dependencies requires UndecidableInstances, type families would ask for them too. As to OverlappingInstances -- that is the key to generic inequality, isn't it? Given the set of pairs of types, the set of the elements representing non-equal types is the complement of the set of elements representing equal types (by equal I mean identical). Overlapping instances is the way to express set complementation. The most general instance is chosen when none of the more specific apply. The set of types chosen by that general instance is the complement for the set of types chosen by the specific instances. > Does "TypeEq a c HFalse" imply proof of inequality, or unprovability > of equality? We are all constructivists here... If 'a' and 'c' are type variables, (TypeEq a c HFalse) is the constraint -- proof obligation if you will -- making sure the variables will never be instantiated to identical types. Strictly speaking, the constraint is discharged if 'a' and 'c' are two ground, and syntactically distinct (non-identical) types. In reality, I think GHC is able to discharge the constraint if 'a' and 'c' are grounded ``sufficiently enough'' for the difference to become apparent (e.g., the head constructors are different).

2 1

ghc api printing of types
by Phyx 17 Jul '10

17 Jul '10

I was wondering how given a Type I can get a pretty printed type out of it. I'm currently using showSDocUnqual . pprType . snd . tidyOpenType emptyTidyEnv But this has the problem that predicates don't get printed, anyone know how GHCi does this? Thanks, Phyx

2 3