August 2014 - Haskell-Cafe

Dealing with encodings
by Dmitry Bogatov 06 Aug '14

06 Aug '14

Hello! How teach hxt to handle "KOI8-R" encoding of input file? And it seems that so many great packages (like hxt, feed, curl) uses String. Is it some work in progress to port them to Text? -- Best regards, Dmitry Bogatov <KAction(a)gnu.org>, Free Software supporter, esperantisto and netiquette guardian. GPG: 54B7F00D

5 7

Applicative & Monad as interaction between categories
by Ruben Astudillo 05 Aug '14

05 Aug '14

Hi all, I recently came across of two blog posts[1][2] by Gabriel Gonzales where he shows motivation for writing compositional functions -- and thus why the concept of category is useful -- and motivation to write functors between them. I wanted to know if something similar for has been done to `Applicative' and `Monad' as design pattern and interaction between categories. I've searching but I always hit the Monad tutorials and more info on how to use them. Thanks in advance. Excuse my English. [1]: http://www.haskellforall.com/2012/08/the-category-design-pattern.html [2]: http://www.haskellforall.com/2012/09/the-functor-design-pattern.html -- Ruben Astudillo. pgp: 0x3C332311 , usala en lo posible :-) Crear un haiku, en diecisiete silabas, es complica...

3 2

[ANN] Hayoo Relaunch
by Sebastian Philipp 05 Aug '14

05 Aug '14

Hi Cafe, Hayoo <http://hayoo.fh-wedel.de> has been relaunched. Hayoo is a search engine for Hackage packages, which allows you to search for functions, data types and packages. It is currently work in progress. Any feedback is greatly appreciated! Hayoo uses Hunt <https://github.com/hunt-framework> for indexing and searching, which is the successor of Holumbus. Hunt is a flexible, lightweight search platform with a powerful query language and JSON API. Example search requests are: * Function names: map <http://hayoo.fh-wedel.de/?query=map> * Function signatures: (a->b)->f a->f b <http://hayoo.fh-wedel.de/?query=%28a-%3Eb%29-%3Ef+a-%3Ef+b> * Module names: Control.Loop <http://hayoo.fh-wedel.de/?query=Control.Loop> Have a look at the examples on <http://hayoo.fh-wedel.de/examples> for some advances queries. The old Hayoo and Holumbus are still online at <http://holumbus-alt.fh-wedel.de/> -- Sebastian Philipp

7 10

Fwd: Default stdout buffering of child process of createProcess
by Chris Myzie 04 Aug '14

04 Aug '14

I'm reposting this message because I think it only went to the google group and not the official haskell-cafe list: On Friday, August 1, 2014 10:06:32 AM UTC-4, Chris Myzie wrote: > > As a workaround, I am able to trick the child haskell process into > thinking it's running in an interactive terminal by wrapping it with > /usr/bin/script: > > createProcess (proc "/usr/bin/script" ["-qc","./A","/dev/null"]) { std_out > = CreatePipe } > > I still think haskell is using screwy defaults for stdout buffering.. > > On Thursday, July 31, 2014 3:24:47 PM UTC-4, Chris Myzie wrote: >> >> Hello, >> >> I'm trying to write a wrapper process that will be able to read any child >> process' stdout. The problem I'm running into is that unless I force the >> child's stdout to LineBuffering, it defaults to BlockBuffering. Is >> BlockBuffering really the default in this case? I don't want to have to >> modify all of the child processes that I want to use with this wrapper. >> >> Below is a simple test case. A.hs is the child process, and B.hs is the >> wrapper. If I run B.hs, I will get no output unless I uncomment the line >> in A.hs. >> >> Thanks, >> Chris >> >> >> ------------------------------ A.hs --------------------------- >> import Control.Concurrent >> import System.IO >> >> main :: IO () >> main = do >> -- hSetBuffering stdout LineBuffering >> putStrLn "test" >> threadDelay 1000000 >> main >> >> >> ------------------------------ B.hs --------------------------- >> import Control.Monad >> import System.IO >> import System.Process >> >> main :: IO () >> main = do >> (_,Just h,_,_) <- createProcess (proc "./A" []) { std_out = CreatePipe } >> hSetBuffering h LineBuffering >> forever $ hGetLine h >>= putStrLn >> >

7 12

[ANN] StrappedTemplates - general purpose templates
by Kyle Hanson 04 Aug '14

04 Aug '14

Hello all, I found that I wasn't quite satisfied by the template offerings (Heist, Hastache, Shakespearean, etc..) of Haskell, so I built my own. The goal is to have simple syntax that non-haskellers would be comfortable using, resembles templates found in other languages (Jinja, Django, etc) and an interface to render them that I don't have to think very hard to use. Features include inheritance, logic, includes, functions, and relatively speedy rendering (much faster than interpreted templates like Heist-I and Hastache) . Its very much a work in progress, but I felt that it was enough for an initial release and get general feedback on it. Comments are appreciated. One thing that I want to improve is how functions are handled in the template and in the "InputBucket", extend the syntax to allow you to do more inside the template and be able to define your own block functions. These will probably be added next major release. http://hackage.haskell.org/package/StrappedTemplates https://github.com/hansonkd/StrappedTemplates Thanks! Kyle Hanson

2 1

New Functional Programming Job Opportunities
by Functional Jobs 04 Aug '14

04 Aug '14

Here are some functional programming job opportunities that were posted recently: Pragmatic Haskell Developer at Anchor Systems http://functionaljobs.com/jobs/8729-pragmatic-haskell-developer-at-anchor-s… Cheers, Sean Murphy FunctionalJobs.com

1 0

Functional programming and William Lawvere's notion of "variable sets"(vs topos of "static sets") in a Topos ...
by Vasili I. Galchin 04 Aug '14

04 Aug '14

Hello Haskellers, I have been re-reading William Lawvere's description of "variable sets" (a functor category) in "a elementary topoi" on a discrete poset/preset vs a "abstract sets" in the topos of abstract sets . In Haskell( and other FPLs .. ) aren't "pure" computations just over the topos of abstract sets and "mutable" /"time-varying/stateful" computations aren't just "variable sets" (i.e. a functor over a discrete preset-poset)?? Please forgive my bad English :-( If uncertain aboyt my Englsih, please say so. Kind regards, Vasya P.S. Yes I am familiar with Eugenio Moggi's papers on using monads for stateful computations ..

2 1

Re: [Haskell-cafe] Fwd: Default stdout buffering of child process of createProcess
by Doug McIlroy 03 Aug '14

03 Aug '14

>> I still think haskell is using screwy defaults for stdout buffering. > Line buffered would be an inefficient default. Stdout is fully buffered in > c99 if it can be determined that it is not attached to a terminal. Putting efficiency before efficacy is un-Haskellian. How can the required determination be implemented? Stdout may be attached to a terminal indirectly through a pipeline. It may be "attached" through a file that another process is watching. It may be attached to some other real-time device or system. Buffering can cause unbounded delays in systems with sporadic input. And it can cause deadlock in systems that employ feedback, as most systems do. Terminals, for example, are unbuffered not merely because users are impatient, but more critically because users are feedback agents. Because the effects of buffering are not transparent, it would be wise to make it an optional optimization, not a default.

5 6

"Standardising" Haskell AST for tooling
by Alan & Kim Zimmerman 03 Aug '14

03 Aug '14

Hi all I am looking at updating HaRe for GHC 7.8.3, and I am fighting my way through the AST changes. One of the things I have had at the back of my mind for a while is to use something like Kure universes or more generally lens to provide a set of isomorphisms from the various flavours of AST into some kind of standardised internal representation, allowing easier tool manipulation. This could potentially decouple the tools from the specific compiler/analyzer, so they could work across GHC 7.6.x, GHC 7.8.x, haskell-src-exts, haste, etc.. Obviously there would be limitations to this where an advance in the language brings in new features, but a lot of useful tool work can be done that does not touch the new stuff. I don't know if anything similar is planned in either ghc-server or another toolchain. My haskell-token-utils is a first limited attempt to bring source code round tripping to a variety of backends, but a more general solution to the AST phase would help tooling in general So, is this worth doing? Is it being done already, or planned? Regards Alan

1 0

Bad interaction of inlinePerformIO and mutable vectors
by Michael Snoyman 03 Aug '14

03 Aug '14

I'm trying to understand why some code isn't behaving as I'd expect, and to determine whether it's a bug or not (and where that bug might be). Here's the simplest version of the code: import Data.ByteString.Internal (inlinePerformIO) import qualified Data.Vector as V import qualified Data.Vector.Mutable as VM main :: IO () main = do vm <- VM.new 1 VM.write vm 0 'A' let x = inlinePerformIO $ VM.write vm 0 'B' x `seq` (V.freeze vm >>= print) A more complete example is available on lpaste[1]. The problem is that I would expect the output to be "B", but in fact "A" is still printed. From the longer paste that I linked to, you can see that: * When using unsafePerformIO and unsafeDupablePerformIO, the semantics work as I would have expected: "B" is printed after forcing evaluation of the result. * If I add a `VM.read vm 0` call after the write, it also works. * Using IORef, the behavior is also as I would have expected: as soon as the result is evaluated, the reference is updated. I'm testing on GHC 7.8.3, Ubuntu 64-bit, and compiling with -O2. I'm curious if anyone has an idea as to why there is this difference in behavior. Michael [1] http://lpaste.net/108483

6 10