December 2014 - Haskell-Cafe

Haskell Package for Fedora 21?
by Dominick Samperi 31 Dec '14

31 Dec '14

It appears that there are some problems using the Haskell Package with Fedora Linux 21. After following the instructions for Debian Linux, an attempt to start ghci yields this: $ ghci Loading packatge ghc-prim ... linking done. Loading package integer-gmp .. <command line>: can't load .so/.DLL for: libgmp.so (cannot open shared object). Adding /usr/lib64 or /usr/lib64/openssl/engines to LD_LIBRARY_PATH doesn't help, even though libgmp.so is located in the latter directory. Are there work-arounds for this? Are there best practices for Fedora 21? Compile from source? Thanks, Dominick

4 4

Haskell Platform for Fedora 21?
by Dominick Samperi 31 Dec '14

31 Dec '14

The thread with title "Haskell Package for Fedora 21?" should have title "Haskell Platform for Fedora 21?" Sorry about the noise...

1 0

Cabal 1.22 RC ready to test
by Johan Tibell 30 Dec '14

30 Dec '14

Hi all, I've now prepared release candidate for Cabal/cabal-install 1.22. Please test them and report bugs here and/or on the bug tracker ( https://github.com/haskell/cabal/issues) To install: cabal install http://johantibell.com/files/Cabal-1.22.0.0-rc.tar.gz http://johantibell.com/files/cabal-install-1.22.0.0-rc.tar.gz Cabal is now such a big project that's used in so many context (e.g. different OSes) that we need your help to help us test, to make sure that new releases work for you. -- Johan

1 0

User-Controlled Defaulting Proposals?
by Andrew Gibiansky 30 Dec '14

30 Dec '14

Hey all, I wish Haskell gave typeclass authors more control over defaulting. I see there are a number of proposals here: https://ghc.haskell.org/trac/haskell-prime/wiki/Defaulting Are any of these being actively pursued? If not, why not, and if so, what proposals will potentially make it into Haskell eventually? Just curious, Andrew

2 2

Encoding for flow charts
by Richard Wallace 30 Dec '14

30 Dec '14

Hello, I've been working to encode a fairly complex flow chart. At first I took the simple approach of having a single data type for the possible states data S = S0 | S1 | S2 and then a simple recursive function to step through it loop :: Monad m => S -> StateT s m Result loop S0 = testS0 >>= \r -> if r then loop S1 else loop S2 loop S1 = s1Result loop S2 = s2Result This works well enough, but with the size of the flow chart I'm working with (50+ decision points) it quickly becomes clear that it's probably not the most best way to do it. One problem that I'd like to solve is being able to test each decision point in isolation, making sure that decision point `m` will go to decision point `n` when it is supposed to, without having to run through the whole rest of the flow chart that would otherwise follow. Something else I'd like to achieve is being able to organize the decisions into separate modules - much better than having just the one big function above. Here's what I've come up with as an alternative so far. data Step a b where TrueFlow :: Decision a d e => a -> Step a b FalseFlow :: Decision b f g => b -> Step a b Done :: Result -> Step a b class Decision a b c | a -> b, a -> c where decide :: (Functor m, Monad m) => a -> StateT s m (Step b c) data S0 = S0 deriving (Show, Eq) instance Decision S0 S1 S2 where decide _ = bool (FalseFlow S2) (TrueFlow S1) <$> testS0 data S1 = S1 deriving (Show, Eq) instance Decision S1 () () where decide _ = s1Result data S2 = S2 deriving (Show, Eq) instance Decision S2 () () where decide _ = s2Result decision :: (Functor m, Monad m, Decision a b c) => a -> StateT s m Result decision s = decide s >>= \case TrueFlow b -> decision b FalseFlow c -> decision c Done s -> return s With this implementation it is easy to test a single decision point, - just use `evalStateT (decide S0) s` and compare the resulting Step to make sure it is either S1 or S2. The logic can also be easily separated and organized into modules. The biggest drawback of this new approach is the need to create a new `Step` record at every decision point. In the end, this may be worth it for the added design benefits, but would love to be able to get rid of it and find a solution that gives the added benefits without incurring any performance penalties. Any suggestions anyone can give on improvements would be greatly appreciated. Thanks! Rich

3 4

macosx-app on Yosemite
by Jeffrey Brown 30 Dec '14

30 Dec '14

Dear list, I tried to install wxHaskell, following these instructions: https://www.haskell.org/haskellwiki/WxHaskell/Mac It lists some configurations known to work, but I am running OS X 10.10,which is not among them. "brew install wxmac --devel" reported no errors. Next, "cabal install wx cabal-macosx" reported no errors. Next, "ghc --make HelloWorld.hs" reported no errors. But when I try to run "macosx-app HelloWorld", I get "-bash: macosx-app: command not found". Below is a transcript of the installation process, tab-indented for ease of viewing. Many thanks, Jeff ---- this worked: Jefferys-MBP:~ jbb$ brew install wxmac --devel ==> Installing dependencies for wxmac: jpeg, libpng, libtiff ==> Installing wxmac dependency: jpeg ==> Downloading https://downloads.sf.net/project/machomebrew/Bottles/jpeg-8d.yosemi ######################################################################## 100.0% ==> Pouring jpeg-8d.yosemite.bottle.2.tar.gz [image: 🍺] /usr/local/Cellar/jpeg/8d: 18 files, 776K ==> Installing wxmac dependency: libpng ==> Downloading https://downloads.sf.net/project/machomebrew/Bottles/libpng-1.6.15. ######################################################################## 100.0% ==> Pouring libpng-1.6.15.yosemite.bottle.tar.gz [image: 🍺] /usr/local/Cellar/libpng/1.6.15: 17 files, 1.3M ==> Installing wxmac dependency: libtiff ==> Downloading https://downloads.sf.net/project/machomebrew/Bottles/libtiff-4.0.3. ######################################################################## 100.0% ==> Pouring libtiff-4.0.3.yosemite.bottle.tar.gz [image: 🍺] /usr/local/Cellar/libtiff/4.0.3: 254 files, 3.8M ==> Installing wxmac ==> Downloading https://downloads.sf.net/project/machomebrew/Bottles/wxmac-3.0.2.yo ######################################################################## 100.0% ==> Pouring wxmac-3.0.2.yosemite.bottle.9.tar.gz [image: 🍺] /usr/local/Cellar/wxmac/3.0.2: 777 files, 41M Jefferys-MBP:~ jbb$ next, this worked: Last login: Sun Dec 28 15:38:02 on ttys000 Jefferys-MBP:~ jbb$ cabal install wx cabal-macosx Resolving dependencies... Downloading fgl-5.4.2.4... Configuring wxc-0.91.0.0... Configuring fgl-5.4.2.4... Building fgl-5.4.2.4... Building wxc-0.91.0.0... Installed fgl-5.4.2.4 Downloading cabal-macosx-0.2.3... Configuring cabal-macosx-0.2.3... Building cabal-macosx-0.2.3... Installed cabal-macosx-0.2.3 Installed wxc-0.91.0.0 Downloading wxcore-0.91.0.0... Configuring wxcore-0.91.0.0... Building wxcore-0.91.0.0... Installed wxcore-0.91.0.0 Downloading wx-0.91.0.0... Configuring wx-0.91.0.0... Building wx-0.91.0.0... Installed wx-0.91.0.0 Updating documentation index /Users/jbb/Library/Haskell/share/doc/index.html Jefferys-MBP:~ jbb$ next, this failed: Jefferys-MBP:wxHaskell jbb$ date Mon Dec 29 16:56:05 PST 2014 Jefferys-MBP:wxHaskell jbb$ pwd /Users/jbb/Mine/fun/computer/Haskell/wxHaskell Jefferys-MBP:wxHaskell jbb$ ls HelloWorld.hs Jefferys-MBP:wxHaskell jbb$ ghc --make HelloWorld.hs [1 of 1] Compiling Main ( HelloWorld.hs, HelloWorld.o ) Linking HelloWorld ... Jefferys-MBP:wxHaskell jbb$ macosx-app HelloWorld -bash: macosx-app: command not found Jefferys-MBP:wxHaskell jbb$

2 2

Do you know the author of the paper "Semantic Lego" ?
by Hidekazu IWAKI 29 Dec '14

29 Dec '14

Dear Sirs. I want to introduce the monad theory as a modular denotational semantics. So I am translating a really nice paper "Semantic Lego" into Japanese. But I have not be able to get the agreement of translating and publishing on the web. I already sent e-mails to two contact address. I know them from his homepage http://groups.csail.mit.edu/mac/users/dae/ However the addresses are not use both them. So anyone know him and his e-mail address, could you please forward the following mail to the author Dr. David Espinosa ? It should be noted he seems to work at Google. Sincerely. ------------------ My name is IWAKI Hidekazu and I am an amateur researcher of computer science and a Japanese Wikipedian. I am a Japanese living in Japan. https://ja.wikipedia.org/wiki/%E5%88%A9%E7%94%A8%E8%80%85:I.hidekazu I am concerning the methods of modular denotational semantics(the theory of monads). But for example Lian,Hudak or Wadler's paper is so difficult or not clear. so I can not explain the theory. Your Ph.D thesis "Semantic Lego" is so nice paper for learning the methods. It is very important for every computer scientists and functional programmer interesting in the theory of Monad. I want to translate this paper to Japanese. So could you give me the agreement ofJapanese translation? I am looking forward to hearing from you. Faithfully yours. -- *:..｡o○☆*ﾟ¨ﾟﾟ･*:..｡o○☆*ﾟ¨ﾟﾟ･*:..｡o○☆*ﾟ¨ﾟﾟﾟ *岩城　秀和（いわき　ひでかず） *メール: i.hidekazu at gmail.com *:..｡o○☆*ﾟ¨ﾟﾟ･*:..｡o○☆*ﾟ¨ﾟﾟ･*:..｡o○☆*ﾟ¨ﾟﾟﾟ

3 2

Is it possible to zip sequences any faster?
by David Feuer 28 Dec '14

28 Dec '14

containers-0.5.6 has a new implementation of zipWith in Data.Sequence that, like the old one, offers O(n) zipping, but unlike the old one offers immediate O((log(min{i,n-i}))^2) access to the elements, where i is the requested index. I haven't been able to think of a way to bring that down from polylogarithmic to logarithmic. The difficulty is that t splits one of the sequences repeatedly, so reaching an element the first time is kind of like pushing a snow shovel all the way down the driveway, rather than throwing snow off to either side. Does anyone have an idea for doing it better, or a good argument that it can't be done? Thanks, David Feuer

1 0

How to call hUpdateAtHNat with a value known only at runtime
by Phil Darnowsky 28 Dec '14

28 Dec '14

Hello all, I'm working on something that involves HLists where I'm making use of the hUpdateAtHNat function for array-like semantics, which works great if you know the index of the element you want at compile time: Prelude> :module +Data.HList Prelude Data.HList> let hl = hEnd $ hBuild "foobar" 123 True Loading [a bunch of packages] Prelude Data.HList> hl H["foobar", 123, True] Prelude Data.HList> hUpdateAtHNat hZero 90909 hl H[90909, 123, True] Prelude Data.HList> hUpdateAtHNat (hSucc hZero) 90909 hl H["foobar", 90909, True] Prelude Data.HList> hUpdateAtHNat (hSucc $ hSucc hZero) 90909 hl H["foobar", 123, 90909] So far so good. What's giving me trouble is figuring out how to update at an arbitrary index, where that index is only known at runtime, because hZero and hSucc are of different types: Prelude Data.HList> :t hZero hZero :: Proxy 'HZero Prelude Data.HList> :t hSucc hZero hSucc hZero :: Proxy ('HSucc 'HZero) What I'd ideally like is some kind of function intToHNatProxy that would take an Int and return a Proxy of the appropriate type, but I so far only kinda sorta understand how proxies and lifted types work, so I haven't yet succeeded in constructing one. Any advice would be very welcome. Thanks, Phil

4 4

ANN: Groundhog-inspector, a tool for generating datatypes from database
by Boris Lykah 28 Dec '14

28 Dec '14

Hello everyone, I am pleased to announce groundhog-inspector. It analyzes database schema and creates corresponding datatypes and mapping configuration for Groundhog. It works with PostgreSQL, MySQL, and Sqlite. Composite keys, constraints, references across schemas, and other details of schema are reflected in the output. Groundhog-inspector can be used as a library and provides a standalone tool for simple scenarios. Here is an example of the standalone tool usage. JSON mapping in the output was manually converted to YAML for brevity. $ sqlite3 dbfile "CREATE TABLE mytable (id INTEGER PRIMARY KEY NOT NULL, str VARCHAR NOT NULL, ref INTEGER references mytable)" $ groundhog_inspector sqlite dbfile data Mytable = Mytable {mytableStr :: String, mytableRef :: (Maybe (AutoKey Mytable))} - entity: "Mytable" dbName: "mytable" constructors: - name: "Mytable" fields: - name: "mytableStr" dbName: "str" - name: "mytableRef" dbName: "ref" * [Hackage] (http://hackage.haskell.org/package/groundhog-inspector) * [GitHub] ( https://github.com/lykahb/groundhog/tree/master/groundhog-inspector/examples ) I would be happy to answer the questions. Enjoy! Regards, Boris Lykah

2 3