November 2009 - Haskell-Cafe

Fwd: (Solved) cabal install with external gcc tool chain not the ghc-bundled one
by Daniel Kahlenberg 12 Nov '09

12 Nov '09

to answer this question myself how the use of another gcc is specified with effect, I used the following options with the 'cabal install' call: --ghc-options="-pgmc e:/programme/ghc/mingw-gcc4/bin/gcc.exe -pgml e:/programme/ghc/mingw-gcc4/bin/gcc.exe" See http://www.haskell.org/ghc/docs/latest/html/users_guide/options-phases.html… (searched in the wrong direction, too many trees...). Slightly tuned, this should be the way to go in all similar cases. One thing I haven't considered yet is if the '--with-ld' and '--with-gcc' options (if curious too, see logs in my previous mail - Subject "[Haskell-cafe] caba install with external gcc toolchain not the ghc-bundled one") only effect what gets written into the setup-config/package.conf file or what other effects these have. Greets Daniel

2 1

socket error
by Alberto G. Corona 12 Nov '09

12 Nov '09

<socket: 1796>: hPutBuf: illegal operation (handle is finalized) I´m a bit lost trying to find the source of this error. I´m running an hack application (package hack). Basically it is a handler of web requests. hack can be used with different web servers: Hyena, simpleserver, HappStack.... all of them produce this error after a few interactions., Supposedly, it happens within the socket module since neither my module, nor hack, nor the hack simpleserver (package hack-handler-simpleserver) call explicitly hPutBuf I tried to reproduce the error under linux, but my ubuntu installation is too old and I´m in the process of reinstalling everything again. In the meantime, Any of you can give me any hint about the error?

2 3

Long running Haskell program
by David Leimbach 12 Nov '09

12 Nov '09

As some of you may know, I've been writing commercial Haskell code for a little bit here (about a year and a half) and I've just recently had to write some code that was going to run have to run for a really long time before being restarted, possibly months or years if all other parts of the system cooperate as it's part of a server infrastructure management system. I recently ran into some serious space leak difficulties that would ultimately cause this program to crash some time after startup (my simulator is also written in Haskell, and runs a LOT faster than the real application ever could, this has enabled me to fast forward a bit the data growth issues and crash in minutes instead of days!) Anyway, rather than try to paste it all here with images and such I thought I'd stick it up on my blog so others could maybe benefit from the anecdote. It's difficult to disclose enough useful information as it is commercial code not under an open source license, but there's neat diagrams and stuff there so hopefully the colors are at least amusing :-) http://leimy9.blogspot.com/2009/11/long-running-haskell-applications.html Dave

10 17

Problem about hidden package again.
by Magicloud Magiclouds 12 Nov '09

12 Nov '09

Hi, Today, when I compiled gtk2hs, I got this: cairo/Graphics/Rendering/Cairo.hs.pp:264:0: Failed to load interface for `Data.Array.Base': it is a member of the hidden package `array-0.2.0.0' Use -v to see a list of the files searched for. As usual, I searched for the resolvement. No luck. And seems like this problem (not particularly to this package) is very common. I wonder how to resolve it? -- 竹密岂妨流水过山高哪阻野云飞

3 5

(state) monad and CPS
by jean-christophe mincke 12 Nov '09

12 Nov '09

Hello, I would like to get some advice about state monad (or any other monad I guess) and CPS. Let's take a simple exemple (see the code below) 'walk' is a function written in CPS that compute the number of nodes & leaves in a tree. It use a counter which is explicitly passed through calls. 'walk2' is does the same using the state monad but is not written in CPS Is it possible to write a function 'walk3' written in CPS and using the state monad? Thank you Regards J-C module M where import Control.Monad.State data Node = Node (Node, Int, Node) |Leaf Int |Empty deriving (Show) walk Empty acc k = k acc walk (Leaf _) acc k = k (acc+1) walk (Node (l, _, r)) acc k = let k1 acc = walk r acc k in walk l (acc+1) k1 nb = Node (Leaf 1, 2, Leaf 3) nd = Node (nb, 4, Empty) nh = Node (Empty, 8, Leaf 9) ng = Node (Leaf 6, 7, nh) ne = Node (nd, 5, ng) r = walk ne 0 id walk2 Empty = return () walk2 (Leaf _ ) = do acc <- get put (acc+1) return () walk2 (Node (l, _, r)) = do acc <- get put (acc+1) walk2 l walk2 r return () r2 = runState (walk2 ne) 0

6 7

lambda-bot installation problem: gentoo trickery problem
by Евгений Пермяков 12 Nov '09

12 Nov '09

When I try cabal-install lambdabot (gentoo linux/amd64, ghc installed with portage), it runs fine until compiler tries to link readline package (some template haskell?). The problem caused by dirty trick, used in gentoo: the /usr/lib64/libreadline is a fake with script, redirecting ld to /lib64 . GHC is not redirected but simply fails with message "can't load .so/.DLL for: readline (/usr/lib64/libreadline.so: invalid ELF header)". So, the question is: is there any workaround? Copying library look like an option, but it is very, very dirty one. Is there a way to say ghc, which libreadline.so it should load?

2 1

Static Linking Problem
by MightyByte 12 Nov '09

12 Nov '09

I am trying to statically compile a simple haskell program so I can use it on a Linux computer without haskell and it's associated libraries. Here is a small example program that illustrates my problem: > module Main where > import Network.Fancy > main = do > withDgram (IP "127.0.0.1" 1234) (flip send "Hello network\n") After a bit of googling, I came to the conclusion that I needed to compile it with "ghc --make -static -optl-static Foo.hs". Using only "-static" or "-optl-static" by themselves did not generate a statically linked binary. But when I compile with both those parameters I get a bunch of linker errors: /home/mightybyte/.cabal/lib/network-fancy-0.1.4/ghc-6.10.4/libHSnetwork-fancy-0.1.4.a(Fancy.o): In function `s6ks_info': (.text+0x3068): warning: Using 'getaddrinfo' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking /usr/lib/ghc-6.10.4/libffi.a(closures.o): In function `init_mparams': (.text+0x3e): undefined reference to `pthread_mutex_lock' /usr/lib/ghc-6.10.4/libffi.a(closures.o): In function `init_mparams': (.text+0x52): undefined reference to `pthread_mutex_unlock' /usr/lib/ghc-6.10.4/libffi.a(closures.o): In function `init_mparams': (.text+0xd3): undefined reference to `pthread_mutex_init' /usr/lib/ghc-6.10.4/libffi.a(closures.o): In function `ffi_closure_free': (.text+0x59c): undefined reference to `pthread_mutex_lock' etc... I've tried this on both Fedora and Arch Linux and I get the same error. Anyone know how to solve this problem?

4 5

"Least common supertype"?
by Sean Leather 12 Nov '09

12 Nov '09

Is there a name for the following concept? Can you point me to any references on it? Suppose I have the following two functions ... > swap1 :: (Int, Char) -> (Char, Int) > swap2 :: (Char, Int) -> (Int, Char) ... and, for some reason, I think I can unify these into a single function. I think, hmm, given that the structure is that same, let's do a first pass: > swap? :: (a, b) -> (c, d) But then I go back to the input types to confirm that this will work, and, alas, it will not, because there are similarities that I missed. This is way too general. I need to ensure that what's an Int stays an Int and likewise for Char. > swap! :: (a, b) -> (b, a) And now I have found a type that is more general than swap1 and swap2 and yet not so general that the shared constraints are left out. This seems somewhat analogous to the least common multiple. Another example is the following: > showFloat :: Float -> String > showBool :: Bool -> String We could say the more general type is ... > show? :: a -> String ... but then we lose the implied constraint that we must know something about 'a' to produce a string. So, we add back such some such constraint: > show! :: (Show a) => a -> String Of course, with all of this, it may not be clear what to do about the definitions of the functions, but I'm curious if there's a name for the concept from a type perspective. Thanks, Sean

7 8

Wrestling the inliner beast
by Bryan O'Sullivan 12 Nov '09

12 Nov '09

I did a little bit of Data.Text benchmarking the other day, and I was shocked to find that decoding a UTF-8 stream proceeded at a sedate 3MB/sec. On investigation, the culprit was that I'd marked both the outer function and the inner (a loop) as INLINE. Because the inner loop was marked as INLINE, GHC wasn't inlining critically important leaf functions into it, so I was getting clobbered to death by boxing and unboxing. Leaving the outer function marked as INLINE, but taking the INLINE off the inner function, seems to cause *both* to get inlined as I originally hoped. This behaviour is all rather mysterious to me. The old "Secrets of the Inliner" paper is very much out of date now, but short of reading the source, I don't know where else to look to turn my voodoo folk intuition into something more solid. Is this all going to change in 6.12 anyway?

1 0

Help Haskell driving Medical Instruments
by Philippos Apolinarius 11 Nov '09

11 Nov '09

Recently, I received as gift medical instruments designed by one of my father former students. There is a description of these instruments on my web page. Here is the address: http://www.discenda.org/med/ By the way, I am not that guy that appears in a picture wearing emg sensors. That said, the instruments and everything else are programmed in Clean. Then I have a new opportunity of translating Clean programs to Haskell and test them in a real application application. Of course, I simplified the programs to see how they work. The medical instruments have on-board computers, that record signals (electromyograms, electroencephalograms, electrocardiograms, end-tydal CO2 partial pressure and temperature), pre-process them and send them to the main computer. The main-computer recognizes patterns in the signals, and use the result to drive a wheelchair, or to call a doctor. On my page you wil find more complete explanations and pictures of the instruments. For ready reference, here is my address: http://www.discenda.org/med/ I decided start my translation work from the most simple programs, the graphical interface and the communication protocol. After substituting a Haskell program for the Clean original, I discovered that the system did not work anymore if I exited the Haskell program. In few words, after leaving the Haskell program without turning off the computer or the sensors, and entering the Haskell program again, Haskell failed to communicate with the sensors. I did what I always do in such a situation: I simplified the program until I reduced it to a few lines. I discovered that Haskell failed to close the serial port. There is a serial to UART-0 driver that allows me to plug the serial cable to a USB port, that both feeds the sensors, and permit communication. I fixed the bug by passing a useless integer argument to the function used to close the port. Since I don't like this kind of patch (useless arguments), I would like to know why the original program does not work, and also I would appreciate if someone could suggest a way to get rid of the argument whose sole job is force Haskell to close the port. The GUI is based on the Small JAPI biding, fixed and incremented with text processing components. Here is the fixed Haskell program: import Gui.Binding import Gui.Types import Gui.Constants import SER.IAL import Control.Monad import Data.Char main = do rv <- j_start frame <- j_frame "Sensors" exit_button <- j_button frame "Exit" j_setpos exit_button 50 50 j_setsize exit_button 80 30 fld <- j_textfield frame 30 j_setpos fld 50 100 j_show frame opencport(4) waitForFrameAction frame fld exit_button let r = closecport 7 {- without the argument, closecport does not work -} print r return j_quit waitForFrameAction :: Frame -> Object -> Object -> IO Bool waitForFrameAction frame f b = do obj <- j_nextaction again <- if obj == event b then return False else do {- nm <- j_gettext f 200 -} tx <- sendMessage 1 "t" let tp= filter (> ' ') tx rx <- sendMessage 1 "x" let rd= filter (> ' ') rx let x = hex2dec rd let tt= (fromIntegral x)*209.0/1024 - 67.5 j_settext f ((show tt)++" ==> "++tp) return True if not again then return True else waitForFrameAction frame f b hex2dec :: String -> Int hex2dec h= sum (zipWith (*) (map (16^) [3,2,1,0]) [digitToInt c | c <- h]) convert d r s0= (fromIntegral (hex2dec d))*r/1024.0- s0 As I told before, let r = closecport 7 did not work until I gave it an argument. Here is the interface between the C-side, and the Haskell-side of the program: {-# LANGUAGE ForeignFunctionInterface #-} module SER.IAL where import Control.Monad import Foreign import Foreign.C.Types import Foreign.C foreign import ccall "rs232.h opencport" opencport :: CInt -> IO () foreign import ccall "rs232.h closecport" closecport :: CInt -> CInt foreign import ccall "rs232.h rdrs232" c_sendmsg :: CInt -> CString -> CString sendMessage :: Int -> String -> IO String sendMessage n msg = withCString msg $ \str -> peekCString (c_sendmsg (fromIntegral n) str) Originally, I had the following line (that did not work properly): foreign import ccall "rs232.h closecport" closecport :: IO () You will find below the C-program. The original program (that did not work) had the following definition for closecport: int closecport() { CloseComport(); return 3; } This deffinition (that did not work) was replaced by the following one: int closecport(int n) { CloseComport(); return n; } Here is the complete C program: #include "serial.h" #include <string.h> #include <stdio.h> /* Possible baudrates on a normal pc: 50, 75, 110, 134, 150, 200, 300, 600, 1200, 1800, 2400, 4800, 9600, 19200, 38400, 57600, 115200 */ #define BAUD "baud=9600 data=8 parity=N stop=1" HANDLE Cport; char comports[16][10]={"\\\\.\\COM1", "\\\\.\\COM2", "\\\\.\\COM3", "\\\\.\\COM4", "\\\\.\\COM5", "\\\\.\\COM6", "\\\\.\\COM7", "\\\\.\\COM8", "\\\\.\\COM9", "\\\\.\\COM10", "\\\\.\\COM11", "\\\\.\\COM12", "\\\\.\\COM13", "\\\\.\\COM14", "\\\\.\\COM15", "\\\\.\\COM16"}; int OpenComport(int comport_number) { if(comport_number>15) { printf("illegal comport number\n"); return(1); } Cport = CreateFileA(comports[comport_number], GENERIC_READ|GENERIC_WRITE, 0, /* no share */ NULL, /* no security */ OPEN_EXISTING, 0, /* no threads */ NULL); /* no templates */ if(Cport==INVALID_HANDLE_VALUE) { printf("unable to open comport\n"); return(1); } DCB port_settings; memset(&port_settings, 0, sizeof(port_settings)); /* clear the new struct */ port_settings.DCBlength = sizeof(port_settings); if(!BuildCommDCBA(BAUD, &port_settings)) { printf("unable to set comport dcb settings\n"); CloseHandle(Cport); return(1); } if(!SetCommState(Cport, &port_settings)) { printf("unable to set comport cfg settings\n"); CloseHandle(Cport); return(1); } COMMTIMEOUTS Cptimeouts; Cptimeouts.ReadIntervalTimeout = MAXDWORD; Cptimeouts.ReadTotalTimeoutMultiplier = 10; Cptimeouts.ReadTotalTimeoutConstant = 10; Cptimeouts.WriteTotalTimeoutMultiplier = 10; Cptimeouts.WriteTotalTimeoutConstant = 10; if(!SetCommTimeouts(Cport, &Cptimeouts)) { printf("unable to set comport time-out settings\n"); CloseHandle(Cport); return(1); } return(0); } int PollComport(unsigned char *buf, int size) { int n; if(size>4096) size = 4096; /* added the void pointer cast, otherwise gcc will complain about */ /* "warning: dereferencing type-punned pointer will break strict aliasing rules" */ ReadFile(Cport, buf, size, (LPDWORD)((void *)&n), NULL); return(n); } int RdByte(unsigned char* m) { int n; ReadFile(Cport, m, 1, (LPDWORD)((void *)&n), NULL); return(n); } int SendByte(unsigned char byte) { int n; WriteFile(Cport, &byte, 1, (LPDWORD)((void *)&n), NULL); if(n<0) return(1); return(0); } int SendBuf(unsigned char *buf, int size) { int n; if(WriteFile(Cport, buf, size, (LPDWORD)((void *)&n), NULL)) { return(n); } return(-1); } int CloseComport(void) { CloseHandle(Cport); return(0); } int IsCTSEnabled(void) { int status; GetCommModemStatus(Cport, (LPDWORD)((void *)&status)); if(status&MS_CTS_ON) return(1); else return(0); } int cprintf(const char *text) /* sends a string to serial port */ { while(*text != 0) SendByte(*(text++)); return(0); } int opencport(int p) { OpenComport(p-1); return 3; } int closecport(int n) { CloseComport(); return n; } char* rdrs232(int n, char* msg) { char *str; char mm; int i, j; for (j=0; j<n; j++) { SendByte(msg[j]); } str = (char *) malloc(16000); i=0; mm=0; while (mm != 10) { RdByte(&mm); str[i]= mm; i= i+1; } if (i>0 && str[0]==0) { str[0]= ' ';} return(str); } Finally, here is the serial.h file: #ifndef rs232_INCLUDED #define rs232_INCLUDED #ifdef __cplusplus extern "C" { #endif #include <stdio.h> #include <string.h> #ifdef __linux__ #include <termios.h> #include <sys/ioctl.h> #include <unistd.h> #include <fcntl.h> #include <sys/types.h> #include <sys/stat.h> #include <limits.h> #else #include <windows.h> #endif int OpenComport(int); int PollComport(unsigned char *, int); int SendByte(unsigned char); int SendBuf(unsigned char *, int); int CloseComport(void); int cprintf(const char *); int IsCTSEnabled(void); char *topa(int n); int opencport(int p); int closecport(int n); char* rdrs232(int n, char* msg); #ifdef __cplusplus } /* extern "C" */ #endif #endif __________________________________________________________________ Make your browsing faster, safer, and easier with the new Internet Explorer® 8. Optimized for Yahoo! Get it Now for Free! at http://downloads.yahoo.com/ca/internetexplorer/

4 8