October 2008 - Haskell-Cafe

Problem with haddock 2.3.0
by Leonidas Fegaras 27 Oct '08

27 Oct '08

You have received a secure message. Read your secure message by opening the attachment, securedoc.html. You will be prompted to open (view) the file or save (download) it to your computer. For best results, save the file first, then open it in a Web browser. To access from a mobile device, forward this message to mobile(a)res.cisco.com to receive a mobile login URL. If you have concerns about the validity of this message, contact the sender directly. First time users - will need to register after opening the attachment. Help - https://res.cisco.com/websafe/help?topic=RegEnvelope About Cisco Registered Email Service - https://res.cisco.com/websafe/about">https://res.cisco.com/websafe/about

1 0

Problem with haddock 2.3.0
by Fegaras, Leonidas 27 Oct '08

27 Oct '08

You have received a secure message. Read your secure message by opening the attachment, securedoc.html. You will be prompted to open (view) the file or save (download) it to your computer. For best results, save the file first, then open it in a Web browser. To access from a mobile device, forward this message to mobile(a)res.cisco.com to receive a mobile login URL. If you have concerns about the validity of this message, contact the sender directly. First time users - will need to register after opening the attachment. Help - https://res.cisco.com/websafe/help?topic=RegEnvelope About Cisco Registered Email Service - https://res.cisco.com/websafe/about">https://res.cisco.com/websafe/about

2 1

Re: [Haskell-cafe] Why 'round' does not just round numbers ?
by Janis Voigtlaender 27 Oct '08

27 Oct '08

Felipe Lessa wrote: > On Mon, Oct 27, 2008 at 8:15 AM, Janis Voigtlaender > <voigt(a)tcs.inf.tu-dresden.de> wrote: > >>That is of course true (and was the topic of heated discussion with my >>fourth grade math teacher), but does not explain 2.5 -> 2. > > > If you round to odd instead of round to even, then 4.5 rounds to 5, Well, of course I did not learn to round to odd. I learned to round .5 to above, but not to do repeated rounding. So we would never have either of > 2.45 -> 2.4 -> 2 > 2.45 -> 2.5 -> 3 as you suggest. Instead, always in one go: 2.45 -> 2 Because of the 4, whereas any digit following it would be ignored: 2.4x -> x In fact, by your explanation, you would have: 3.46 -> 3.5 -> 4 Which is most odd. 3.46 is definitely closer to 3 than to 4. -- Dr. Janis Voigtlaender http://wwwtcs.inf.tu-dresden.de/~voigt/ mailto:voigt@tcs.inf.tu-dresden.de

7 11

Haskell on the JVM
by David Leimbach 27 Oct '08

27 Oct '08

Are there, or have there been Haskell ports to the JVM? Are any of them alive and well? Is there an interest in hosting GHC on the JVM (besides my own). I know I don't have time to tackle such a project but I'd be extremely interested in using such a thing. Especially in my day to day work. I currently do a lot of Java <-> Erlang via JInterface, and have been for months, yet there's many tasks I could do better in a language like Haskell.

9 12

Re: [Haskell] Re: Current XML libraries status
by Uwe Schmidt 27 Oct '08

27 Oct '08

Hello David, > I tried to use HXT's readDocument with its tagsoup option for my > application. I couldn't find a way to construct the operation that > didn't run out of memory. I'll attach some code using HaXml's > saxParse so you can see what I want. Is that easy to do in HXT? > I simply want the text of <PMID> and <AbstractText> elements. here's an example, that reads the input in a lazy way. I ran this in ghci with a file containing 2^20 XML Elements. The file was about 18Mb in size. A normal parse with the standard parsec parser ran out of memory on my 1Gb box. This one used within ghci about 200Mb max. ------------------------------------ module Main where import Text.XML.HXT.Arrow import System main = do mapM_ main' names main = do (name:_) <- getArgs runX ( readDoc name >>> fromLA (deep (hasName "PIMD" -- select the nodes <+> hasName "AbstractText" ) >>> getChildren -- get the text >>> getText ) >>> arrIO putStrLn ) putStrLn "main finished" readDoc = readDocument [ (a_tagsoup, v_1) , (a_parse_xml, v_1) , (a_remove_whitespace, v_1) , (a_encoding, isoLatin1) , (a_issue_warnings, v_0) , (a_trace, "1") ] --------------------- Cheers, Uwe Schmidt -- Uwe Schmidt Web: http://www.fh-wedel.de/~si/

1 0

External Sort: Sort a 10-million integer file with just 256M of ram.
by Thomas Hartman 27 Oct '08

27 Oct '08

Now on hackage: cabal install external-sort demo (included in distribution): thartman@thartman-laptop:~/external-sort>cat demo.hs {-# LANGUAGE PatternSignatures #-} import Algorithms.ExternalSort import Data.List import System.IO import System.Environment (getArgs) import System.Time import HSH -- to do: compare speed against unix sort util on a 10 million line file. -- pure in-memory prelude sort will crash your computer when the list gets over a million elements or so -- externalsort caches the sublists used in the sort algorithm on your hard drive, so you can sort a much larger list. {- The behavior below was on a demo executable, compiled. (In ghci, even last on a 10 million element list caused an out of memory error.) The test computer had 256M physical ram and was ulimited to 256M cache. *Main>:! ulimit -v 262144 For 10 million element list: *Main> :! time ./demo preludesort 7 demo: out of memory (requested 1048576 bytes) Command exited with non-zero status 1 4.88user 0.68system 0:21.11elapsed 26%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+64817minor)pagefaults 0swaps *Main> :! time ./demo externalsort 7 10000000 73.87user 1.96system 1:24.25elapsed 90%CPU (0avgtext+0avgdata 0maxresident)k 792inputs+156280outputs (6major+16739minor)pagefaults 0swaps ./demo unixsort 7 wrote bigfile, time: Mon Oct 20 15:25:26 CEST 2008 demo: out of memory (requested 1048576 bytes) For 100 million element list, external sort failed. Can it be made to work? maybe in some future version. time ./demo externalsort 8 demo: out of memory (requested 1048576 bytes) real 10m14.061s user 8m26.712s sys 0m11.793s thartman@thartman-laptop:~/external-sort>ls -lh ExternalSort.bin -rw-r--r-- 1 thartman thartman 764M Oct 20 15:50 ExternalSort.bin The problem is not fitting a 10^8 element list in memory, the following works fine (when compiled, though not in ghci): t = putStrLn . show . last $ [1..10^8::Int] Maybe think about this more later. -} main = do [s,e] <- getArgs let exp = read e case s of "preludesort" -> sortwith exp $ return . sort "externalsort" -> sortwith exp externalSort "unixsort" -> unixsort exp _ -> let msg = "usage: ./demo preludesort 7 or ./demo externalsort 7 or ./demo unixsort 7 \ \(sort 10 million element list)" in fail msg sortwith exp s = putStrLn =<< return . show . last =<< s ([1..10^exp ::Int]) unixsort exp = do let fn = "bigfile" withFile fn AppendMode (\h -> (mapM_ (hPutStrLn h . show) ([1..10^exp::Int]) ) ) putStrLn . ( ("wrote " ++ fn ++ ", time: ") ++ ) . show =<< getClockTime run $ "time tail -n1 | sort " ++fn :: IO String return () t = putStrLn . show . last $ [1..10^8::Int]

5 6

Instances of Lattice?
by Galchin, Vasili 27 Oct '08

27 Oct '08

Hi Henning, I am rereading my emails and I don't believe I got an examples of "instance Lattice". E.g. instance Lattice Bool. ?? Thanks, Vasili

2 1

Examples of Mutually Recursive Types
by Hugo Pacheco 26 Oct '08

26 Oct '08

Hi all, I have been searching for examples of Haskell "real scenarios" that employ mutually recursive datatype definitions. Does anyone know some interesting libraries or structures that I could play with? Thanks, hugo -- www.di.uminho.pt/~hpacheco

6 6

problem using ST monad
by Paul L 26 Oct '08

26 Oct '08

I'm have some trouble using the ST monad, and I think I'm confused about its use of existential type. > {-# OPTIONS -XRankNTypes #-} > import Control.Monad.ST > import Data.Array.ST I want to implement a map function that unfold all ST monads in a list: > mapST :: (a -> (forall s . ST s b)) -> [a] -> [b] > mapST f (x:xs) = runST (f x) : mapST f xs > mapST f [] = [] This is fine. However, it wouldn't typecheck if I had declared its type differently: mapST :: (a -> ST s b) -> [a] -> [b] I first thought the reason could be that runST has to take something of type (forall s . ST s b), but apparently this is not the case: > g :: (STArray s Int Int -> ST s a) -> ST s a > g f = newArray (0,0) 0 >>= f with this definition, runST (t (flip readArray 0)) happily returns me 0, despite the fact that the "s" in the type of "g" is not existentially quantified. Then when I try this one: mapST g [flip readArray 0] It fails to typecheck. Why? Is it possible, if at all, to implement a generic mapST? -- Regards, Paul Liu Yale Haskell Group http://www.haskell.org/yale

4 6

Slavomir Kaslev has invited you to radiusIM.com
by Slavomir Kaslev 26 Oct '08

26 Oct '08

2 1