- Glasgow-haskell-users

RE: Problems creating a DLL with ghc-5.02.1
by Simon Peyton-Jones 06 Dec '01

06 Dec '01

I know that Sigbjorn is working on this. We will do 5.02.2 in due course, but I'm not yet sure how dll-friendly is will be. I don't understand all the issues yet. Simon | -----Original Message----- | From: Lehti, Patrick [mailto:Patrick.Lehti@softwareag.com] | Sent: 06 December 2001 14:05 | To: 'glasgow-haskell-users(a)haskell.org' | Subject: Problems creating a DLL with ghc-5.02.1 | | | Hi, | | I tried to wrap a Haskell application in a DLL using | ghc-5.02.1. But this is not working, with cygwin in path I | got this error message: | /usr/lib/mingw/libmingw32.a(gccmain.o)(.text+0x88):gccmain.c: | multiple definition of `__main' | /usr/lib/libm.a(ds00023.o)(.text+0x0): first defined here | | ... and without cygwin in path, it could not find "dllwrap" | and "dlltool". | | Is there any solution for this problem? | | Patrick | | _______________________________________________ | Glasgow-haskell-users mailing list | Glasgow-haskell-users(a)haskell.org | http://www.haskell.org/mailman/listinfo/glasgow-| haskell-users |

1 0

Communicating with other programs (Win32)
by Jan de Wit 05 Dec '01

05 Dec '01

Hi all, In my free time, I'm writing a Go program in Haskell. Main platform is GHC and GHCi under Win32. It is coming along nicely: it has a GUI and the moves it comes up with are starting to make some sense. For testing (and maybe later training :) purposes I want it to be able to play against GnuGo (the strongest Open Source Go program) on its own. Currently, to get GnuGo to think up a move, I use something like (SGF is the standard way of outputting Go games): | do writeFile "in.sgf" (showAsSGF game) | let cmd = "echo quit > gnugo --quiet --infile in.sgf --color " ++ | (show computerColor) ++ | " --mode ascii > out.txt" | system cmd | result <- readFile "out.txt" | return (parseResultMove result) I think this both ugly and inefficient: for every move a new instance of GnuGo gets started and an SGF file is output! Instead, I want to start GnuGo only once per game and communicate with it like a human player would, by typing commands at it; or by using the Go Text Protocol which is specifically designed for computer-computer play. Does anybody here know a way - under Win32 - to start a process in the background and communicate with it via its stdin/stdout? I've poked around a bit in the Win32 library source, looking for an FFI interface to CreateProcess (the API function which does what I want) but found nothing. The Posix library looks like it should be up to the job, but unfortunately that isn't supported under Win32... (I'm not using Cygwin) Any help would be appreciated! Cheers, Jan de Wit

2 1

Re: [qforeign] CursesTest segfaults
by Marcin 'Qrczak' Kowalczyk 04 Dec '01

04 Dec '01

29 Nov 2001 18:12:08 +0900, Jens Petersen <juhp(a)01.246.ne.jp> pisze: > I built qforeign from cvs under Linux. Any idea why > tests/CursesTest segfaults? No idea - works for me now. ghc-5.03 ncurses-5.2 glibc-2.2.3 gmp-3.1.1 readline-4.2 -- __("< Marcin Kowalczyk * qrczak(a)knm.org.pl http://qrczak.ids.net.pl/ \__/ ^^ QRCZAK

2 1

RE: GHC 5.02.1 on Mac OS X
by Simon Peyton-Jones 03 Dec '01

03 Dec '01

Sebastien Carlier [sebc(a)posse42.net] was at one time working on a MacOSX port, but I don't know how far he got. There was one other person too, but I can't lay my hands on appropriate email. Simon | -----Original Message----- | From: Ashley Yakeley [mailto:ashley@semantic.org] | Sent: 02 December 2001 22:50 | To: Nicolas Oury; GHC List | Subject: Re: GHC 5.02.1 on Mac OS X | | | At 2001-12-02 14:39, Nicolas Oury wrote: | | >do you know if there is a port of gh 5.02.1 on MacOSX? | | The latest I know of is GHC 5.00.2. And even that is incomplete, it | doesn't seem to include a createAdjustor for 'foreign export dynamic' | functions (if you happen to need that, as I do). | | -- | Ashley Yakeley, Seattle WA | | | _______________________________________________ | Glasgow-haskell-users mailing list | Glasgow-haskell-users(a)haskell.org | http://www.haskell.org/mailman/listinfo/glasgow-| haskell-users |

4 7

RE: ghc and hierarchical module names
by Simon Marlow 03 Dec '01

03 Dec '01

[ thread moved from glasgow-haskell-users(a)haskell.org to libraries(a)haskell.org ] > "Simon Marlow" <simonmar(a)microsoft.com> writes: > > > You must give the full module name in the header, eg. > > > module A.B.C where > > This surprised me somewhat, and I must have missed any > discussion/rationale. Could you elaborate/point me at the relevant > information? The discussion was on libraries(a)haskell.org (take a look at the archives), and the motivation was to make the smallest change possible to Haskell itself. With the current proposal, no changes are required to the Haskell report except for the 1-character change required to the lexical syntax of module names. I agree that it might be useful to be able to abbreviate module names by leaving out an initial prefix, so that entire subtrees can be moved without changing the source code. However, there are reasons why this approach is not so nice: as you point out, this introduces a dependency between the source code and the external environment, such that the meaning of a program depends on the directory layout and options to the compiler (currently the meaning depends on just the source to the collection of modules constituting the program), and also there is an inconsistency in that you would declare a module as 'C' but import it as 'A.B.C'. Cheers, Simon

1 0

Re: ghc and hierarchical module names
by Ketil Z Malde 03 Dec '01

03 Dec '01

"Simon Marlow" <simonmar(a)microsoft.com> writes: > You must give the full module name in the header, eg. > module A.B.C where This surprised me somewhat, and I must have missed any discussion/rationale. Could you elaborate/point me at the relevant information? It seems to me that A.B could be implicit (leaving "module C where"), from the path. This might help resolve point 3.2[0] Renaming of Subtrees, since one could add -I (or equivalent options, or modify a path in an env var) for commonly used directories. (Okay, so it would introduce a dependency between the importing source and the build process. Is that it?) -kzm [0] ...suggesting a syntax along import Org.Com.Microsoft.* as Microsoft.* import Microsoft.Foo import Microsoft.J++ as Java giving a shorthand access to deeply nested libraries. -- If I haven't seen further, it is by standing in the footprints of giants

1 0

RE: ghc and hierarchical module names
by Simon Marlow 03 Dec '01

03 Dec '01

> The release notes for ghc 5.02.1 say: > > 1.4.4. New experimental features > > Partial support for hierarchical module names. > > - but I could not find a document describing how to use this feature. > > Is it possible to use hierarchical module names with ghc at > present, and if so, > how does one write and use a hierarchically-named module ? Module names may include the '.' character with -fglsagow-exts on. Searching for interfaces to include is extended as follows: if our search path contains directories D1...Dn, then a module A.B.C is searched for in D1/A/B/C.hi ... Dn/A/B/C.hi. You must give the full module name in the header, eg. module A.B.C where This support *is* experimental though, so it is subject to change. Also note that certain parts of the hierarchy are reserved for system-supplied libraries - see the current proposal at http://www.haskell.org/~simonmar/libraries/libraries.html Cheers, Simon

1 0

Re: GHC 5.02.1 on Mac OS X
by Ashley Yakeley 02 Dec '01

02 Dec '01

At 2001-12-02 14:39, Nicolas Oury wrote: >do you know if there is a port of gh 5.02.1 on MacOSX? The latest I know of is GHC 5.00.2. And even that is incomplete, it doesn't seem to include a createAdjustor for 'foreign export dynamic' functions (if you happen to need that, as I do). -- Ashley Yakeley, Seattle WA

1 0

GHC 5.02.1 on Mac OS X
by Nicolas Oury 02 Dec '01

02 Dec '01

Hi everybody, do you know if there is a port of gh 5.02.1 on MacOSX? How difficult is it to do a port? Just compiling the source or modifying code of the RTS and the code generation? Is this support planned to be kept in the future? Best regards, Nicolas Oury

1 0

Re: FORTRAN and HASKELL
by Sigbjorn Finne 02 Dec '01

02 Dec '01

Hi, you're in luck then, as the basic ABI is identical that of C (gcc), since g77 shares the same backend. Here's the simple example I tested it with: foo$ cat square.f SUBROUTINE SQUARE(N,M) C COMPUTES THE SQUARE OF N, RETURNS IN M M=N*N RETURN END C foo$ cat main.hs module Main where import Ptr import Storable import MarshalUtils import MarshalAlloc foreign import "square_" unsafe square_ :: Ptr Int -> Ptr Int -> IO Int square :: Int -> IO Int square x = withObject x $ \ ptr_x -> alloca $ \ ptr_res -> do square_ ptr_x ptr_res peek ptr_res main = square 11 >>= print {- --------------------- -} foo$ g77 -c square.f foo$ ghc -o main main.hs square.o -fglasgow-exts -package lang To get the name mangling and the details of passing arguments to the Fortran subroutine right, I looked at the output of "f2c -P", which gives back C prototypes for Fortran function/subs. You could certainly imagine a tool that would automate all this.. hth --sigbjorn ----- Original Message ----- From: "Heron" <heron_carvalho(a)bol.com.br> To: "Sigbjorn Finne" <sof(a)galois.com> Sent: Sunday, December 02, 2001 02:43 Subject: Re: FORTRAN and HASKELL > Thank you for your comments. Answering your question ... > > > you don't say what platform this is on (or what Fortran compiler > > you're using). > > We are using GNU Fortran 77 Compiler (f77) on a Linux PC cluster based > on Red Hat 6.2. The GHC version used is 5.02. > > Best Regards, > Heron de Carvalho

1 0