March 2009 - Haskell-Cafe

How to compare PortNumbers or Bug in Network.Socket?
by Stefan Schmidt 13 Mar '09

13 Mar '09

Hello, I want to compare two PortNumber-Values from the Network.Socket module and I think I've experienced some unexpected behavior of the derived ordering methods. I'm using the ghc-6.10.1 and the network library 2.2.0.1 on a x86 32-Bit machine. The following program creates two pairs of PortNumber values and compares them with each other. module Main(main) where import Network.Socket -- Block A p1 :: PortNumber p1 = fromIntegral 9000 p2 :: PortNumber p2 = fromIntegral 10000 -- Block B pp1 :: PortNumber pp1 = PortNum 9000 pp2 :: PortNumber pp2 = PortNum 10000 main :: IO () main = do putStrLn $ "A - " ++ (show $ p1 > p2) putStrLn $ "B - " ++ (show $ pp1 > pp2) On my machine, I get the result: A - True A - False At the first glance, it seemed to me, that there is a bug in the fromIntegral Implementation. But after I looked into the code, I saw that fromIntegral calls the system function "htons" to change the byte order from my machine to the network byte order, which on my machine is different. Because of this the values 9000 and 10000 are transferred, so that their ordering changes. Block B preservers this ordering, because the Constructor PortNum does not call htons, but since the byte order is wrong, I cannot use pp1 and pp2 to address port 9000 and 10000. As a work around, I could convert two PortNumbers back to Int-Values before comparing them, but I think the ordering functions for the PortNumber-Type do not work as expected. Or am I wrong? Stefan Schmidt

4 5

Re: [Haskell-cafe] Using a monad to decompose a function into functions
by Marcin Kosiba 13 Mar '09

13 Mar '09

On Thursday 12 March 2009, you wrote: > 2009/3/12 Marcin Kosiba <marcin.kosiba(a)gmail.com>: > > Hi, > > I'm doing a bit of research into mobility models and I'm currently > > exploring implementation language choices for the simulator > > (*snip*) > > The simulation algorithm requires expressing > > the node's mobility so that it is "stateless". The mobility model > > algorithm's type should be something like: > > mobility_model :: WorldState -> NodeState -> OtherInput -> (Action, > > NodeState) > > > > where Action can alter WorldState and the second NodeState is an altered > > input NodeState. I perform a form of speculative execution on > > mobility_model so sometimes I need to backtrack to a previous world and > > node state. This is all fairly simple stuff, and was just an > > introduction. What I do now is store an enum in NodeState and implement > > mobility_model as one big case statement. Yes, this is very imperative of > > me, I know. What I'd like to do is to express mobility_model, so that the > > code would look like: > > > > mobility_model world node input = do > > do_calculus > > emit_action > > if something > > then emit_action > > else emit_action > > do_calculus > > emit_action > > mobility_model world node input > > Hi, > > It seems you can use > http://hackage.haskell.org/packages/archive/mtl/latest/doc/html/Control-Mon >ad-State-Lazy.html Just have a look at the exemple : > > tick :: State Int Int > tick = do n <- get > put (n+1) > return n > > your code would become something like > mobility_model :: OtherInput -> State (WorldState,NodeState) () > mobility_model input = do > world <- gets fst > node <- gets snd > .... > let (world',node') = ... > put (world',node') ok, that solves part of the problem. what this doesn't solve is that somewhere between these lines (which corespond to emit_action in my example) > let (world',node') = ... > put (world',node') I need to return a value and an Action and NodeState to the simulation algorithm. and then, after the simulation algorithm calculates a new WorldState it will want the mobility_model to where it left off, but with a new WorldState. I hope I'm clear about what I wish to achieve: each emit_action should return a value (Action, NodeState) and maybe a function mobility_model_cont which I then could call with the new WorldState to continue from where emit_action returned. Thanks, Marcin Kosiba

2 1

ANNOUNCE: Vintage BASIC 1.0
by Lyle Kopnicky 13 Mar '09

13 Mar '09

I am pleased to announce the initial release of Vintage BASIC, an interpreter for microcomputer-era BASIC. Fully unit-tested, it faithfully implements the common elements of the language. On the web site, you can find 102 games from the classic book BASIC Computer Games, all of which run flawlessly. Have fun! This is a standalone interpreter, operating on text files. Although not an embedding, like Lennart Augustsson's clever implementation, it does use a custom BASIC monad in order to execute the code. A unique feature of this implementation is that control structures such as FOR are implemented using resumable exceptions: FOR is a handler and NEXT throws an exception. A Developer's Guide is included with the source. This is my first public release of open source software. I have been working on this project since 2003. Home page: http://www.vintage-basic.net Also available on Hackage<http://hackage.haskell.org/cgi-bin/hackage-scripts/package/vintage-basic>and patch-tag.com <http://patch-tag.com/repo/vintage-basic/home>. Please e-mail me at lyle(a)vintage-basic.net with any questions/comments. - Lyle

2 1

Distributing Linux binaries
by Lyle Kopnicky 13 Mar '09

13 Mar '09

Hi folks, I've got an application to release. I'm releasing the source, but I also wanted to release binary versions for people that don't have GHC. I developed on Windows, so making a Windows executable was simple. I also have access to an Ubuntu Linux box, on which I can easily build and test my app. But, I'm a bit confused about how to release binaries for Linux. It's not so simple as on Windows. Or, maybe it is but I don't know it. What do people recommend as a way of distributing Linux binaries? I tried to make a Debian package and couldn't figure it out, but maybe that's overkill. GHC itself has some kind of tarball for the binary distributions, one for Debian and one for Red Hat. How would I make such a package and what would go into it? There's only one executable, so it shouldn't be too complicated. I figure the main issue on Linux is that when you compile it, it's linked to specific shared libraries. Don't have a Mac myself but I'm happy to have Mac folks submit binaries. I'll worry about that later. - Lyle

6 8

Natural Numbers: Best implementation?
by Mark Spezzano 12 Mar '09

12 Mar '09

Hi, I was wondering what the best way to implement Natural number would be. Is there a package which already does this? Here are some options: 1. Don’t bother. Just use Integer. 2. Use the type data Natural = Zero | Succ !Natural 3. Use the following definition taken from the Gentle Introduction to Haskell 98 newtype Natural = MakeNatural Integer toNatural ::Integer-> Integer toNatural x | x < 0 = error “Can’t create negative naturals!” | otherwise = MakeNatural x fromNatural :: Natural -> Integer fromNatural (MakeNatural i) = i and then... instance Num Natural where fromInteger = toNAtural x + y = toNatural (fromNatural x + fromNatural y) x – y = etc.. x * y = etc... Which method is best? So far, I’ve been picking option #1 – just leaving things as they are and using Integer to keep things simple. I’ve got that feeling that [2] would be fast and [3] would be slow. Comment appreciated on the merits of each. Cheers, Mark Spezzano No virus found in this outgoing message. Checked by AVG. Version: 7.5.557 / Virus Database: 270.11.12/1998 - Release Date: 12/03/2009 6:23 PM

1 0

Combining haddock documentation from multiple packages
by Heikki Aitakangas 12 Mar '09

12 Mar '09

Is it possible to generate a "unified" documentation tree similar to http://www.haskell.org/ghc/docs/latest/html/libraries/index.html from packages installed via cabal from hackage? And if so, how does one go about doing it? -- Heikki Aitakangas

1 0

Using a monad to decompose a function into functions
by Marcin Kosiba 12 Mar '09

12 Mar '09

Hi, I'm doing a bit of research into mobility models and I'm currently exploring implementation language choices for the simulator (yes, sadly it needs to be a custom one). I've been using Haskell here and there for some small tasks, and thought I should consider it as an implementation language for the simulator. While I already have an working implementation in Haskell, there is one thing that I would like to express in a more elegant way, but just can't figure out. The simulation algorithm requires expressing the node's mobility so that it is "stateless". The mobility model algorithm's type should be something like: mobility_model :: WorldState -> NodeState -> OtherInput -> (Action, NodeState) where Action can alter WorldState and the second NodeState is an altered input NodeState. I perform a form of speculative execution on mobility_model so sometimes I need to backtrack to a previous world and node state. This is all fairly simple stuff, and was just an introduction. What I do now is store an enum in NodeState and implement mobility_model as one big case statement. Yes, this is very imperative of me, I know. What I'd like to do is to express mobility_model, so that the code would look like: mobility_model world node input = do do_calculus emit_action if something then emit_action else emit_action do_calculus emit_action mobility_model world node input but I'd like to be able to alter world and node state before continuing from emit_action. I've tried to get this working by using the idea from http://www.haskell.org/pipermail/haskell/2005-April/015684.html but couldn't get the state-altering behavior I was looking for. I've also taken a look at http://monadicheadaches.blogspot.com/2008/01/python-25s-iterators-in-haskel…, the unified concurrency model and Control.Coroutine, but couldn't get the behavior I was going for. Thanks! Marcin Kosiba

2 1

DSLs with {in,}equalities
by Andrew Hunter 12 Mar '09

12 Mar '09

Several times now I've had to define an EDSL for working with (vaguely) numeric expressions. For stuff like 2*X+Y, this is easy, looking pretty much like: > data Expr = Const Integer | Plus Expr Expr | Times Expr Expr > > instance Num Expr where > fromInterger = Const > (+) = Plus > (*) = Times &c. This lets me get a perfectly nice AST, which is what I want. When I want to be able to express and work with inequalities and equalities, this breaks. Suppose I want to write 2*X + Y < 3. I either have to: a) Hide Prelude.(<) and define a simple < that builds the AST term I want. b) Come up with a new symbol for it that doesn't look totally awful. Neither of these work decently well. Hiding Eq and Ord operators, which is what I effectively have to do for a), is pretty much a nonstarter--we'll have to use them too much for that to be practical. On the other hand, b) works...but is about as ugly as it gets. We have lots and lots of symbols that are already taken for important purposes that are syntactically "near" <,<=,==, and the like: << and >> and >>= for monads, >>> for arrows, etc. There...are not good choices that I know of for the symbols that don't defeat the purpose of making a nice clean EDSL for expressions; I might as well use 3*X + Y `lessthan` 3, which is just not cool. Does anyone know of a good solution, here? Are there good substitutions for all the six operators that are important (<,>,>=,<=,==,/=), that are close enough to be pretty-looking but not used for other important modules? Better yet, though a little harder, is there a nice type trick I'm not thinking of? This works for Num methods but not for Ord methods because: (+) :: (Num a) => a -> a -> a (<) :: (Ord a) => a -> a -> Bool i.e. the return type of comparisons is totally fixed. I don't suppose there's a good way to...well, I don't know what the *right* answer is, but maybe define a new typeclass with a more flexible type for < that lets both standard types return Bool and my expressions return Expr? Any good solution would be appreciated. Thanks, AHH

8 10

Hand calculation of Bird's definition of zip using foldr
by R J 12 Mar '09

12 Mar '09

Can someone provide a complete hand calculation of zip [1,2,3] [4,5,6] using the following definition of zip, based on foldr: zip :: [a] -> [b] -> [(a, b)] zip = foldr f e where e ys = [] f x g [ ] = [] f x g (y : ys) = (x , y) : g ys foldr :: (a -> b -> b) -> b -> ([a] -> b) foldr _ e [] = e foldr f e (x : xs) = f x (foldr f e xs) This implementation of zip produces the expected result [(1, 4), (2, 5), (3, 6)], but I'm unable to do the hand calculation and don't understand why it works. Part of my problem is that "e" is defined as a function that takes one argument, I don't see how that fits in with the usual scheme for foldr, which, as I understand it, is: foldr f e [x1, x2, ...] = f x1 (f x2 (f x3 ...(f xn e)))... Thanks, as always, to all in this great community. _________________________________________________________________ Windows Live™: Life without walls. http://windowslive.com/explore?ocid=TXT_TAGLM_WL_allup_1a_explore_032009

3 2

Re: [Haskell-cafe] ThreadScope: Request for features for the performance tuning of parallel and concurrent Haskell programs
by Don Stewart 12 Mar '09

12 Mar '09

karel.gardas: > Don Stewart wrote: > > marlowsd: > >> Ben Lippmeier wrote: > >> > >>> On 12/03/2009, at 12:24 AM, Satnam Singh wrote: > >>>> Before making the release I thought it would be an idea to ask people > >>>> what other features people would find useful or performance tuning. > >>>> So if you have any suggestions please do let us know! > >>>> > >>> Is it available in a branch somewhere to try out? > >> There are three parts to it: > >> > >> - some patches to GHC to generate the log files. The patches are > >> not yet in, but I hope to get them in in the next couple of weeks. > > > > Just as a meta-point, it makes a *lot* of sense for the runtime to > > support proper logging -- it's practically a microkernel after all, so > > logging makes as much sense for GHC's rts as it does for regular > > kernels. > > > > Now we just need a /proc for the rts, so I can peek at GC stats live. > > Seeing this thread and discussion direction I cannot resist to > temptation to ask: is anybody going to write rts DTrace[1] provider? I think this is clear next step in the investigations. -- Don

1 0