- Haskell-Cafe - Haskell.org

C global vars local to threads
by Maurício 29 Jun '09

29 Jun '09

I would like to write a wrapper to C global variable, but that variable is unique to each thread. Is there some native support on that in Haskell FFI? (I imagine I probably should write a C function to get its pointer as an 'IO (Ptr ...)', but I would like to check if there's a more standard way to do that.) Thanks, Maurício

1 0

test-framework success
by Simon Michael 28 Jun '09

28 Jun '09

Max - I was thinking about how to drive my new shell tests with your framework. I have: data ShellTest = ShellTest { filename ∷ String ,command ∷ String ,stdin ∷ Maybe String ,stdoutExpected ∷ Maybe String ,stderrExpected ∷ Maybe String ,exitCodeExpected ∷ Maybe ExitCode } deriving (Show) parseShellTest ∷ FilePath → IO ShellTest runShellTest ∷ ShellTest → IO Bool I converted to hunit tests with shellTestToHUnitTest t = filename t ~: do {r ← runShellTest t; assertBool "" r} and tied test-framework into my test runner with main = do args ← getArgs let (opts,files) = partition ((=="-").take 1) args hunittests ← mapM (λf → parseShellTest f >>= (return ∘ shellTestToHUnitTest)) files withArgs opts $ defaultMain $ concatMap hUnitTestToTests hunittests With surprisingly little code and pain, the new runner runs my 14 shell tests, with colored output, in the same time as the old one - 3.5s on this macbook. But, I can now add -j8 and get the same results output in.. 0.13s. This quite surprised me, and now I want to say: thank you very much! :) Best, -Simon $ time tools/shelltest2 tests/*.test -j8 :tests/eliding-balance.test: [OK] :tests/missing-real-and-virtual-amt.test: [OK] :tests/null-accountname-component.test: [OK] :tests/parens-in-account-name.test: [OK] :tests/sample-balance-depth.test: [OK] :tests/sample-balance-o.test: [OK] :tests/sample-balance.test: [OK] :tests/unbalanced.test: [OK] :tests/unicode-account-matching.test: [OK] :tests/unicode-balance.test: [OK] :tests/unicode-description-matching.test: [OK] :tests/unicode-print.test: [OK] :tests/unicode-register.test: [OK] :tests/virtual.test: [OK] Test Cases Total Passed 14 14 Failed 0 0 Total 14 14 real 0m0.132s user 0m0.104s sys 0m0.104s

3 3

What is an "expected type" ...
by michael rice 28 Jun '09

28 Jun '09

as opposed to an "inferred type"? Michael

10 11

"shadowing" keywords like "otherwise"
by Vasili I. Galchin 28 Jun '09

28 Jun '09

Hello, I am currently using ghc 6.8.2. With the following, swishParse :: String -> String -> SwishStateIO (Maybe RDFGraph) swishParse fnam inp = do { fmt <- gets $ format ; case fmt of N3 -> swishParseN3 fnam inp otherwise -> do { swishError ("Unsupported file format: "++(show fmt)) 4 ; return Nothing } } I am receiving a shadow warning: Swish/HaskellRDF/SwishCommands.hs:304:12: Warning: Defined but not used: `otherwise' It seems to me that in the code base somewhere that there is a "redefine" of the keyword"otherwise". I haven't read the Haskell 98 Report but I thought that it was not possible to redefine keywords. ?? Kind regards, Vasili

6 7

A Reader Monad Tutorial
by Henry Laxen 28 Jun '09

28 Jun '09

Dear Group, If any of you are struggling with understanding monads, I've tried to put together a pretty through explanation of what is behind the Reader monad. If you're interested, have a look at: http://www.maztravel.com/haskell/readerMonad.html Enjoy. Henry Laxen

2 2

Re: [Haskell-cafe] Rigid type variable error
by Darryn 28 Jun '09

28 Jun '09

Thanks for the help previously received, but I still cannot seem to get on top of this. The types for the constructor K will not resolve and I'm at a loss to work out what to do with it. If anyone can offer a further explanation and help I would be very grateful. My code (File Test5.hs): ---------------------------- {-# LANGUAGE ExistentialQuantification #-} class A a where a1 :: a a2 :: a -> a a3 :: (B b) => b -> a class B b where b1 :: Int -> b --data Ainst b = I | J (Ainst b) | K b -- a3 :: (B b, A a) => b -> a -- yet without the constraint on K, K :: b -> Ainst b -- so the above data definition fails. Trying to -- existentially quantify K below seems to make -- sense, but also fails ... data Ainst b = I | J (Ainst b) | (B b) => K b instance (B b) => A (Ainst b) where a1 = I a2 = J a3 = K -- Reported line of the error data Binst = Val Int instance B Binst where b1 = Val ------------------------------- The error from ghci is as follows: Test5.hs:25:9: Couldn't match expected type `b' against inferred type `b1' `b' is a rigid type variable bound by the type signature for `a3' at Test5.hs:7:13 `b1' is a rigid type variable bound by the instance declaration at Test5.hs:16:12 Expected type: b -> Ainst b1 Inferred type: b1 -> Ainst b1 In the expression: K In the definition of `a3': a3 = K Failed, modules loaded: none. Thanks in advance for any help. Apologies if what I am doing is odd or the answer is obvious, I'm still very new to Haskell. Darryn.

3 2

Re: [Haskell-cafe] "shadowing" keywords like "otherwise"
by Janis Voigtlaender 28 Jun '09

28 Jun '09

Vasili I. Galchin wrote: > so Janis I have to look where the original author overrode "otherwise"? > If so, just great .. a ton of code. ;^( No, you don't have to look for this. It is in the code snippet you sent: swishParse :: String -> String -> SwishStateIO (Maybe RDFGraph) swishParse fnam inp = do { fmt <- gets $ format ; case fmt of N3 -> swishParseN3 fnam inp otherwise -> do { swishError ("Unsupported file format: "++(show fmt)) 4 ; return Nothing } } In the case expression you use "otherwise" as a variable. But you do not use that variable in that branch. Hence the warning: Swish/HaskellRDF/SwishCommands.hs:304:12: Warning: Defined but not used: `otherwise' This has nothing to do with someone else shadowing the definition of "otherwise". Ciao, Janis. -- Dr. Janis Voigtlaender http://wwwtcs.inf.tu-dresden.de/~voigt/ mailto:voigt@tcs.inf.tu-dresden.de

1 0

bizarre syntax error
by Geoffrey Irving 28 Jun '09

28 Jun '09

I ran into a unfortunate syntax error just now. I figured I'd share it because it's the weirdest message I've ever gotten out of ghc. The broken code is case t of TyApply tv types -> do (tvl, cases) <- lookupDatatype prog tv let tenv = Map.fromList (zip tvl types) -- GHC POINTS HERE caseType (c,vl,e') | Just tl <- List.lookup c cases = if length vl == a then expr prog global (foldl (\e (v,t) -> Map.insert v t e) env (zip vl (map (subst tenv) tl))) e' else typeError ("arity mismatch in pattern: "++show (pretty c)++" expected "++show a++" argument"++(if a == 1 then "" else "s") ++" but got ["++concat (intersperse ", " (map (show . pretty) vl))++"]") where a = length tl -- THIS IS THE PROBLEM | otherwise = typeError ("datatype "++show (pretty tv)++" has no constructor "++show (pretty c)) defaultType Nothing = return [] defaultType (Just (v,e')) = expr prog global (Map.insert v t env) e' >>=. \t -> [t] join t1 t2 | Just t <- unifyS t1 t2 = return t | otherwise = typeError ("failed to unify types "++show (pretty t1)++" and "++show (pretty t2)++" from different case branches") caseResults <- mapM caseType pl defaultResults <- defaultType def foldM1 join (caseResults ++ defaultResults) _ -> typeError ("expected datatype, got "++show (pretty t)) ghc complains that "The last statement in a 'do' construct must be an expression" and points to the line declaring "tenv". The actual problem is that you can't put a "where" block in between two pattern guards. :) Geoffrey

2 1

ANNOUNCE: Cal3D animation library
by Gregory D. Weber 27 Jun '09

27 Jun '09

The Cal3D for Haskell project provides a partial binding to the C++ Cal3D animation library. The project homepage is http://haskell.org/haskellwiki/Cal3d_animation There are three packages available on hackage: * cal3d-0.1 is a Haskell binding to the Cal3D library itself. * cal3d-opengl-0.1 adds a few functions for using Cal3D with OpenGL. * cal3d-examples-0.1 provides a simple example based on the Cally Demo. Cal3D is a C++ library for skeletal-based character animation. It is platform-independent and independent of any particular graphics API. In fact, it does not actually render the animation, but provides hooks that the graphics API (such as OpenGL) can use to do the actual drawing. (As far as I know, though, the only graphics API available for Haskell is OpenGL.) For more information about the (C++) Cal3D library, please see * Cal3D FAQ: http://cal3d.sourceforge.net/docs/api/html/cal3dfaq.html * Cal3D Homepage: http://home.gna.org/cal3d/ -- ___ ___ __ _ / _ \ / _ \| | | | Gregory D. Weber, Associate Professor / /_\// / | | | /\ | | Indiana University East / /_\\/ /__| | |/ \| | http://mypage.iu.edu/~gdweber/ \____/\_____/\___/\__/ Tel. (765) 973-8420; FAX (765) 973-8550

1 0

Trivial pivoting for the DSP lu decomposition
by Fernan Bolando 27 Jun '09

27 Jun '09

Hi all I created the beginnings of a simple circuit simulator using haskell. It can be downloaded from. http://plan9.bell-labs.com/sources/contrib/fernan/escomma.tar.bz2 It uses a modified version of the haskell DSP library matrix. I extended it with a simple pivoting Its not very haskelly, but it works with matrices that didn't work with the unmodified version. thanks for those that answered my list processing questions. I used that code to create a permutation vector for the pivoting. I am now looking into non-linear matrix solver. does anybody have a reference code for that? fernan -- http://www.fernski.com

1 0

Rigid type variable error
by Darryn 27 Jun '09

27 Jun '09

From: Darryn <djreid(a)aapt.net.au> To: beginners(a)haskell.org Subject: Rigid type variables match error Date: Sat, 27 Jun 2009 10:18:13 +0930 Hi, I wonder if anyone can explain what is going on here and what to do about it. I'm fairly new to Haskell, so apologies in advance if my question seems naive. I've cut my code down to a minimum that reproduces the problem. I receive an error in GHCI in the following code, complaining that it cannot match the rigid type variables for the instance definition for Ainst for the function a3. Can anyone advise about what to do about it? ------------------------------------------------ class A a where a1 :: a a2 :: a -> a a3 :: (B b) => b -> a class B b where b1 :: Int -> b data (B b) => Ainst b = I | J (Ainst b) | K b instance (B b) => A (Ainst b) where a1 = I -- a2 :: (B b, A a) => a -> a a2 = J -- a3 :: (B b, A a) => b -> a a3 = K -- Error! -- a3 = K `asTypeOf` a3 -- Error even with this! data Binst = Val Int instance B Binst where b1 = Val ------------------------------------------------ Test5.hs:17:9: Couldn't match expected type `b1' against inferred type `b' `b1' is a rigid type variable bound by the instance declaration at Test5.hs:12:12 `b' is a rigid type variable bound by the type signature for `a3' at Test5.hs:5:13 Expected type: b -> Ainst b1 Inferred type: b -> Ainst b In the expression: K `asTypeOf` a3 In the definition of `a3': a3 = K `asTypeOf` a3 Failed, modules loaded: none. Thanks in advance for any help anyone can provide.

2 1

System.Vacuum.Ubigraph
by Cetin Sert 26 Jun '09

26 Jun '09

Hi, The package seems to have a connection problem on my machine. Prelude> import System.Vacuum.Ubigraph Prelude System.Vacuum.Ubigraph> view (0,1) Loading package syb ... linking ... done. Loading package base-3.0.3.1 ... linking ... done. Loading package parsec-2.1.0.1 ... linking ... done. Loading package network-2.2.1.3 ... linking ... done. Loading package array-0.2.0.0 ... linking ... done. Loading package containers-0.2.0.1 ... linking ... done. Loading package bytestring-0.9.1.4 ... linking ... done. Loading package old-locale-1.0.0.1 ... linking ... done. Loading package old-time-1.0.0.2 ... linking ... done. Loading package pretty-1.0.1.0 ... linking ... done. Loading package filepath-1.1.0.2 ... linking ... done. Loading package unix-2.3.2.0 ... linking ... done. Loading package directory-1.0.0.3 ... linking ... done. Loading package process-1.0.1.1 ... linking ... done. Loading package random-1.0.0.1 ... linking ... done. Loading package haskell98 ... linking ... done. Loading package mtl-1.1.0.2 ... linking ... done. Loading package time-1.1.3 ... linking ... done. Loading package vacuum-0.0.95 ... linking ... done. Loading package HTTP-4000.0.7 ... linking ... done. Loading package HaXml-1.13.3 ... linking ... done. Loading package dataenc-0.13.0.0 ... linking ... done. Loading package haxr-3000.2.1 ... linking ... done. Loading package vacuum-ubigraph-0.1.0.3 ... linking ... done. *** Exception: user error (Error calling ubigraph.clear: �: openFile: does not exist (No such file or directory)) In ubigraph.clear: �: openFile: the part between clear and openFile is not always the same character or sequence of characters! I guess it's a problem with Network.XmlRpc.Client but couldn't locate it. Anyone else that has (had) this problem? Best Regards, Cetin Sert

1 0

download, download-curl ...
by Keith Sheppard 26 Jun '09

26 Jun '09

Hello Cafe, I'm looking to make it possible for people to use urls directly in my haskell program (TxtSushi) and I'd like your suggestions if you have any. I really like the API's for download and download-curl, but I'm wondering what the practical differences are between the two? Is download-curl a replacement for download? Will it build OK on any platform or do I have to use something like the HTTP library if I want it to work for people with Windows? (I really like the fact that the "download" APIs support FTP too though) Thank You Keith -- keithsheppard.name

2 1

Multi-module scripts
by Maurício 26 Jun '09

26 Jun '09

Is it possible to write a Haskell script that uses a module that is also going to be interpreted? Like, say: #!/usr/bin/runhaskell module Main (main) where import OtherModule main = (...) ---- where OtherModule is available in, say, OtherModule.hs and not in the already compiled and installed library? Thanks, Maurício

2 1

Why is Bool no instance of Num and Bits?
by Stephan Friedrichs 26 Jun '09

26 Jun '09

Hi! When looking for an xor function, I found one in Data.Bits but couldn't use it for Bool, because Bool is no instance of Bits and of Num (which would be necessary, because it's "class (Num b) => Bits b"). My question is: Why not? We could declare instance Num Bool where (+) False = id (+) True = not (*) True True = True (*) _ _ = False (-) = (+) negate = id abs = id signum = const True fromInteger = not . even which basically implements the field with 2 elements and instance Bits Bool where bitSize = const 1 isSigned = const False (.&.) = (&&) (.|.) = (||) xor = (+) complement = not shift = const shiftL = const shiftR = const rotate = const rotateL = const rotateR = const bit = (==0) setBit _ 0 = True setBit b _ = b clearBit _ 0 = False clearBit b _ = b complementBit b 0 = not b complementBit b _ = b testBit b 0 = b testBit _ _ = False quite trivial... Why is this not part of base? Or am I missing something? //Stephan -- Früher hieß es ja: Ich denke, also bin ich. Heute weiß man: Es geht auch so. - Dieter Nuhr

12 16

Hack (web) and apache configuration
by Henry Laxen 26 Jun '09

26 Jun '09

Dear Group, I am posting this here even though it probably belongs on the apache list because I suspect other haskell users will be able to find it here more easily. I am playing around with hack, and am having trouble with configuring apache with fastcgi to make things work. My understanding of the "hack" concept is that it provides a stardardized interface that lets you glue together web "Applications". It also provides several front-ends, such as happs, and fastcgi etc. Now based on looking at the Middleware supplied with hack, it seems to be trying to dispatch based on the contents of the pathInfo field of the Env record. So, my question is, how do we configure Apache2 with the fastcgi handler so that something appears in the pathInfo field? I have tried several things, the most recent being: RewriteEngine on RewriteRule ^/(.*)$ /hackTest?input=$1 [T=application/x-httpd-cgi] <Location /> SetHandler fastcgi-script Options ExecCGI FollowSymLinks </Location> but the pathInfo field is always null. Env {requestMethod = GET, scriptName = "/lambda", pathInfo = "", queryString = "input=lambda", serverName = "127.0.0.1", serverPort = 80, http = [("FCGI_ROLE","RESPONDER"),("SCRIPT_URL","/lambda"), ("SCRIPT_URI","http://127.0.0.1/lambda"),("User-Agent","curl/7.18.2 (x86_64-pc-linux-gnu) libcurl/7.18.2 OpenSSL/0.9.8g zlib/1.2.3.3 libidn/1.8 libssh2/0.18"), ("Host","127.0.0.1"),("Accept","*/*"),("PATH","/usr/local/bin:/usr/bin:/bin"), ("SERVER_SIGNATURE","<address>Apache/2.2.9 (Debian) mod_fastcgi/2.4.6 proxy_html/3.0.0 mod_apreq2-20051231/2.6.0 mod_perl/2.0.4 Perl/v5.10.0 Server at 127.0.0.1 Port 80</address>\n"), ("SERVER_SOFTWARE","Apache/2.2.9 (Debian) mod_fastcgi/2.4.6 proxy_html/3.0.0 mod_apreq2-20051231/2.6.0 mod_perl/2.0.4 Perl/v5.10.0"), ("SERVER_NAME","127.0.0.1"),("SERVER_ADDR","127.0.0.1"), ("SERVER_PORT","80"),("REMOTE_ADDR","127.0.0.1"), ("DOCUMENT_ROOT","/home/henry/maztrave2/www/fcgi"), ("SERVER_ADMIN","[no address given]"), ("SCRIPT_FILENAME","/home/henry/maztrave2/www/fcgi/hackTest"), ("REMOTE_PORT","44936"), ("GATEWAY_INTERFACE","CGI/1.1"),("SERVER_PROTOCOL","HTTP/1.1"), ("REQUEST_METHOD","GET"),("QUERY_STRING","input=lambda"), ("REQUEST_URI","/lambda"),("SCRIPT_NAME","/lambda")], hackVersion = [2009,5,19], hackUrlScheme = HTTP, hackInput = Empty, hackErrors = HackErrors, hackHeaders = []}% I think what I want is to have all URLS, such as: http://127.0.0.1/lambda be dispatched though my hackTest executable, without having to go through the rewrite, but I can't convice apache to do that. In the interest of completeness, my hackTest.hs file is the following: import Hack import Hack.Handler.FastCGI import Data.ByteString.Lazy.Char8 (pack) import Hack.Contrib.Middleware.Lambda app :: Application app = \env -> return $ Response { status = 200 , headers = [ ("Content-Type", "text/plain") ] , body = pack $ show env } main = runFastCGIorCGI $ lambda app ------------------------------------------------------ One final comment to the authors of hack. Would you please consider renaming this project. hack is such a common word that has nothing to do with this project that it make searching the web with google, etc. almost useless. I realize it is a clever respelling of the ruby version "rack", but please consider naming it something more unique, while it is still relatively new on the web. Thanks in advance for your help. Best wishes, Henry Laxen

4 4

Generic Graph Class
by Ivan Lazar Miljenovic 26 Jun '09

26 Jun '09

At the moment, there are at least three ways people use graph data-structures in Haskell: * Data.Graph from containers * Data.Graph.Inductive from FGL * A custom job (usually using something like IntMap). If we look on Hackage, there are a number of graph-related packages there, each of which uses one of the above approaches. However, for the more "generic" packages that operate _on_ graphs (rather than using graphs as an internal data structure), there is (as far as I can tell, anyway) usually no reason why at least most of that package cannot work on _any_ of these types of graphs. Examples include (in no particular order; found by looking for "graph" on the main hackage package list page): * hgal {uses Data.Graph} * vacuum and related packages {custom, [(a, [a])]} * PkgGraph {custom, a -> [b] AFAICT} * cabalgraph {custom, a -> b AFAICT} * graphviz {uses FGL} * SceneGraph {FGL} * data-reify {custom} * hsc3 and related packages {custom, not sure how} * dotgen {custom} * fenfire {custom} * GraphSCC {Data.Graph} * Graphalyze {FGL} * SourceGraph {FGL via Graphalyze} Just to give you an idea, out of these the following implement some form of graph -> .dot function (for GraphViz): * vacuum * PkgGraph * cabalgraph * graphviz (well, that's the point of this package... used by Graphalyze and SourceGraph) * SceneGraph * hsc3-dot * dotgen (From its description: "This package provides a simple interface for building .dot graph files...") Why do we duplicate so much work and effort? Not just in terms of generating Dot representations for use in GraphViz, but also algorithm development? For example, I can't see any reason why hgal shouldn't also work for any generic graph type, possibly even GraphSCC (note that sometimes specialisation isn't always avoidable: my Graphalyze library uses the inductive nature of FGL for its algorithms). I thus propose that we work out a generic graph class that can be used by the various libraries we have and use, to avoid this duplication of effort (I have already proposed that I intended to add such functionality to the graphviz library, but I'm throwing open the design of such a class to the general community). This means that even if you have to use some custom graph-like data structure in your program, you can take advantage of one of the libraries available (e.g. graphviz) without having to write your own graph functions for common tasks. Here is my current thinking on how such a class could be defined. Note that I'm defining it for _directed_ graphs, as all graph definitions used seem to match this mould (and we can emulate undirected graphs using directed graphs but not the other way around). ====================================================================== type Vertex = Int class GenericGraph g where vertices :: g -> [Vertex] edges :: g -> [(Vertex, Vertex)] order :: g -> Int order = length . vertices size :: g -> Int size = length . edges -- All vertices v' such that there's an edge (v',v) adjacentIn :: g -> Vertex -> [Vertex] adjacentIn g v = map fst . filter ((==) v . snd) $ edges g degIn :: g -> Vertex -> Int degIn g = length . adjacentIn g -- All vertices v' such that there's an edge (v,v') adjacentOut :: g -> Vertex -> [Vertex] adjacentOut g v = map snd . filter ((==) v . fst) $ edges g degOut :: g -> Vertex -> Int degOut g = length . adjacentIn g adjacent :: g -> Vertex -> [Vertex] adjacent g v = adjacentIn g v ++ adjacentOut g v degree :: g -> Vertex -> [Vertex] degree g = length . adjacent g -- For this function, maybe use as a default something based on the -- hidden isUndir function in graphviz's Data.GraphViz isDirected :: g -> Bool ====================================================================== I'm not saying that the above class is perfect (though I can't think of any other generic graph-related functions we might want to include in there off the top of my head). However, I do think it's a good starting point. Once we have defined such a class, how should we package it? In particular, how should we define instances for pre-packaged (i.e. Data.Graph and FGL) graph data structures? Method 1: Define the instances in the same package. --------------------------------------------------- If we include the instances in the same package, then all libraries that use this class will require FGL, etc. to be installed (e.g. if you want to use the graphviz library for your own custom class, then thanks to the static linking used by GHC you will still have an unwanted FGL dependency in your program). Method 2: Have seperate instance packages. ------------------------------------------ No unwanted dependencies, but this means we have what are basically useless packages containing nothing but an orphan instance. Method 3: Define the instance in Data.Graph and FGL --------------------------------------------------- I'm not sure if the maintainers of these packages (let alone the GHC maintainers who require containers) will appreciate us giving them another dependency for their packages. Furthermore, this limits users of libraries that use this class to using versions of the data structures contained in libraries new enough to have this instance defined in them. Method 4: Give a default instance in the documentation. ------------------------------------------------------- Users can copy/paste the instance definition from the documentation and put it in their code. This will allow them to tweak the definitions if they so desire, but will lead to code duplication. Please tell me any ideas/criticisms you have of this proposal. My main reason for wanting this at the moment is to define it in graphviz, though I can think of some other uses I might have of it. If we all agree to define this type of class in a custom package (rather than me just defining it in graphviz and leaving it up to other users to define instances). I'd like to have this resolved sooner rather than later, as I want to work on graphviz soonish so I can start working on Graphalyze and SourceGraph again. -- Ivan Lazar Miljenovic Ivan.Miljenovic(a)gmail.com IvanMiljenovic.wordpress.com

3 3

combining monads with IO
by Richard Silverman 25 Jun '09

25 Jun '09

Hi all, I'm puzzled by something. Suppose I have some code that does lots of IO, and also occasionally refers to some global state. No problem, use ReaderT for the state, combining with the IO monad. Except... since IO is on the bottom, simple uses of do-notation such as "foo <- ask" work in the Reader monad, and to access the IO monad, I need to lift, e.g. (bar <- liftIO getLine). If my code does lots of IO, this is *very* ugly -- the code is littered with lift functions! Is there no cleaner way to do this? Thanks, - Richard

5 6

Re: [Haskell-cafe] GHCi infers a type but refuses it as type signature
by papa.eric＠free.fr 25 Jun '09

25 Jun '09

> Simple: the definition of MonadState uses those extensions. Thanks, yes it helps and explains all. :^) I suppose then that if -XFlexibleContexts is indeed required by the standard libraries, it is a "safe" extension, meaning supported by all compilers? Are many such extensions de-facto standard that anyone can enable by default?

6 8

[ANN] full-sessions: yet another implementation of session types
by Keigo Imai 25 Jun '09

25 Jun '09

Hello haskell-cafe, I would like to announce the pre-release of full-sessions, yet another implementation of session types in Haskell. Session types are used to statically check the safe and consistent use of communication channels according to protocols. Our work is quite similar to the Session monad in Haskell'08 paper of Pucella and Tov, however, a notable advantage of our implementation is that it requires almost no type annotation or term annotation in a source code, and at the same time provides full functionality of session types including channel-generation and channel-passing. It can be obtained from the website. A few examples are also provided. http://www.agusa.i.is.nagoya-u.ac.jp/person/sydney/full-sessions.html Any comments are much appreciated. Cheers, - Keigo IMAI

2 2

Re: [Haskell-cafe] unique identity and name shadowing during type inference
by Geoffrey Irving 25 Jun '09

25 Jun '09

Thanks. I'll go with the monad for now. Geoffrey On Sat, Jun 20, 2009 at 4:40 PM, Lennart Augustsson<lennart(a)augustsson.net> wrote: > Use 1. You'll probably need a monad in the type checker soon or later > anyway, e.g., for handling errors. On Sun, Jun 21, 2009 at 5:13 AM, Zsolt Dollenstein<zsol.zsol(a)gmail.com> wrote: > I think you should also take a look at the value-supply package. At least > that was my solution for scope analysis issues.

1 0

Parsing .dot files?
by Lee Pike 25 Jun '09

25 Jun '09

Hi, Does anybody know of a Haskell library for parsing .dot graph files? (I know Andy Gill wrote dotgen for *generating* .dot files <http://hackage.haskell.org/package/dotgen >.) Thanks, Lee -- <http://www.cs.indiana.edu/~lepike/>

3 2

Can you determine a constructor's arity using Data.Typeable and Data.Data?
by David Fox 25 Jun '09

25 Jun '09

Is it possible to determine the arity of a value's constructor? Suppose I have a value x of type data A = B Int | C They typeOf function returns its TypeRep, which contains its type constructor, but I don't see how to decide whether that constructor's arity is 0 or 1. If the type has field names I can look at those using Data.Data.conFields and count them, but if it doesn't I don't see how to do it. -david

2 2

Using unsafePerformIO safely
by gladstein＠gladstein.com 25 Jun '09

25 Jun '09

Regarding how to make the show instructions cause printout as soon as they are executed: If you write your interpreter to return a list of printout lines (strings), you get this behavior for free. Haskell's laziness enables the printing to start right away, while in an imperative language the list of strings wouldn't be returned until it was complete. To see this in action, try this program: main = mapM putStrLn myList myList = ["first", "second", loop, "third"] loop = loop Despite the infinite loop, "first" and "second" print out.

1 0

Another question about unsafePerformIO
by Matthias Görgens 25 Jun '09

25 Jun '09

I have a program that optimizes train schedules. It employs an external solver for Integer Linear Programs. The solve function has the following type: > solve :: Constraints -> IO (Maybe Solution) And this works. However, my external solver also behaves like a pure function from input to output. I wonder whether this guarantee should be reflected in the type system. I'd also appreciate if the compiler would be able to eliminate some calls to the solver. > solvePure :: Constraints -> Maybe Solution > solvePure = unsafePerformIO . solve Is this a good idea?

4 4

How to fork a process?
by Magicloud Magiclouds 25 Jun '09

25 Jun '09

Hi, In many language, both thread and process are supported. But in haskell's document, the only thing I could find called fork is to make a thread. So how to fork the program itself, like fork () in C? Thanks. -- 竹密岂妨流水过山高哪阻野云飞

2 1

Code walking off the right edge of the screen
by Deniz Dogan 25 Jun '09

25 Jun '09

I (too) often find myself writing code such as this: if something then putStrLn "howdy there!" else if somethingElse then putStrLn "howdy ho!" else ... I recall reading some tutorial about how you can use the Maybe monad if your code starts looking like this, but as you can see, that doesn't really apply here. "something" and "somethingElse" are simply booleans and each of them have different actions to take if either of them is True. So how do I make code like this prettier? Thanks, Deniz Dogan

14 16

Using unsafePerformIO safely
by Hector Guilarte 25 Jun '09

25 Jun '09

Thanks for answering so fast. Yes, GCL == Guarded Command Language... It is for an assigment I have in my Languages and Machines Course. About the nicer/Haskellier solution you proposed: If there is a way of printing right in the moment the Interpreter finds the show instruction then I don't think the teacher is gonna like this soluttion (I would bet on that) so, can you (or somebody) explain a little bit better how would the ugly solution be? As I said earlier, I'm no expert in monads, actually the truth is that I know almost nothing about monads, so please, explain it to me as if you are explaining it to a Monads newbie... Also, can somebody explain me how would it be using the Writer Monad? remember is for a Monads newbie... Thanks a lot! Hector Guilarte On Thu, Jun 25, 2009 at 6:04 PM, Jochem Berndsen <jochem(a)functor.nl> wrote: > Hector Guilarte wrote: > > I made a GCL compiler using Alex and Happy and now I'm making the > > interpreter to that program. Here's the deal: > > > > > First of all, I'm no expert in the usage of monads. Now: > > > > Whenever a "show" instruction is found in any GCL program while the > > interpretation is being done it is supposed to print on the stdout the > > string or the aritmetic expresion it was called with, so I guessed I need > to > > run an IO operation and continue the interpretation of my program. I > managed > > to do this using unsafePerformIO and `seq` like is shown below. My > question > > is: Is it safe to use it this way? So far it is working great, but I need > to > > be sure I'm using it in a "safe" way. Like I said, I'm no expert in > monads > > and the System.IO.Unsafe documentation says: > > > > " > > *unsafePerformIO* :: > > IO< > http://www.haskell.org/ghc/docs/latest/html/libraries/base/System-IO.html#t… > >a > > -> a > > This is the "back door" into the > > IO< > http://www.haskell.org/ghc/docs/latest/html/libraries/base/System-IO.html#t… > >monad, > > allowing > > IO< > http://www.haskell.org/ghc/docs/latest/html/libraries/base/System-IO.html#t… > >computation > > to be performed at any time. For this to be safe, the > > IO< > http://www.haskell.org/ghc/docs/latest/html/libraries/base/System-IO.html#t… > >computation > > should be free of side effects and independent of its > > environment. > > " > > > > I don't know if the IO computation I'm doing is free of side effects and > > independent of its enviroment :s. (is just hPutStr stdout ....) > > Well, writing to the standard output is certainly a side effect. (This > does not mean that you cannot use unsafePerformIO. The compiler, > however, may assume that any value is free from side effects. This means > that you could get, in theory, less or more output from your program > than you want. In this sense it is not "safe".) > > > Also I've read something about my code not being executed for sure or > > something like that. Can somebody check the code and tell me if I'm > "safe" > > with it? > > It's "safe" in the sense that it probably won't blow up your computer. > It may also work. On the other hand, I would not recommend using > unsafePerformIO in this way. > > I see two possibilities for resolving this issue: > * (ugly) run your GCL (Guarded Command Language?) interpreter in the IO > monad, and using "print"/"putStr"/... whenever you encounter a 'show' > statement in the GCL program. > * (nicer/Haskellier) adapt your interpreter such that it returns a list > of Strings to output. You have then a purely functional interpreter, and > in the main function of your program you can print this list. This will > be lazily evaluated as the GCL program runs. You now have a very nice > separation of clean, pure code, and impure code in the IO monad (your > "main" function, which can be pretty small in your case). To avoid > boilerplate, you can use the Writer monad, for example, but others may > have better suggestions. > > Kind regards, > > -- > Jochem Berndsen | jochem(a)functor.nl > GPG: 0xE6FABFAB >

5 8

Error in array index.
by Jason Dusek 25 Jun '09

25 Jun '09

I'm using tuples to index into a `UArray` created like this: fromList :: [(Word16, Word16)] -> UArray (Word16, Word16) Bool fromList list = runSTUArray $ do arr <- empty return arr empty :: ST s (STUArray s (Word16, Word16) Bool) empty = newArray ((0,0), (0xFFFF,0xFFFF)) False You'll notice this example doesn't do much; I'm not able to get much mileage so far. For example: > Array16Bit.fromList [(1,2), (2,3)] array ((0,0),(65535,65535)) [((0,0),*** Exception: Error in array index What do I need to do to debug this? -- Jason Dusek

6 13

Re: [Haskell] ANNOUNCE: jhc 0.6.1
by David Barton 25 Jun '09

25 Jun '09

Switching to Haskell Cafe; I hope you read that list, John, since it seems more suitable to this kind of question. John Meacham wrote: > Hi, this is to announce the release of jhc 0.6.1. The jhc homepage with > distribution information is at http://repetae.net/computer/jhc/ > > The main new feature in this release is a much simplified > cross-compilation mechanism. While cross-compilation was always possible > with jhc, it used to involve manually copying the C file and calling gcc > with the right options on it, now this is taken care of by jhc. > > A (popular) example would be setting up an iPhone cross compilation > target. For instance with the SDK setup I have, I would simply add the > following to a file ~/.jhc/targets.ini > > [iphone] > cc=arm-apple-darwin > cflags+=-I/usr/local/arm-apple-darwin/include > merge=le32 > > then you can compile iphone binaries with > > ; jhc --cross -miphone HelloWorld.hs > > the targets mechanism is extensible at run-time and I have included > native unix, win32, osx-intel and osx-powerpc targets. But certainly > many more interesting ones are possible. Some I have tested have been a > nokia N770 as a target and an atheros MIPS based router running dd-wrt. Maximum coolness! When you were targeting the nokia, how did you handle the radically different user interface? Did you have to establish a mapping from one of the Haskell UI packages to the Nokia equivalents? If so, which one did you pick, and how much time did it take? I'd love it if I could get this working for the Palm.... Dave Barton (soon to be) University of Toronto

2 1

Optimizing spelling correction program
by Kamil Dworakowski 25 Jun '09

25 Jun '09

Hi all, I want to learn how to optimize Haskell code on the Norvig's spelling correction program [1]. Peter Norvig has written the program in Python, quite a literate translation to Haskell (thanks to Grzegorz Chrupala) [2] was also available. Both versions are about 23 LOCs, and Haskell version is four times slower, 2m25s versus 36s. All the numbers I give come from running the program on a list of 806 misspellings, Haskell version compiled with - O2 and -fforce-recomp flags. I started with trial and error approach. Just preventing a repetitive call to keysSet brought the running time down to 1m48s. I also restructured the edits1 function and dropped all uses of sets, which haven't been pulling their own weight. This brought the running time down to 55s. Not quite there yet but close. Reading the Profiling chapter in the RWH book enabled me to do some more informed optimizing. At this point the GC time was about 2% of the overall runnig time, Grzegorz has done a good job of using strict versions of functions where necessary. The timing per expression however showed something useful, 70% of the time was being spent around determining membership in the known words dictionary (Data.Map). There are about 30 000 items in the dictionary. Right... Python uses hashtables while here I have a tree with log n access time. I did not want to use the Data.HashTable, it would pervade my program with IO. The alternative is an ideal hashmap that never gets changed. This program creates a dictionary at start which then is only being used to read from: an ideal application for the Data.PerfectHash by Mark Wotton available on Hackage [3]. The PerfectHash is a dictionary keyed by ByteStrings, so I had to migrate the program to use these instead of Strings. Sadly it increased the running time to 59s. Adding PerfectHash brought the running time down to 39s. I have done some code restructuring with the view to publishing it, which brought the running time up to 45s. The code is available on patch-tag [4] for anyone to scrutinize and hopefully point out mistakes (it uses big.txt and misspellings file which are linked to from Norvig's page [1]). At this point [5] I don't know how to proceed. The program is still slower than Python. The -ddump-simpl dump does not give me any clues, there is too much of it and I don't understand most of it. The GC time is about 10% of the total, and according to the profiler almost half the time is spent in the member function, and almost all the rest in edits1. Shouldn't it run much faster than the interpreted Python? Frankly, I am not impressed with the PerfectHash performance; It looks as if it is only two times faster than Data.Map. Is this because of the number of elements? 1. http://www.norvig.com/spell-correct.html 2. http://pitekus.blogspot.com/2007/04/norvigs-spelling-corrector-in-haskell.h… 3. http://hackage.haskell.org/package/PerfectHash 4. http://patch-tag.com/r/spellcorrect/ 5. http://patch-tag.com/r/spellcorrect/snapshot/hash/20090621192727-a99e5-fed4…

8 23

Re: [Haskell-cafe] Suggestion for Network.Socket in regards to PortNumber
by Johan Tibell 24 Jun '09

24 Jun '09

On Mon, Jun 22, 2009 at 7:27 PM, Tim Sheard <sheard(a)cs.pdx.edu> wrote: > My idea is to use types to ensure that any > sequence of operations adheres to a path on > a finite state automata. For example > if we wanted to implement the following > automata (This needs to read in a fixed width font) > Here's the reply Oleg gave me when I asked about using monadic regions to enforce similar constraints (published with Oleg's permission): Oleg writes: > First of all, protocols of using API functions are essentially > session types, right? Thus various implementations of session types > could be taken advantage of (including the one presented by Jesse Tov > on Haskell 2008 Symposium). Aliasing is responsible for a lot of the > complexity: if the variable 's' is bound to a value of the type socket > and we execute "do s' <- bindS s" so to bind the socket, the old > variable s is still available and could be used. But it should not be > used because bindS "consumed" the unbound socket and produced the new > one. To properly describe these resource restrictions (a resource can > be consumed and after that it is not available any more), we need > linear, affine or various other kinds of substructural logics and type > systems. It seems they all can be emulated in Haskell, but with some > hackery (see for example Sec 6 of our monadic regions > paper). Parameterized monads are necessary. > > The lightweight monadic regions paper relied on a simple > session protocol: open -> (read+write)* -> close. That protocol can be > ensured without parameterized monads: the only aliasing to mind is > the accessibility of the file handle after the closing operation. We > can get rid of `close' (make it implicit ar region's exit) and use > rank-2 types to prevent file handles from leaking out. There is some > dynamic overhead, in the function open. However, the most frequent > operation, reading and writing from the handle, involves no overhead > and its correctness ensured statically. > > That design can be extended to sockets. For example: > > data Socket sort = Socket Int > > data UnconnectedSocket > data ConnectedSocket s > > connect :: RMonadIO (m s) => SocketProto -> SAddress -> > m s (Socket (ConnectedSocket s)) > > sendto :: RMonadIO (m s) => (Socket (ConnectedSocket s)) -> Buffer > -> m s (ReturnCode) > > -- polymorphic in the socket sort > send :: MonadIO m => Socket sort -> Buffer -> m ReturnCode > > > In this design, the operation of creating a socket and connecting it > are bundled together -- which is very common and is the case for > System V, I think (where it is called openActive, I think). If you for > some reason want to keep 'socket' and 'connect' separate, then > 'connect' has to check at run-time that the socket in question is not > connected yet. Still, the most common operations like send and write > should be overhead-free. Cheers, Johan

3 5

Using unsafePerformIO safely
by Hector Guilarte 24 Jun '09

24 Jun '09

Hello, I made a GCL compiler using Alex and Happy and now I'm making the interpreter to that program. Here's the deal: First of all, I'm no expert in the usage of monads. Now: Whenever a "show" instruction is found in any GCL program while the interpretation is being done it is supposed to print on the stdout the string or the aritmetic expresion it was called with, so I guessed I need to run an IO operation and continue the interpretation of my program. I managed to do this using unsafePerformIO and `seq` like is shown below. My question is: Is it safe to use it this way? So far it is working great, but I need to be sure I'm using it in a "safe" way. Like I said, I'm no expert in monads and the System.IO.Unsafe documentation says: " *unsafePerformIO* :: IO<http://www.haskell.org/ghc/docs/latest/html/libraries/base/System-IO.html#t…>a -> a This is the "back door" into the IO<http://www.haskell.org/ghc/docs/latest/html/libraries/base/System-IO.html#t…>monad, allowing IO<http://www.haskell.org/ghc/docs/latest/html/libraries/base/System-IO.html#t…>computation to be performed at any time. For this to be safe, the IO<http://www.haskell.org/ghc/docs/latest/html/libraries/base/System-IO.html#t…>computation should be free of side effects and independent of its environment. " I don't know if the IO computation I'm doing is free of side effects and independent of its enviroment :s. (is just hPutStr stdout ....) Also I've read something about my code not being executed for sure or something like that. Can somebody check the code and tell me if I'm "safe" with it? Thanks a lot! Here's the code: -- Tabla is my Symbol Table of the program being interpreted evalInstruccion:: Instruccion -> Tabla -> Tabla evalInstruccion (ShowY showY) tabla = myRunIO (evalShow showY tabla) evalInstruccion _ tabla = tabla -- There are many other Instructions here missing wich are not relevant to my question {-# NOINLINE myRunIO #-} myRunIO:: (Tabla, IO()) -> Tabla myRunIO tupla = ((unsafePerformIO (snd tupla)) `seq` (fst tupla)) -- Here's the unsafePerformIO.... Am I safe? evalShow:: ShowY -> Tabla -> (Tabla, IO()) evalShow (ShowS string) tabla = (tabla,(hPutStr stdout string)) evalShow (ShowE expr) tabla = (tabla,(hPutStr stdout (show (evalExpr expr tabla)))) -- Don't worry about evalExpr, it works and returns an Int

2 1

Which one is right?
by Linker 24 Jun '09

24 Jun '09

Hi,All, Construct a list: [0,0.1..1] ==================================== || || || || || || ||__ Hugs 98: Based on the Haskell 98 standard ||___|| ||__|| ||__|| __|| Copyright (c) 1994-2005 ||---|| ___|| World Wide Web: http://haskell.org/hugs || || Bugs: http://hackage.haskell.org/trac/hugs || || Version: September 2006 _________________________________________ Haskell 98 mode: Restart with command line option -98 to enable extensions Type :? for help Hugs> [0,0.1..1] [0.0,0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0] Hugs> ==================================== GHCi, version 6.8.2: http://www.haskell.org/ghc/ :? for help Loading package base ... linking ... done. Prelude> [0,0.1..1] [0.0,0.1,0.2,0.30000000000000004,0.4,0.5,0.6,0.7,0.7999999999999999,0.8999999999999999,0.9999999999999999] Prelude> ==================================== Thanks. -- Regards, Linker Lin linker.m.lin(a)gmail.com

5 4

Suggestion for Network.Socket in regards to PortNumber
by John Van Enk 24 Jun '09

24 Jun '09

Hello List, Is there any one else in favor of hiding the PortNum constructor in Network.Socket in the next release? Here's why: Prelude> :m + Network.Socket Data.Word Prelude Network.Socket Data.Word> let p = 5000 :: Word16 Prelude Network.Socket Data.Word> let p' = PortNum p Prelude Network.Socket Data.Word> let p'' = fromIntegral p :: PortNumber Loading package parsec-2.1.0.1 ... linking ... done. Loading package network-2.2.1.1 ... linking ... done. Prelude Network.Socket Data.Word> p 5000 Prelude Network.Socket Data.Word> p' 34835 Prelude Network.Socket Data.Word> p'' 5000 Notice that using the PortNum constructor does not at all do what the user expects. This really should be a hidden constructor. Perhaps we can add the following: mkPortNum :: (Num a) => a -> PortNumber mkPortNum p = fromIntegral p /jve

4 12

Re: [Haskell-cafe] GHCi infers a type but refuses it as type signature
by papa.eric＠free.fr 24 Jun '09

24 Jun '09

> That depends on the monad library. In transformers it would be: > > (Monad m, Eq a) => a -> StateT [a] m Bool > > State s is just a type synonym for StateT s Identity, so this works for State as well. I like it, it is simpler than the one using MonadState. I had read many tutorials and wiki articles, and saw nowhere that there were different implementations of what seemed to be standard libraries. > On a higher level, in case you are interested, here's a description of > how I would model your problem. Actually using a monad transformer was the goal of it all, tutorial-like. The example was contrived. Another question would be: when is it relevant to use monad transformers? I'm not sure to have enough haskell practice to be ready for that one, though. :^) And for sure, my example didn't need it. Thanks Eric

2 1

Haskell on the iPhone
by Ryan Trinkle 24 Jun '09

24 Jun '09

Dear Haskellers, Recently, there's been a groundswell of activity in the Haskell community regarding the Haskell's use in developing iPhone games. The iPhone is a powerful, innovative platform (with a great monetization scheme, to boot), and it's not surprising that many of us would want to develop apps for it in our favorite language. I am proud to announce today that my company, iPwn Studios Inc., is currently preparing to release an open source patch to GHC that allows it to output binaries for iPhone OS. The patch will be released under a BSD license as soon as possible and hopefully integrated into the GHC main-line in the near future. As the first (to my knowledge) Haskell-based game studio, iPwn Studios is committed to giving back to the Haskell community through open source - contributing to a rising tide that lifts us all. I would like to take this opportunity to propose the creation of a haskell-iphone mailing list, so that all Haskellers working with the iPhone - whether for profit or for pleasure - can come together to make Haskell a force to be reckoned with in the burgeoning iPhone App marketplace. Best wishes, Ryan Trinkle President, iPwn Studios Inc. P.S.: If you wish to be involved in the preparation of the GHC patch or in the creation of iPwn Studios' first game, don't hesitate to contact me by email (ryant5000(a)gmail.com) AIM (RyanT5000), or IRC (RyanT5000 on irc.freenode.net)

10 15

rewrite rules
by Sjoerd Visscher 24 Jun '09

24 Jun '09

Hi all, I have a rewrite rule as follows: {-# RULES "transform/transform" forall (f::forall m. Monoid m => (a -> m) -> (b - > m)) (g::forall m. Monoid m => (b -> m) -> (c -> m)) (l::FMList c). transform f (transform g l) = transform (g.f) l #-} It fires on this code: print $ transform (. (*2)) (transform (. (+1)) (upto 10)) But it doesn't fire on this code: print $ map (*2) (map (+1) (upto 10))) with map g x = transform (. g) x and with or without {-# INLINE map #-}. What am I doing wrong? -- Sjoerd Visscher sjoerd(a)w3future.com

3 6

ICFP contest
by Rafael Gustavo da Cunha Pereira Pinto 24 Jun '09

24 Jun '09

Hi, The good folks here are really quiet about the ICFP Contest. Is anyone participating this year? As in the previous four years, I will probably download the problem, take a look, and try for a while... Having a family don't allow me to go much further, especially when teaming up with myself! Best regards, Rafael Gustavo da Cunha Pereira Pinto

4 4

ANN: darcs 2.3 beta 1
by Petr Rockai 24 Jun '09

24 Jun '09

Hi! I'd like to announce immediate availability of a first beta release of darcs 2.3. There is a number of improvements and bugfixes over the last stable release (2.2). Moreover, work has been done on performance of "darcs whatsnew" for large repositories. This has also introduced a slight risk of regressions, but please note that all of the disruptive changes are in read-only code paths: the new code will never touch your repository, so it is unable to cause permanent harm. The worst that could happen is that you get no or bad diff from "darcs whatsnew". (This is also a reason why we need your testing!) There is only a single installation package for this release of darcs: cabalised source. (Please note that the final version with also come with the legacy autoconf-based buildsystem, for the last time.) You can either download a tarball from <http://repos.mornfall.net/darcs/darcs-2.2.98.1.tar.gz> and build manually (see the build instructions in README inside the tarball), or, alternatively, you can use cabal-install to obtain a copy (the beta release is now available on Hackage): $ cabal update $ cabal install darcs-beta This should give you a darcs binary in `~/.cabal/bin` -- you should probably add that to your PATH. (Note: The package name on Hackage is different, since people installing darcs from Hackage are expecting a stable version. The name change means that you cannot use the Hackage version to build other packages that depend on the darcs library: you either need the tarball for this, or you can use the stable version from Hackage.) There is a quick (possibly incomplete) list of important changes: * lots and lots of documentation changes (Trent) * haskeline improvements (Judah) * cabal as default buildsystem (many contributors) * fixes in darcs check/repair memory usage (Bertram, David) * performance improvement in subtree record (Reinier) * --summary --xml (Florian Gilcher) * changes --max-count (Eric and myself) * fix changes --only-to-files for renames (Dmitry) * performance fix in "darcs changes" (Benedikt) * hardlinks on NTFS (Salvatore) * coalesce more changes when creating rollbacks (David) * new unit test runner (Reinier) * darcs-shell in contrib (László, Trent) * .authorspellings (Simon) -- I find this to be controversial though * working directory index and substantial "darcs wh" optimisation (myself) * gzip CRC checking and repair feature (Ganesh) and there is a number of issues that has been resolved since 2.2: * 948: darcsman (Trent) * 1206: countable nouns (Trent) * 1285: cabal test v. cabal clean (Trent) * 1302: use resolved, not resolved-in-unstable (Trent) * 1235: obliterate --summary (Rob) * 1270: no MOTD for --xml-output (Lele) * 1311: cover more timezones (Dave) * 1292: re-encoding haskeline input (Judah) * 1313: clickable ToC and refs in PDF manual Trent) * 1310: create merged \darcsCommand{add} (Trent) * 1333: better "cannot push to current repository" warning (Petr) * 1347: (autoconf) check for unsafeMMapFile if mmap use enabled (Dave) * 1361: specify required includes for curl in cabal file (Reinier) * 1379: remove libwww support (Trent) * 1366: remove unreachable code for direct ncurses use (Trent) * 1271: do not install two copies of darcs.pdf (Trent) * 1358: encode non-ASCII characters in mail headers (Reinier) * 1393: swap "darcs mv" and "darcs move" (Trent) * 1405: improve discoverability of global author file (Trent) * 1402: don't "phone home" about bugs (Trent) * 1301: remove obsolete zsh completion scripts (Trent) * 1162: makeAbsolute is now a total function (Ben F) * 1269: setpref predist - exitcode ignored bug (Ben M) * 1415: --edit-long-comment, not --edit-description, in help (Trent) * 1413: remove duplicate documentation (Trent) * 1423: complain about empty add/remove (Trent) * 1437: Implement darcs changes --max-count (Eric) * 1430: lazy pattern matching in (-:-) from Changes command module (Dmitry) * 1434: refactor example test (Trent) * 1432: refer to %APPDATA%, not %USERPROFILE% (Trent) * 1186: give a chance to abort if user did not edit description file (Dmitry) * 1446: make amend-record -m foo replace only the patch name (Dmitry) * 1435: default to get --hashed from a darcs-1.0 source (Trent) * 1312: update and reduce build notes (Trent) * 1351: fix repository path handling on Windows (Salvatore) * 1173: support hard links on NTFS (Salvatore) * 1248: support compressed inventories for darcs-1 repos (Ganesh) * 1455: implement "darcs help environment" (Trent) Yours, Petr. -- Petr Ročkai | http://web.mornfall.net A physicist is an atom's way of knowing about atoms. (George Wald)

1 0

Forward compatibility (was Re: [Haskell-cafe] GHCi infers a type but refuses it as type signature)
by Luke Palmer 24 Jun '09

24 Jun '09

On Tue, Jun 23, 2009 at 2:20 AM, <papa.eric(a)free.fr> wrote: > > Simple: the definition of MonadState uses those extensions. > > Thanks, yes it helps and explains all. :^) > > I suppose then that if -XFlexibleContexts is indeed required by the > standard libraries, it is a "safe" extension, meaning supported by all > compilers? Are many such extensions de-facto standard that anyone can enable > by default? You should treat "all compilers" the same way as a universal quantifier in mathematics. It's not just all the compilers you know about, but all possible compilers. In which case the answer to your question is of course no. (Beware: incoming my "recurring rant of the month", which I am trying ferociously to cram down the throats of my peers :-p) By measuring your software's portability on this stricter criterion, you contribute to the exponential growth of our field. It means that *every *future compiler can compile your code for 10% less effort*. *A future innovation may rely on one such compiler, which means it can use your code 10% sooner -- which may be the difference between your good code being used and it being badly reimplemented (and future bug fixes in the new code does not benefit yours). And so on, ad infinitum. When you think about potability, think about it from the perspective of twenty years in the future. Fundeps are becoming obsolete now, will your code still work when they are gone? Will it still work when the typeclass resolution algorithm is obsoleted by a superior algorithm (I'm looking at you, Oleg! :-)? When Haskell is obsolete, how hard will it be to port? Of course, there is definitely a place for extensions -- we should be testing them, to help refine the language that will one day obsolete Haskell. A culture of extension users helps us to know that, for example, even though implicit params are cool, nobody cares. I think "works on all existing compilers" is a fine definition of "portable" -- let's just realize that there is a level of portability beyond that, a level which gives our whole field a little boost. </rant> Luke

4 6

What happened to the installation chapter in the GHC User's Guide?
by Benjamin L.Russell 24 Jun '09

24 Jun '09

Up to GHC 6.10.2, there used to be an installation chapter in the GHC User's Guide (see "Chapter 2. Installing GHC" at http://www.haskell.org/ghc/docs/6.10.2/html/users_guide/installing-bin-dist…) This section was very useful, for example, in determining whether it was possible to move the GHC tree around in Windows, or, by extension, whether it was necessary to uninstall GHC in order to install a later version. However, starting with GHC 6.10.3, this chapter seems to be missing. When I asked about this issue earlier on this mailing list a few months ago, IIRC, one response was that the chapter was outdated. Are there any plans to update this chapter? Every once in a while, I forget whether GHC needs to be uninstalled in order to be updated on Windows, and am forced to consult an earlier version of this chapter. -- Benjamin L. Russell -- Benjamin L. Russell / DekuDekuplex at Yahoo dot com http://dekudekuplex.wordpress.com/ Translator/Interpreter / Mobile: +011 81 80-3603-6725 "Furuike ya, kawazu tobikomu mizu no oto." -- Matsuo Basho^

2 1

Weird thing happened when I use gtk2hs with concurrent.
by Magicloud Magiclouds 24 Jun '09

24 Jun '09

This is the weird code, please see line 61. 20 cpuNew :: IO CPU 21 cpuNew = do 22 self <- meterNew 23 usage <- newIORef $ CPUusage 0 0 0 0 0 0 0 0 24 self `on` realize $ do 25 forkIO $ redraw self usage 33 return () 34 return self 35 where 56 redraw self usage = do 57 newUsage <- getUsage 58 oldUsage <- readIORef usage 59 postGUISync $ meterDraw self "Cpu" $ computeIdle $ diff newUsage oldUsage 60 writeIORef usage newUsage 61 threadDelay 10000 -- ten seconds, but it just refresh crazy! 62 redraw self usage

2 1

Using Parsec with ByteString ?
by Fernand 24 Jun '09

24 Jun '09

Hi, May be it's a frequently asked question, but I have been unable to find my way in the parsec-3.0.0 library when trying to write a small parser with Parsec that could work with ByteStrings. I mean, there is the Text.Parsec.ByteString module which allows to parse a file using ByteStrings, but the parser one needs to write must parse ByteStrings instead of Strings (that is, something like having a "Parsec Bytestring ()" type, unless I'm completely misunderstanding the situation). My problem is that I do not manage to write primitive parsing combinators (like string, satisfy, letter, etc.) to define my language's tokens. I'm surely missing some primitive somewhere (may be the Text.Parsec.Prim.token, though I do not really manage to use it). Thanks in advance, F.

2 2

Type system trickery
by Andrew Coppin 24 Jun '09

24 Jun '09

I have a datatype with about a dozen constructors. I'd like to find a way to use the type system to prevent one of the constructors from being used in certain places. But I can't think of a way to do that. data Foobar = Foo Foobar | Bar Foobar | Zoo Foobar I want the type system to track whether or not Zoo has been used in a specific value. Sure, you can check for it at runtime, but I'd be happier if the type system can guarantee its absence when required. I tried this: data Zoo x = Zoo x | NoZoo x data GeneralFoobar f = Foo f | Bar f type Foobar = GeneralFoobar Foobar type FoobarZ = GeneralFoobar (Zoo FoobarZ) but the type checker seems to not like it. (Plus I now have to wade through miles of NoZoo wrappers everywhere.) The other alternative is to just duplicate the entire data declaration, sans the Zoo constructor... but I'd obviously prefer not to have to do that. (E.g., if I ever need to alter the declaration, I now have two copies to keep in sync.) Suggestions?

8 17

(fwd) Haskell logo fail
by Benjamin L.Russell 24 Jun '09

24 Jun '09

On Tue, 16 Jun 2009 14:06:42 -0700 (PDT), in comp.lang.haskell Robert Hicks <sigzero(a)gmail.com> wrote: >http://blog.plover.com/prog/haskell/logo.html >From the site referenced at the above-mentioned URL: >Tue, 16 Jun 2009 > >Haskell logo fail >The Haskell folks have chosen a new logo. > ><image of new Haskell logo> > ><image of Amtrak logo> > >ouch The two logos do look disturbingly similar.... -- Benjamin L. Russell -- Benjamin L. Russell / DekuDekuplex at Yahoo dot com http://dekudekuplex.wordpress.com/ Translator/Interpreter / Mobile: +011 81 80-3603-6725 "Furuike ya, kawazu tobikomu mizu no oto." -- Matsuo Basho^

6 6

How to determine if a FilePath is a directory name or regular file?
by Colin Paul Adams 24 Jun '09

24 Jun '09

I've been hoogling like bad to try to determine if a function like this exists. getDirectoryContents returns sub-directories as well as file names. I want only the latter, so I'm looking for a suitable filter. -- Colin Adams Preston Lancashire

9 15

Generate Fortran 77 parser from a yacc grammar using Happy
by Cetin Sert 24 Jun '09

24 Jun '09

I have stumbled upon the following F77 yacc grammar: http://yaxx.cvs.sourceforge.net/viewvc/yaxx/yaxx/fortran/fortran.y?revision… . How can I make a Fortran 77 parser out of this file using Happy<http://www.haskell.org/happy/> ? Why is there some C?/C++? code in that .y file?

2 1

HaRe (the Haskell Refactorer) in action - short screencast
by Claus Reinke 24 Jun '09

24 Jun '09

I've heard that many Haskellers know HaRe only as a rumour. It has been many years since the original project finished, and HaRe hasn't been maintained for quite some time, so just pointing to the sources isn't quite the right answer. The sources are still available, and build with GHC 6.8.3 (I had to fix one lineending issue on windows, iirc, and copy one old bug fix that hadn't made it into the latest release), but there is currently noone with the time or funding for maintenance, fixing bugs, making releases, or ironing out practical issues. If anyone would provide funding, people to do the work could be found, but the effort would probably be better spent on reimplementing the ideas in a GHC / Cabal environment (instead of the Haskell'98 environment targetted by our "Refactoring Functional Programs" project). If you've got the funding, please get in touch - even a three month run could get something started at least!-) In principle, the project experiences and lessons learned are quite well documented at the project site http://www.cs.kent.ac.uk/projects/refactor-fp/ but that doesn't give anyone an idea of what working with HaRe was like. With the recent interest in screencasts, I thought I'd make a little demo, for archival purposes. Nothing fancy, using only features that were already present in HaRe 0.3 (end of 2004), and not all of those, on a tiny 2-module example (screenspace is a bit crowded to keep the text readable on YouTube). I hope it might give a rough idea of what the papers, slides and reports are talking about, for Haskellers who weren't around at the time: http://www.youtube.com/watch?v=4I7VZV7elnY For the old HaRe team, Claus ------------------- YouTube video description: HaRe - the Haskell Refactorer (a mini demo) [4:10] The Haskell Refactorer HaRe was developed in our EPSRC project "Refactoring Functional Programs" http://www.cs.kent.ac.uk/projects/refactor-fp/ Building on Programatica's Haskell-in-Haskell frontend and Strafunski's generic programming library, it supported module-aware refactorings over the full Haskell'98 language standard. Interfaces to the refactoring engine were provided for both Vim and Emacs (this demo uses HaRe via GVim on Windows). While HaRe has continued to see occasional contributions by students and researchers, who use its Haskell program transformation API as a platform for their own work, it is not currently maintained. As the Haskell environment marches on, this demo is meant to record a snapshot of what working with HaRe could be like when it still built (here with GHC 6.8.3). The lessons learned (note, eg, the preservation of comments, and the limited use of pretty-printing, to minimize layout changes) are well documented at the project site, and should be taken into account when porting the ideas to the GHC Api, or other Haskell frontends.

3 2

GHCi infers a type but refuses it as type signature
by Eric 24 Jun '09

24 Jun '09

Hi haskell helpers, I am new to haskell (but enthusiast). I have begun to play with State and StateT, but this very simple exercice has led me to a strange situation where GHCi recognises and accepts the type of a function but GHC won't allow it as a type signature. Here is the example (which is also a try at literate haskell). > import Control.Monad.State > > type Play = Char > type Game = [Play] -- a game is a series of plays Now, you play by issuing a char, and you win if you have played an already played char. That was my first "play" function, returning True if you win. > play :: Play -> State Game Bool > play p = do ps <- get > put (p:ps) > return $ p `elem` ps I tested this one this way: > play_abc :: State Game Bool > play_abc = do play 'a' > play 'b' > play 'c' > > play_abca :: State Game Bool > play_abca = play_abc >> play 'a' Now in GHCi, "runState play_abc []" yields (False, "cba") "runState play_abca []" yields (True, "acba") Good! I was happy, now trying StateT to add IO to get console input: > type IOGame = StateT Game IO > > run_io :: IOGame Bool -> IO (Bool, Game) > run_io x = runStateT x [] Now I wanted a version of "play" that reads a char, on the following model: > play_once :: IOGame Bool > play_once = do x <- liftIO getChar > play' x Of course I first tried to express play' using play, and failed. Eventually I copied-and-pasted play, only changing the type signature, and it worked: > play' :: Play -> IOGame Bool > play' x = do xs <- get > put (x:xs) > return $ x `elem` xs Now, on GHCi I was happy to type: "run_io $ play_once" or even "run_io $ play_once >> play_once >> play_once" However, how to avoid the code duplication? I just tried to remove the type signature, and yes, play2 suddenly works in both State Game and StateT Game IO: > play2 x = do xs <- get > put (x:xs) > return $ x `elem` xs > > play2_abc :: State Game Bool > play2_abc = play2 'a' >> play2 'b' >> play2 'c' > > play2_once :: IOGame Bool > play2_once = do x <- liftIO getChar > play2 x So what's the type of play2? :t play2 yields: (MonadState [a] m, Eq a) => a -> m Bool Wow, nice, I get it now! However if I try to add that type signature, or even the second, more specific one below, ghci fails (signatures commented out because of that): > -- play3 :: (MonadState [a] m, Eq a) => a -> m Bool > -- play3 :: (MonadState Game m) => Play -> m Bool > play3 x = do xs <- get > put (x:xs) > return $ x `elem` xs It seems that an extension is required: Non type-variable argument in the constraint: MonadState [a] m (Use -XFlexibleContexts to permit this) In the type signature for `play3': play3 :: (MonadState [a] m, Eq a) => a -> m Bool So how is it possible that GHCi can infer (and use) a type that you cannot use as signature? And is it really non standard to avoid such code duplication? I was a bit surprised that such simple example should require a compiler extension. Thanks! Eric

3 2