- Haskell-Cafe - Haskell.org

Re: [Haskell-cafe] Newbie: State monad example questions
by Olivier Boudry 23 May '08

23 May '08

On Fri, May 23, 2008 at 2:20 PM, Thomas Hartman <tphyahoo(a)gmail.com> wrote: > > The big benefit I got from using the State Monad was that I was able to > reorder the functions > by just copy/pasting the function name from one > place to another. > > I don't understand... why do you need state to do this? Couldn't you > have a function pipeline using dots for composition like > > .... ( > ... > parseAttn . > parsePoBox . > ... > ) address ... > > and have the functions be equally switchable? (well, the first and > last can't quite be copy pasted, but close enough.) > > Introducing state seems like a lot of trouble to me if the only one is > easier reorderability of lines. > Agreed, in fact I started with a function pipeline and then switched to using the State Monad. As the program was written months ago I don't remember exactly why. Maybe I don't like to read backwards. ;-) Funtions running in the state monad can call other functions with the same `State s a` signature (and so on as deep as you want). You never have to care about passing parameters and restarting a new pipeline. But of course you can easily do without the State Monad. I don't think the State Monad allows to do thing that you can't do with basic Haskell. It just makes code more readable (in my opinion at least). Olivier.

1 0

RE: [Haskell-cafe] force inlining in GHC (fwd)
by Henning Thielemann 23 May '08

23 May '08

maybe also of interest in the ghc-core thread ---------- Forwarded message ---------- Date: Fri, 23 May 2008 21:21:03 +0200 (CEST) From: Henning Thielemann <lemming(a)henning-thielemann.de> To: Simon Peyton-Jones <simonpj(a)microsoft.com> Cc: glasgow-haskell-users(a)haskell.org Subject: RE: [Haskell-cafe] force inlining in GHC On Wed, 30 Apr 2008, Henning Thielemann wrote: > On Tue, 29 Apr 2008, Simon Peyton-Jones wrote: > > > As luck would have it, I'm working on INLINE pragmas for Roman right at > this moment. > > > > Could you spare a moment to give me a concrete test case, to make sure I > > hit your case too? If you can give me a program that doesn't optimise as > > you expect, I'm much more likely to get it right. I examined some more examples and found out that the behaviours of GHC-6.4.1 and GHC-6.8.2 are quite consistent, that is, accelerations I achieved for GHC-6.8.2 also worked for GHC-6.4.1. Main problem in both GHC-6.4.1 and GHC-6.8.2 remains that sometimes GHC decides to SPECIALISE a function that is tagged with INLINE. I think that SPECIALISE should either copy the INLINE tag to the specialised function (confluent rewriting, right?) or should not SPECIALISE an INLINE function at all. I can now present a simple example, however it needs a (self-contained) package: http://darcs.haskell.org/storablevector/ module Main where import qualified Data.StorableVector.Lazy as SV zipData :: SV.Vector Double zipData = SV.take 500000 $ SV.zipWith (+) (SV.iterate SV.defaultChunkSize ((1-1e-6)*) 0.5) (SV.iterate SV.defaultChunkSize (1e-6 +) 0) main :: IO () main = do SV.writeFile "test-data" zipData ghc-core -o dist/build/fusiontest/fusiontest -O -Wall -fexcess-precision -package storablevector speedtest/SpecialiseTest.hs I get quite at the beginning of the Core: a1 :: Data.StorableVector.Lazy.ChunkSize -> (Double -> Data.Maybe.Maybe (Double, Double)) -> Double -> [Data.StorableVector.Base.Vector Double] This looks much like specialised SV.unfoldr, the expansion of SV.iterate. It seems to be that GHC detects that SV.unfoldr is called twice with the same type and thus specialises it to Double - while forgetting that SV.unfoldr should be inlined anyway.

1 0

Re: throwDynIO ?
by Aleksey Nogin 23 May '08

23 May '08

Back in October'06 Carlos Pita sent a message to this list: > Currently I'm trying to grasp haskell exception handling and although > I find it mostly clear I'm a bit confused regarding the lack of a > throwDynIO action "paralelling" throwIO, as there are dynamic variants > for the other throw functions: > > throwDyn - throw > throwDynTo - throwTo > > I understand that the same effect could be conjured up by > > throwIO (DynException (...))) > > magic. But the lack of symmetry still disturbes me. Am I missing something? As far as I can tell, his question was left unanswered. As another Haskell newcomer (crossing over from the OCaml world), I found the absence of throwDynIO weird as well - as far as I understand things, throwing an error in the IO monad seems like a safer thing to do (as I'd know precisely when to expect it to be raised) than throwing it from a pure code and then relying on the "imprecise exceptions" mechanism to get it right. Looking at the ghc-6.8.2/libraries/base/Control/Exception.hs in the GHC source code, I see the following definitions: throw = unsafePerformIO . throwIO throwDyn exception = throw (DynException (toDyn exception)) so I guess I can get define throwDynIO myself by "taking back" the unsafePerformIO: throwDynIO exception = throwIO (DynException (toDyn exception)) but I am still wondering whether the library is simply incomplete here, or whether there is some reason for not defining/using throwDynIO. Thanks in advance for any clarifications! Aleksey

1 0

Trying to write a very simple HTTP Client
by Eric 23 May '08

23 May '08

Hi all, I've written the following program to connect to a submit an HTTP GET request to a server and print the response: module Main where import Network import System.IO main = withSocketsDo go go = do putStrLn "Connecting..." out <- connectTo "haskell.org" (PortNumber 80) hPutStrLn out "GET /\r" hPutStrLn out "Host: haskell.org\r" hPutStrLn out "\r" cs <- hGetLine out hClose out print cs When I run the program, however, I get the following error: HTTPClient: <socket: 1872>: hGetLine: end of file Can anyone see what's wrong? E.

4 4

Newbie: State monad example questions
by Dmitri O.Kondratiev 23 May '08

23 May '08

I am trying to understand State monad example15 at: http://www.haskell.org/all_about_monads/html/statemonad.html Example 15 uses getAny that I don't understand at all how it works: getAny :: (Random a) => State StdGen a getAny = do g <- get (x,g') <- return $ random g put g' return x Questions: 1) random has type: random :: (Random a, RandomGen g) => g -> (a, g) and for State monad: return a = State (\s -> (a, s)) then: return (random g) = State (\s -> ((a,g), s)) Is it correct? 2) What x and g' will match to in: do ... (x,g') <- return $ random g which, as I understand equals to: do ... (x,g') <- State (\s -> ((a,g), s)) What x and g' will match to in the last expression? 3) In general, in do expression (pseudo): do { x <- State (\s -> (a, s)); ...} What x will refer to? Will x stand for a whole lambda function: \s -> (a, s) ? 4) How 'g <- get' works in this function (getAny) ? 5) Why we need 'put g'? Thanks!

8 16

What to do when GHC works, RUNGHC fails
by Peter Verswyvelen 23 May '08

23 May '08

I'm experimenting creating a Cabal script on Windows that installs a module that refers to some DLLs (freetype, png, ...) via C wrappers When configuring, building, and installing using cabal, a test program that uses this installed module runs fine when compiled with GHC --make, but fails miserably when run with RUNGHC (or GHCi) The first thing I noticed is that RUNGHC looks for a DLL called "png.dll" (which it doesn't find and then bails out with an error), while the executable build with GHC uses the correct "libpng3.dll". When I rename the libpng3.dll into png.dll, RUNGHC goes a bit further but then fails with an "unknown symbol". Obviously I did something wrong in my Cabal script, but I have no idea what it is. There seems to be a big difference when it comes to linking between of GHCi and GHC . Any guidelines on what to do in cases like the one above? I really like to use the interpreter because it provides much shorter turnaround times, and things like this stop me from using it. Thanks, Peter

2 3

London Haskell Get Together
by Dominic Steinitz 23 May '08

23 May '08

For information, not an "official" user group get together but I believe the evening was enjoyed by all participants. Dominic. http://enhyper.blogspot.com/2008/05/haskell-get-together-22nd-may-2008.html

1 0

whhaskell on osx 10.5
by Kirk Peterson 23 May '08

23 May '08

I had a difficult time getting wxhaskell installed and working on a mac running os x 10.5. I did a quick write up of the process I got working here: http://necrobious.blogspot.com/2008/05/wxhaskell-go-go.html Im pretty new to haskell, so if if you see anything that needs correcting please let me know, hope this helps cheers, -kirk

4 4

MD5 performance optimizations, and GHC -via-C producing segfaulting binary
by Salvatore Insalaco 22 May '08

22 May '08

2008/5/17 Andrew Coppin <andrewcoppin(a)btinternet.com>: > Hi folks. > > OK, try this: > > darcs get http://darcs.orphi.me.uk/MyMD5 > cd MyMD5 > ghc -O2 --make md5sum > md5sum <some large filename> I've got some time to take a look at the code. It's very nice, readable and declarative, but obviously not optimized for "raw speed". There're a few things to do to have a speed-up of 2x, without going "low-level". 1) There's a lazyness leak in Pad.hs. The sum in: else make_block_bs bs0 : work (n + block_size_bytes) bs1 is not strict. With a very large file (e.g. 100 mb) it means stack overflow. [roxas:~/Desktop/test2/MyMD5] kirby% ghc -O2 --make md5sum.hs [roxas:~/Desktop/test2/MyMD5] kirby% time ./md5sum ../../jboss-4.2.2.GA.zip Stack space overflow: current size 8388608 bytes. Use `+RTS -Ksize' to increase it. To solve it just add a bang (!) before the n parameter: work !n bs = You've got to add {-# LANGUAGE BangPatterns #-} at the top of the file too. There're solutions that don't imply BangPatterns, and use only seq, but I like bang patterns! Now it works: [roxas:~/Desktop/test2/MyMD5] kirby% time ./md5sum ../../jboss-4.2.2.GA.zip E6542028D538BAEC180A96A5D1E6EC3A 11.958u 0.210s 0:12.19 99.7% 0+0k 0+2io 0pf+0w 2) make_block_bs is sub-optimal, and very critical to performance. I decided to use Data.Binary for it (it's also more readable, and you get rid of the unsafeWrite): import Data.Binary import Data.Binary.Get import Control.Monad // ... instance Binary Block where put _ = undefined get = do xs <- replicateM 16 getWord32le return $ Block $ listArray (0, 15) xs make_block_bs :: B.ByteString -> Block make_block_bs = decode Now we are getting faster: [roxas:~/Desktop/test2/MyMD5] kirby% time ./md5sum ../../jboss-4.2.2.GA.zip E6542028D538BAEC180A96A5D1E6EC3A 8.063u 0.174s 0:08.23 100.0% 0+0k 0+0io 0pf+0w 3) You are doing a lot of access to fields of a strict data type (State). You can at least ask the compiler to help you a bit with -funbox-strict-fields. Now we have: [roxas:~/Desktop/test2/MyMD5] kirby% time ./md5sum ../../jboss-4.2.2.GA.zip E6542028D538BAEC180A96A5D1E6EC3A 6.780u 0.133s 0:06.91 100.0% 0+0k 0+1io 0pf+0w We have got a good, nearly 2x, speed-up with very few optimizations, and we run in a very small constant amount of memory. Probably compiling with -fvia-C could help even more, but strangely: [roxas:~/Desktop/test2/MyMD5] kirby% ghc -fvia-C -funbox-strict-fields -O2 --make md5sum.hs [roxas:~/Desktop/test2/MyMD5] kirby% time ./md5sum ../../jboss-4.2.2.GA.zip Segmentation fault Is this supposed to happen? There're no "unsafe" functions or imports used in the program. Maybe a bug in GHC? Salvatore

7 20

Network.CGI and error handling
by Ketil Malde 22 May '08

22 May '08

Hi, I'm trying to implement a CGI, but I have encountered some problems with handling program errors properly. I think it boils down to this: The first program from the documentation at http://hackage.haskell.org/packages/archive/cgi/3001.1.5.2/doc/html/Network… import Network.CGI cgiMain = output "Hello World!" main = runCGI (handleErrors cgiMain) works fine, and ouputs a page with the expected test. Replacing line 2 with: cgiMain = output $ error "Hello World!" returns a blank page, although according to the documentation, 'handleErrors' should produce an error page (500) instead. The error is recorded in the log, where one can read: [Thu May 22 14:35:27 2008] [error] [client 10.1.9.227] test.cgi: Hello World! I've toyed with various approaches, including using catchCGI and replacing 'error' with other run-time failures, but only getting the same result. Any hints? (ghc 6.8.1, cgi-3001.1.5.1, apache 2.0, linux.) -k -- If I haven't seen further, it is by standing in the footprints of giants

1 2