September 2011 - Haskell-Cafe

Can't resolve class dependences (making Boolean class)
by Grigory Sarnitskiy 13 Sep '11

13 Sep '11

Hello! I'm trying to have a class for booleans called Boolean (the methods are not complete): class MyEq a where (===) :: (Boolean b) => a -> a -> b class (MyEq a) => Boolean a where (/\) :: a -> a -> a instance MyEq Bool where x === y = x==y instance Boolean Bool where (/\) = (&&) However, to make Bool an instance of Boolean I need to make it an instance of MyEq first, which I can't, because to define === I need Bool to be in Boolean. Indeed the code above give the error: Could not deduce (b ~ Bool) from the context (Boolean b) bound by the type signature for === :: Boolean b => Bool -> Bool -> b at 1.hs:8:5-18 `b' is a rigid type variable bound by the type signature for === :: Boolean b => Bool -> Bool -> b at 1.hs:8:5 In the expression: x == y In an equation for `===': x === y = x == y In the instance declaration for `MyEq Bool' Failed, modules loaded: none. How can I overcome the issue?

3 3

extending and reusing cmdargs option specs ?
by Simon Michael 13 Sep '11

13 Sep '11

Hi Neil, I just spent a day converting hledger from getopt to cmdargs. cmdargs feels more high level and featureful and nicer. And yet... I haven't reduced the line count that much - nothing like your HLint 3:1 ratio. And, I may have made things worse for myself in the reuse/avoiding boilerplate department: I'm not sure how to reuse a cmdargs options data structure, extending it with a few more options. Using getopt I was able to do this without repeating myself (as long as I defined the full set of Opt constructors in one place.) I've looked at the more advanced cmdargs api, but don't see a way yet - would you have any ideas ? I want this because I have multiple executables (hledger, hledger-vty, hledger-web etc.) which should share most (but not all) options. Also, I'd like to move a generic subset of report options, without the ui- specific ones, into hledger-lib for use by all apps. Also, the hledger executable has multiple commands, so I'd like to define a mode for each, but not have to redeclare all the same options for each mode - I didn't see how to do that. As always, thanks a lot for cmdargs! -Simon

4 6

Can't link OpenCL on Windows
by Jason Dagit 13 Sep '11

13 Sep '11

Hello, I'm trying to get the OpenCLRaw bindings to a point where I can use them on windows. I've forked the the OpenCLRaw repo on github so I can make modifications as needed. My branch is here: https://github.com/dagit/OpenCLRaw I've been mostly working out of my "FunPtr" branch. The issue I'm having is this: I installed AMD's OpenCL SDK, converted their Visual Studio specific .lib file to a file that gcc can handle (.a file), but ghc can't seem to link with it. I get undefined symbols for everything I use in the OpenCL API. I was able to build a "trivial" C program and link it using the .a file that I generated and gcc from mingw (not from the Haskell install). I'm using the latest windows release of the Haskell platform. These are the steps I used for generating the .a file: http://forums.amd.com/forum/messageview.cfm?catid=390&threadid=138890 I used the commands in the example script (e.g., gendef and dlltool). I've tried to use 32bit everything as much as possible, as I know that GHC will want everything to be 32bit, so I don't think it's a 32bit vs. 64bit issue. Does anyone know if there is something different about invoking gcc under ghc instead of the gcc that I get from mingw? I've also played with the ghc command line (I used cabal-dev --verbose=3 to inspect the command line) and I'm still unable to massage it into a working state. Any help would be appreciated! Thanks, Jason

1 1

Haskell Platform Older Releases for Linux
by Luis Cabellos 12 Sep '11

12 Sep '11

Hi, I search the Haskell Platform webpage for older releases, I found it but maybe it's worth to put a table of older releases as tar.gz file with their ghc dependency, like (versions totally make up): HP 2011.2.0.0 source, March 2011. -> ghc 13.1 HP 2010.2.0.0 source, July 2010. -> ghc 3.14 HP 2010.1.0.0 source, March 2010. -> ghc 616 HP 2009.2.0.2 source, July 2009. -> ghc 1 HP 2009.2.0.1 source, June 2009. -> ghc 0 HP 2009.2.0 source, May 2009. -> ghc -1 Thanks, Luis

3 2

Package descriptions on hackage
by Joachim Breitner 12 Sep '11

12 Sep '11

Dear hackage package authors, this is a short message from your distribution package creators: Please, if possible, write good, not too short descriptions, and also keep them up to date. Of course, users browsing hackage will benefit as well. Keep in mind that the users do not know what other packages this relates to. So if the package is part of a larger set of corresponding packages, include a general blob and then explain this particular packages. This is especially true for packages that contain the base types for a suite of packages. Also think about all the possible keywords that someone might enter in a search engine when looking for something like your package, and ideally explain how your library relates to these keywords so that the user can estimate whether it useful for him even before reading the API. Here a few negative examples, not to expose particular authors, but selected because they were the most resent that the Debian Haskell Team had packaged. ========= The http-enumerator package This package uses attoparsec for parsing the actual contents of the HTTP connection. The only gotcha is the withHttpEnumerator function, otherwise should do exactly what you expect. (A list of supported features would be great, maybe a short note what enumerator means. Also, the second sentence is incomprehensible to me and the latest version does not contain a withHttpEnumerator function any more) ========= The representable-functors package Representable functors (Clearly not helpful. Here is what Iulian came up for for the Debian package, from the individual module documentation, but I still think the author could do better This package provides a generalized Store comonad, parameterized by a Representable functor (the representation of that functor serves as the index of the store) and a generalized State monad, parameterized by a Representable functor (the representation of that functor serves as the state). Representable functors on Hask all monads, being isomorphic to a reader monad. Representable contravariant endofunctors over the category of Haskell types are isomorphic to (_ -> r) and resemble mappings to a fixed range. Representable endofunctors over the category of Haskell types are isomorphic to the reader monad and so inherit a very large number of properties for free.) ========= The data-lens package Haskell 98 Lenses (Also not helpful. And probably a very useful package! Iulian wrote „A lense is a composable notion of a purely functional field accessor.“ which is still quite short, but explains the use of the package much better) ========= The algebra package Constructive abstract algebra (Seems to be the “main” package of a series of packages. Yet another reason for having a good description. Here a suggestion, again by Iulian: “This package provides algebraic structures, such as groups, fields, rings, modules. It also provides bands, also known as idempotent semigroups (band is a semigroup where every element is equal to its own square), coalgebras, semirings (rigs), idempotent semirings, also known as dioids.“) ========= I know that most people have better things to do than to write API documentation, so it is not unexpected that, after having written the API documentation after all, they don’t spend too much time on the description. But think about it: It is the first thing possible users are going to see from your package. So if you want to attract users, do take your time to write a comprehensive and comprehensible description. Thanks, Joachim -- Joachim "nomeata" Breitner mail(a)joachim-breitner.de | nomeata(a)debian.org | GPG: 0x4743206C xmpp: nomeata(a)joachim-breitner.de | http://www.joachim-breitner.de/

7 7

TypeOperators and Unicode
by Grigory Sarnitskiy 12 Sep '11

12 Sep '11

I want to have Unicode symbols for type operator: {-# LANGUAGE TypeOperators #-} data a ── b = Foo a b But it doesn't work. Any suggestions?

4 4

matrix vector product
by kolli kolli 11 Sep '11

11 Sep '11

Hi, can anyone help me with the matrix vector product ....

2 1

Parsec: non-greedy 'between'
by Scott Lawrence 11 Sep '11

11 Sep '11

Hey all, Trying to match C-style comments, I have: between (string "/*") (string "*/") $ many anyChar Which doesn't work, because it is equivalent (ignoring returned values) to do {string "/*"; many anyChar; string "*/"} If the termination criterion was a single character, then I could use noneOf or (satisfy . not), but that doesn't help here. So... what am I missing? Thanks in advance. -- Scott Lawrence

3 3

Functional Programming eXchange 2012: Call for abstracts
by Robert Pickering 11 Sep '11

11 Sep '11

Hello all, I’m sending this email to several mailing lists so my apologies if you see this twice. Functional Programming eXchange is a developer conference that focuses on functional programming. The 2012 edition will take place on Friday March 16th March 2012, at the Skills Matter eXchange, in London. Typically talks are explore concepts that often crop up in functional programming such as parallelism or domain specific languages, look at frameworks built in functional languages or are cases studies of using functional programming in industry. Previous speakers have included: Simon Peyton Jones, Miles Sabin, David Pollak, Adam Granicz, Antonio Cisternino, Tomas Petricek, Jonas Boner, and Viktor Klang. Videos of the talks at the previous events can be found here: http://skillsmatter.com/event/scala/functionalpx-2011 http://skillsmatter.com/event/cloud-grid/functional-programming-exchange-20… If you are interested at speaking at this conference please send a talk abstract to talks_functionalpx(a)skillsmatter.com. I’m aiming to make the decision about which talks have been chosen around 14th October so if you submit an abstract expect to hear from me around this time. Kind regards, Robert

1 0

ANN: Netwire 1.2.4
by Ertugrul Soeylemez 11 Sep '11

11 Sep '11

Hello there, version 1.2.4 of netwire is out. Major changes: * Changed the semantics of time. Previously if a wire was not run (because of an earlier inhibiting wire), then its local time was suspended. Example: proc _ -> do t1 <- time -< () fps1 <- avgFps 1000 -< () require_ <<< wackelkontakt -< () t2 <- time -< () fps2 <- avgFps 1000 -< () identity -< printf "Time: %8.2f %8.2f\nFPS : %8.2f %8.2f" t1 t2 fps1 fps2 In the previous version t2 would be about half of t1, because the {require_ <<< wackelkontakt} wire inhibits for about half of all instances. Further fps1 and fps2 would be about equal. In other words, time was local, while FPS were global. This is not very useful in practical applications. With the new semantics t1 and t2 will be equal. Instead the framerate will drop after the inhibiting wire, such that fps2 will be about half of fps1. So now time is global and FPS are local. This is useful e.g. in a stateless network application (like a web app) to ban users, who are trying logins too fast. The following wire allows one login trial per second in average over the last 10 trials (the faster you try the longer you have to wait): login :: Wire MyApp IpAddress User login = context $ proc ipAddr -> do fps <- avgFps 10 -< () require_ -< fps < 1 {- login procedure -} Used with ArrowPlus this easily lets you fetch the currently logged in user from the session, and if there is none, request a login: getUser :: Wire MyApp IpAddress User getUser = getCurrentUser <+> login * Context-sensitive wires. There is now a new wire transformer called 'context', which is essentially a context switcher establishing multiple wire threads of control: context :: (Ord a, Monad m) => Wire m a b -> Wire m a b It takes a base wire and evolves it individually for each possible input value. This is a very convenient alternative to parallel routing switches. The 'context' wire transformer never forgets a context. For many applications (like sessions in a web application) this is not desirable. For them there is an alternative transformer 'contextLimited', which allows you to garbage-collect. Minor changes: * More responsive calculus wires. Reflect changes immediately without the one-instant delay. If you need the delay for feedback, use 'delay' explicitly. * New analysis wires 'collect' and 'lastSeen'. * New inhibition wires 'forbid_', 'inhibit_' and 'require_'. * Separate session starting/cleaning functions in FRP.NetWire.Session. However, the exception-safe withWire function is the preferred way. * Optimized internal representation. There are now only two constructors WGen and WArr for the Wire type instead of four. I dropped WConst and WId, because they are representable as special cases of WArr without noticable performance loss. This improves overall performance, because now there are only two constructors to pattern-match against. Right now I'm working intensively on netwire, so if you have any feature requests, just mail me. There is a good chance to find your desired feature implemented quickly. Note: I have been too quick with version numbers. Major version 2 will be the one with the stabilized API, so for now with increasing minor version numbers (but not patch levels!) you have to be prepared for API changes, although I don't expect big changes anymore. Enjoy and keep the feedback coming! Greets, Ertugrul -- nightmare = unsafePerformIO (getWrongWife >>= sex) http://ertes.de/

1 0