- Haskell-Cafe - Haskell.org

[ANN] Twitter Client
by Chris Eidhof 10 May '08

10 May '08

Hey everyone, I was tired of all those graphical Twitter clients that aren't usable from my Terminal, so I wrote my own. It's still very much alpha, but comments or improvements are more than welcome. You can install it by downloading the twitter-package from hackage. In order to get it to work, create a file "~/.twitter" with the contents username:password, and you should be set. The source code isn't polished yet, but it makes heavy use of the new curl, json and xml libraries from Galois, so it might be interesting to browse the source. -chris

3 3

ANN: unix-2.4.0.0 additional POSIX real time support
by Galchin, Vasili 10 May '08

10 May '08

Hello, Additional POSIX realtime (POSIX 1003.1b) support has been added: 1) MQueue.hsc - POSIX message queues Supported FFI bindings for: - mq_open - mq_close - mq_unlink - mq_getattributes - mq_setattributes - mq_send Not yet supported bindings: - mq_rcv .. still debugging a mysterious problem - mq_notify - will support when asynchronous I/O is implemented 2) RTsched.hsc - real time scheduling Supported bindings for: - sched_yield - sched_get_param - sched_set_param - sched _get_scheduler - sched_set_scheduler - sched_get_priority_min - sched_get_priority_max Unsupported currently: - sched_rr_get_interval - still looking for the TimeSpec data type in order to get "peek" working 3) Files.hsc Added bindings for: - fsync - fdatasync I am new to Haskell but would like to make a lot of contributions to insure Haskell success in the marketplace (of ideas ;^)). I apologize for any glitches and look for feedback. Kind regards, Vasili

2 1

ANN: Roguestar 0.2
by Christopher Lane Hinson 10 May '08

10 May '08

website: http://roguestar.downstairspeople.org darcs 2: darcs get --lazy http://www.downstairspeople.org/darcs/roguestar tarball: http://www.downstairspeople.org/roguestar/roguestar-0.2.tar.gz After much delay, I'm happy to announce Roguestar 0.2. Roguestar is a science fiction themed roguelike (turn-based, chessboard-tiled, role playing) game written in Haskell. Roguestar uses OpenGL for graphics. This is still a very early release and several important things don't work. This initial release allows you to play one of six alien races. You begin the game stranded on an alien planet, fighting off an endless hoard of hostile robots. RSAGL, the RogueStar Animation and Graphics Library, includes a domain-specific monad for 3D modelling of arbitrary parametric surfaces, as well as an animation monad and arrow, which is more or less like YAMPA as a stack of arrow transformers. RSAGL was specifically designed for roguestar, but every effort has been made (including the license) to make it accessable to other projects that might benefit from it. Performance is an issue, but there is a lot of low-hanging fruit in this area. Roguestar is released under the Affero General Public License version 3. However, RSAGL uses a permissive license. I would be absolutely thrilled to receive any feedback anyone may have on the design and aesthetics of the game, especially in the form of source code.

1 0

Re: [Haskell-cafe] inserting values in a binary tree
by PR Stanley 09 May '08

09 May '08

>Are there any invariants you wish to maintain when inserting? If >not, it's rather trivial. Paul: The idea is to find a value in the tree greater than the new value and then placing the new value before it on the tree. duplicates are ignored and if no value greater than he new value is found then it is appended as a new node to the end of the last node checked. in C you'd fiddle with pointers and Bob's your uncle. Here I'm not sure how to piece that tree back together again with the new element after having expanded it recursively. Cheers Paul

2 1

(Num t) => [t] or [Int]
by PR Stanley 09 May '08

09 May '08

Hi I thought [1, 2, 3] :: [Int] GHC says it's (Num t) > [t] Okay, when it comes to 3.3:[1,2,3] (Num t) => [t] makes more sense. Is that the only reason? Cheers, Paul

3 2

Seeking advice on best practices with FFI
by PJ Durai 09 May '08

09 May '08

Hello, I am trying to write FFI wrappers for a dynamic library in windows. I would like to get some help in how to go about doing things. For example for this C function, the import and wrapper I came up with are listed below. -- STATUS LNPUBLIC NSFDbOpen (char far *PathName, DBHANDLE far *rethDB); foreign import stdcall "nsfdb.h NSFDbOpen" nsfDbOpen' ::CString -> Ptr CInt -> IO CInt type DbHandle = CInt nsfDbOpen :: String -> IO DbHandle nsfDbOpen dbname = alloca $ \ptr -> do db <- newCString dbname rc <- nsfDbOpen' db ptr handle <- peek ptr if rc /= 0 then error "dbopen failed" else return handle Q: Does that look ok? Is the db <- newCString leaking ? Q: How would I import and call a function with signature like this calling C code: char str1[MAX_PATH]; char str2[MAX_PATH]; GetStringValues( str1, str2); called C function signature is: void GetStringValues( char* s1, char* s2); i.e. how would the haskell import (and the wrapper function) look like in situation like this (where I have to pass preallocated char array to C code and get it filled up)? Appreciate your time. thanks pj

2 1

MonadPlus
by Andrew Coppin 09 May '08

09 May '08

OK, so I feel I understand monads fine. I regularly use Maybe, [] and IO, and I've even constructed a few monads of my own. But here's a question: what is the purpose of the MonadPlus class? Clearly it defines a binary operation over monadic values and an identity element for that operation. But... what is this supposed to *do*? (For example, (>>) has an almost identical signature to mplus. But presumably they don't both do the same thing...) What functionallity is this supposed to give you? [In a somewhat unrelated question... I saw some code the other day that used Either as if it were a monad. And yet, I don't see an instance given in the standard libraries - even though there should be one. I can see Functor (Either a), but not Monad (Either a) or even Monad (Either String)...]

9 12

Haskell PNG Writer
by Nahuel Rullo 09 May '08

09 May '08

Hi list, i am new in Haskell. I need to make images (PNG, JPEG) with haskell, if you can give me a tutorial, thanks! -- Nahuel Rullo

8 7

Re: [Haskell-cafe] Induction (help!)
by PR Stanley 09 May '08

09 May '08

You've got the right idea. Paul: At long last! :-) I should point out that it doesn't make sense to say p(Succ n) = Succ(p(n)), p(x) represents some statement that is either true or false, so it doesn't make sense to say Succ(p(n)). . Paul: okay, da capo: We prove/test through case analysis that the predicate p holds true for the first/starting case/element in the sequence. When dealing with natural numbers this could be 0 or 1. We try the formula with 0 and if it returns the desired result we move onto the next stage. If the formula doesn't work with 0 and so the predicate does not hold true for the base case then we've proved that it's a nonstarter. In the inductive step we'll make a couple of assumptions: we'll imagine that p(j). We'll also assume that p holds true for the successor of j - p(j+1). Then with the help of rules and the protocol available to us we'll try to establish whether the formula (f) gives us f(j) = f(j+1) - f(1) So, we know that the predicate holds for 0 or at least one element. By the way, could we have 3 or 4 or any element other than 0? Anyway, p(0). Then we set out to find out if p holds for the successor of 0 followed by the successor of the successor of 0 and so forth. However, rather than laboriously applying p to every natural number we innstead try to find out if f(j+1) - f(1) will take us back to fj). I think this was the bit I wasn't getting. The assumptions in the inductive step and the algebraic procedures are not to prove the formula or premise per se. That's sort of been done with the base case. Rather, they help us to illustrate that f remains consistent while allowing for any random element to be succeeded or broken down a successive step at a time until we reach the base/starting element/value. Okay so far? Cheers Paul

3 5

[Haskell-cafe] Announce: SCC, Streaming Component Combinator library
by Mario Blazevic 09 May '08

09 May '08

I'm pleased to announce the 0.1 release of Streaming Component Combinators in Haskell, based on my earlier work done in OmniMark and published in paper Mario Blažević, Streaming component combinators, Extreme Markup Languages, 2006. http://www.idealliance.org/papers/extreme/proceedings/html/2006/Blazevic01/… The release consists of the following: - Yet another monad transformer with coroutine functionality. Its central feature is the function pipe :: forall context x m r1 r2. Monad m => Producer m x r1 -> Consumer m x r2 -> Pipe context m (r1, r2) - Based on the above, streaming component types Transducer and Splitter are defined, as well as a number of primitive component instances. - Next are component combinators that can be used to synthesize larger components. - Finally, there is a rudimentary executable, shsh (Streaming Haskell SHell), that exposes most of the above within a command-line shell. It's useful for experiments and testing of new components. Here are a few simple command examples: wc -c count | show | concatenate wc -l foreach line then substitute x else suppress end | count | show | concatenate grep "foo" foreach line having substring "foo" then append "\n" else suppress end sed "s:foo:bar:" foreach substring "foo" then substitute "bar" end sed "s:foo:[\\&]:" foreach substring "foo" then prepend "[" | append "]" end The SCC library and shell are available on Hackage, at http://hackage.haskell.org/cgi-bin/hackage-scripts/package/scc-0.1

1 0