- Haskell-Cafe - Haskell.org

Cabal and package changelog
by Manlio Perillo 09 Mar '09

09 Mar '09

Hi. The Cabal package description includes a property `license-file`. This property may be used by Hackage to provide a link to the license file. However I would like to have a `changelog-file`, too. So that from Hackage I can quickly check what's changed with the new version of a package. Is this feasible? Thanks Manlio Perillo

4 4

Looking for literature
by Gü?nther Schmidt 09 Mar '09

09 Mar '09

Hi all, I'm trying to catch up with all the wonderful Haskell Types, classes, Abstract Data Types, Algebraic Data Types, Types that give peoples headaches and all the other, deeper stuff I have been happily putting off. Now to look smart I have a bookshelf full of books for smart people and TaPL is amoung them. Should I try that or will RWH teach me enough to get by? Günther

5 4

cabal-install 0.6.2 does not bootstrap with ghc-6.10.1 debian distribution
by Ahn, Ki Yung 09 Mar '09

09 Mar '09

Dear Haskellers and especially who are working on cabal-install and debian packaging, I sometimes clean up .ghc and .cabal in my home directory to start from scratch because of dependency loopholes (cabal-install does not have remove option yet, so it's hard to fix when such loophole happens). Today, I had some time in the airport and decided to start from scratch again because of the dependency loophole with process 1.0.1.1 and haddock. I downloaded the most recent version of cabal-install the version 0.6.2, and found out that the ./bootstrap.sh does not work. So, I had to bootstrap from version 0.6.2 and do "cabal update" and "cabal upgrade cabal-install" to upgrade to 0.6.2. I am not sure whether this is a cabal-install problem or debian dstribution ghc-6.10.1 packaging probelm, since I have not tried to test this with any other ghc-6.10.1 distribution. If anyone who are not using debian distribution ghc-6.10.1 (e.g., general linux binary ghc-6.10.1 or source compiled one) can try bootstrapping cabal-install 0.6.2 from scratch also finds the same problem, I think someone should make a ticket for cabal-install. Thanks, Ahn, Ki Yung

4 6

Looking for a co-founder for a startup using Haskell
by Ed McCaffrey 08 Mar '09

08 Mar '09

Hello, I'm turning a project involving music into a startup, and I will be using Haskell for most and possibly all of it. I had an angel investor interested until the market collapse forced him to turn his focus away from new investments. Other investors I have spoken with want me to contact them again when it is further developed; basically, I think that I am optimistic at the prospect of funding with the right team. I am looking for one or more co-founders in the Boston, MA, area who would be willing to work on this full-time in the near future, after funding is secured, and in their spare time until then. Since many investors are currently moving towards larger investments in growth-stage companies, I am going to apply to incubators who specialize in early-stage companies. Anyone who wants more info send a reply to ed(a)edmccaffrey.net Thanks, Ed McCaffrey ed(a)edmccaffrey.net

3 2

Re: [Haskell] string type class
by Duncan Coutts 08 Mar '09

08 Mar '09

On Fri, 2009-03-06 at 19:16 +0000, Chris Kuklewicz wrote: > Matthew Pocock wrote: > > It seems every time I look at hackage there is yet another stringy > > datatype. For lots of apps, the particular stringy datatype you use > > matters for performance but not algorithmic reasons. Perhaps this is a > > good time for someone to propose a stringy class? > > Not likely. > > I did define my own (private) class for regular expressions, to abstract over > String, the ByteStrings, and Seq Char. But it is used in one place and is a > wart that should be removed. > > The simple task of looping over the contents of a String (once, forward) is > quite different from a Strict ByteString (using an index and a lookup). > > This means for decent efficiency I need two copies of my code, hand specialized > to each case. > > "tail" or "(x:xs)" : very efficient for String (no allocation) > "tail" or "uncons" : not efficient for ByteString (allocation, might as well > convert to [Char] head/tail/uncons on a strict or lazy ByteString is perfectly ok. I would recommend it over using length+index. If you are using tail in a tail recursion then ghc can almost always optimise it to unbox the various ByteString components and so there is no allocation of fresh BS constructors, just adjusting offsets and lengths held in registers. It's also possible to use unsafeHead and unsafeTail if you've checked for null. Duncan

2 1

3D Rendering Engine in Haskell (Re: [Haskell-cafe] FRP + physics / status of hpysics)
by Peter Verswyvelen 08 Mar '09

08 Mar '09

Awesome! When you say you took Ogre as a base, do you mean your engine will be able to load the meshes and material files, or are you also wrapping the C++ code base? Did you also consider Collada? 2009/3/8 Csaba Hruska <csaba.hruska(a)gmail.com> > Hi! > > This should be a bit offtopic, but now I'd like to announce that I'm > working on a 3D rendering rengine in Haskell. > I take ogre3d (www.ogre3d.org) as a base. My code can import import a > subset of ogre model files (currently xml version) and material scripts > (http://www.ogre3d.org/docs/manual/manual_14.html#SEC23) > You can easily import an existing 3d content or create new one because > there are ogre exporter plugins for every major modeler programs. > Currently I have some working code. I hope I'll be ready to release it in a > month. It should act as graphical backend for FRP libraries. > I use yampa in the example program but it's just an option, only example > program depends on it. > > Cheers, > Csaba Hruska > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe > >

4 5

Arch Linux: New package updates for Mar 8, 2008
by Don Stewart 08 Mar '09

08 Mar '09

New Haskell packages for the week ending Mar 8. http://archhaskell.wordpress.com/2009/03/08/arch-haskell-news-mar-8-2009/ Arch now has 962 Haskell packages in AUR. That’s 17 new packages this week, and lots of updates as well. Notable releases this week * htar-0.3: Command-line tar archive utility * haskell-llvm-0.6.5.0: Bindings to the LLVM compiler toolkit * haskell-harpy-0.4.1: Runtime code generation for x86 machine code * haskell-future-1.0.0: Supposed to mimics and enhance proposed C++ ”future” features * haskell-holumbus-mapreduce-0.0.1: a distributed MapReduce framework * haskell-holumbus-storage-0.0.1: a distributed storage system * haskell-holumbus-distribution-0.0.1: intra- and inter-program communication * hevolisa-dph-0.0.1: Genetic Mona Lisa problem in Haskell - using Data Parallel Haskell * doctest-0.0.0: Test interactive Haskell examples * haskell-happstack-0.2.1: The haskell application server stack + code generation See the full list of new Haskell packages this week on the blog, broken down by category, http://archhaskell.wordpress.com/2009/03/08/arch-haskell-news-mar-8-2009/

1 0

derive & Drift - some help please
by Gü?nther Schmidt 08 Mar '09

08 Mar '09

Hi, I need to derive Data, Typeable for Data.Time.LocalTime. I had hoped to use either derive or DrIFT to generate the instance declarations for it. I tried to understand the respective user manuals but it's still all Greek to me. Someone please help! Günther

4 6

Help with c2hs, structs and typedefs?
by Antoine Latter 08 Mar '09

08 Mar '09

Hey folks, I was trying to create some FFI bindings using c2hs and I ran into trouble: I can't use the 'sizeof' and 'set' hooks for structs that aren't defined inside a typedef. I don't know a whole lot about C - is there any declaration I can make to make something like my second struct work in c2hs? Because the second struct is what's provided in <sys/uio.h> :-) In the following code, if you change all of the directives in the haskell code from my_iovec_2 to my_iovec, the "unknown indentifier" errors go away. Thanks, Antoine >>>>> -- Works! #c typedef struct { void * my_iov_base; int my_iov_len; } my_iovec; #endc -- Broken! I get lots of "unknown identifier" errors for -- the #sizeof and #set hooks - but the #pointer hook seems to work. #c struct my_iovec_2 { void * my_iov_base; int my_iov_len; }; #endc {#pointer *my_iovec_2 as IOVec newtype#} withIOVec :: [(Ptr a, CSize)] -> (Ptr IOVec -> Int -> IO b) -> IO b withIOVec [] f = f nullPtr 0 withIOVec bufers f = let len = length buffers in alocaBytes (length bufers * {#sizeof my_iovec_2#}) $ \ptr -> vecAry <- castPtr ptr foldM (go vecAry) 0 bufers f vecAry len where go vecAry off (ptr, size) = do vec <- vecAry `plusPtr` off * {#sizeof my_iovec_2#} {#set my_iovec_2->my_iov_base#} vec $ castPtr ptr {#set my_iovec_2->my_iov_len#} vec $ cIntConv size return $ off + 1 <<<<<

2 2

ANNOUNCE: Holumbus-MapReduce 0.0.1
by Stefan Schmidt 08 Mar '09

08 Mar '09

The Holumbus framework is known for building specialized search-engines like the Hayoo! Haskell API search. Now, I'd like to announce the newest libraries of the Holumbus framework: Holumbus-MapReduce 0.0.1 Holumbus-Distribution 0.0.1 Holumbus-Storage 0.0.1 As part of my Master's Thesis at the University of Applied Sciences in Wedel, Germany, I have developed a library for building distributed MapReduce systems in Haskell. The existing Holumbus-Searchengine library already offers a MapReduce implementation, but it does not provide mechanisms for distributed computing. The newly developed Holumbus-MapReduce library contains of tools and modules for building individualized distributed MapReduce systems. Holumbus-MapReduce is based on two libraries: Holumbus-Distributed and Holumbus-Storage. Maybe they could be useful for building other distributed systems, not only a MapReduce system. Holumbus-Distribution consists of modules and tools for the implementation of distributed systems in general. Besides common data types and small helper functions, this library provides a mailbox-based data transfer mechanism. It offers an easy-to-use interface for inter- and intra-process communication like in Erlang or Mozart/Oz. Holumbus-Storage is designed to build a distributed storage system which can be linked into other programs. It is used by the MapReduce system for the data management. All three libraries are still under heavy development and should not be considered for creating applications in productive environments. But I think, it's time to give them to the community for suggestions on further development and features. Please email comments, bugs, etc. to sts(a)holumbus.org Stefan Schmidt Holumbus on Hackage: http://hackage.haskell.org/cgi-bin/hackage-scripts/package/Holumbus-MapRedu… http://hackage.haskell.org/cgi-bin/hackage-scripts/package/Holumbus-Distrib… http://hackage.haskell.org/cgi-bin/hackage-scripts/package/Holumbus-Storage link to my thesis: http://www.holumbus.org/redmine/attachments/37/thesis-mapreduce.pdf project homepage: http://www.holumbus.org/redmine/wiki/holumbus darcs repository (latest snapshots): http://holumbus.fh-wedel.de/holumbus/src

2 1