November 2001 - Glasgow-haskell-users

RE: GHC parsers
by Simon Marlow 22 Nov '01

22 Nov '01

[ moved to glasgow-haskell-users(a)haskell.org ] > we would like to adapt a Haskell parser for whatever > purposes, and looked > for one in the ghc distribution. However, in ghc 5.02, we found two > happy-generated parsers, one in ghc/compiler/parser and one in > hslibs/hssource which both look similar, but seem to be > slightly different > (at least their version numbers are). Could someone tell me what the > differences are and which one we really ought to use? The parser in hslibs/hssource is designed to be portable (it's written in standard Haskell 98). GHC's parser, and in particular the lexer, by contrast uses several GHC extensions. Also, the hssource parser doesn't handle any GHC extensions to the Haskell syntax. If you're looking for a standalone Haskell parser, then hslibs/hssource is almost certainly the way to go. Cheers, Simon

1 0

RE: hdirect-0.18 comserv example
by Lehti, Patrick 22 Nov '01

22 Nov '01

I could figure out the problems mentioned earlier. The type library problem is easy fixed by switching on the SUPPORT_TYPELIBS option in the src/Makefile during the build of hdirect. Note that you cannot switch on this option in a first time build. The link error has something to do with cygwin, when I remove cygwin from my PATH, this error disappears. But then unfortunately no dll file is built at all, without any error message! So I am still waiting for a solution... Patrick -----Original Message----- From: Janna Khegai [mailto:janna@cs.chalmers.se] Sent: Montag, 12. November 2001 00:47 To: glasgow-haskell-users(a)haskell.org Cc: Patrick.Lehti(a)softwareag.com Subject: hdirect-0.18 comserv example I've tried to build the comserv example with hDirect-0.18 and ghc-5.02 under Windows ME. I've encountered the problems similar to ones mentioned earlier in: http://www.haskell.org/pipermail/glasgow-haskell-users/2001-October/000974.h tml <http://www.haskell.org/pipermail/glasgow-haskell-users/2001-October/000974. html> : >problems generating the type library: it is simply not created. and http://www.haskell.org/pipermail/glasgow-haskell-users/2001-October/000979.h tml <http://www.haskell.org/pipermail/glasgow-haskell-users/2001-October/000979. html> : > a problem linking the DLL, I always get 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 If anybody already knows the cure to these problems, help me, please! thanks in advance, Janna Khegai

1 0

Re: Socket library ghc 5.02.1
by Sven Eric Panitz 22 Nov '01

22 Nov '01

From: Volker Stolz <stolz(a)hyperion.informatik.rwth-aachen.de> Cc: glasgow-haskell-users(a)haskell.org In local.glasgow-haskell-users, you wrote: >> main = >> do >> d <- connectTo "localhost" (PortNumber 80) >> hPutStr d "GET / HTTP/1.0\n\n" >> hFlush d >> c <- hGetContents d >> putStr c > > whereas on Linux I get the following error: > *** Exception: failed > Action: connect > Reason: Unknown error 142731264 Please run the program with 'strace', e.g. strace -o log ./a.out and post the log when it fails. Be aware that 'hGetContents' is broken on 5.02.1 if you use it extensively (the GC will close already-reused fds and cause your program to crash!), the patch was recently postet to -bugs by Simon Marlow and works. OTOH, this is not the bug you're seeing. As I just figured out, reported error above only occurs within ghci. Otherwise Sockets seem to work on ghc 5.02.1 for Linux. Sven Eric

1 0

hsc: Passing -1 as CString/Ptr a?
by Volker Stolz 22 Nov '01

22 Nov '01

Hi, I want to invoke the C-function 'dlsym', which has three possible options for a parameter of type (void*): - NULL - void* to a string, i.e. CString in Haskell - RTLD_NEXT = (void *)(-1) The first two ones are no problem, I can use a CString and 'nullPtr'. For the last one, I have the problem that in the "Next"-case I cannot write pack :: Flag -> CString pack Null = nullPtr pack Name str = str pack Next = #const RTLD_NEXT because there's no way to cast Int to a Ptr -- at least I didn't find any. Unluckily, the C-codes (void *)(-1) even gets expanded by hsc to the literal "4294967295" which can't be right either. So currently I'm facing two possible solutions: pack Next = nullPtr `plusPtr` 4294967295 (expanded from "... #const RTLD_NEXT") or pack Next = nullPtr `plusPtr` (-1) (which is not a good idea since it doesn't use the C-constant) "Ptr" is abstract so I cannot create this Ptr myself. Maybe the Ptr-module just lacks a 'minusOnePtr' similar to 'nullPtr'? -- Volker Stolz * vs(a)foldr.org * PGP + S/MIME

1 0

RE: Binary library
by Jan de Wit 22 Nov '01

22 Nov '01

> > Does Malcolm's Binary library exist for ghc? If not, is there > > a standard > > way (ie using Haskell that works on any compiler) to dump a data > > structure into a file that another program or part of the same program > > can then pick up and read (cf "serialize" in Java)? I'm > > reluctant to go > > with a solution that's compiler-specific. > > Using Show & Read is the only really portable way to do this, > unfortunately. [snip] Well, it's not really portable... Try the program at the end of this file, by first executing hugsMain in hugs, exiting and then running main in ghc or ghci, and you'll see: | Cons{hd='a',tl=(Cons{hd='b',tl=Nil})} | *** Exception: Prelude.read: no parse So, at least ghc and hugs disagree... My ghc version is 5.02 and my hugs is February 2000/2001 (both give the same output). Cheers, Jan de Wit ----8<---- module Test where data List a = Nil | Cons { hd :: a, tl :: List a } deriving (Show,Read,Eq) theList = Cons 'a' (Cons 'b' Nil) hugsMain = do writeFile "test.txt" (show theList) main main = do s <- readFile "test.txt" let theListFromFile = read s print theList print theListFromFile print $ theList == theListFromFile writeFile "test.txt" (show theList)

3 5

RE: Binary library
by Jan de Wit 21 Nov '01

21 Nov '01

> Can anyone spot the fault in the derived instance of Read produced > by DrIFT? > > Regards, > Malcolm The precedence of things printed or read after a label should IMO be lower than 10, probably even 0. In other words: calls to showsPrec/readsPrec 10 should be replaced by showsPrec/readsPrec 0 in the code below. I haven't checked this though, however I've written a generic version of read and show using Generic Haskell which omits the superfluous parentheses by using precedence 0. Cheers, Jan > > > data List a = Nil | Cons { hd :: a, tl :: List a } > > > > {-* Generated by DrIFT-v1.0 : Look, but Don't Touch. *-} > > instance (Show a) => Show (List a) where > > showsPrec d (Nil) = showString "Nil" > > showsPrec d (Cons aa ab) = showParen (d >= 10) > > (showString "Cons" . showChar '{' . > > showString "hd". showChar '='. showsPrec 10 aa. showChar',' . > > showString "tl" . showChar '=' . showsPrec 10 ab > > . showChar '}') > > > > instance (Read a) => Read (List a) where > > readsPrec d input = > > (\ inp -> [((Nil) , rest) | ("Nil" , rest) <- lex inp]) input > > ++ > > readParen (d > 9) > > (\ inp -> > > [((Cons aa ab) , rest) | ("Cons" , inp) <- lex inp , > > ("{" , inp) <- lex inp , ("hd" , inp) <- lex inp , > > ("=" , inp) <- lex inp , (aa , inp) <- readsPrec 10 inp , > > ("," , inp) <- lex inp , ("tl" , inp) <- lex inp , > > ("=" , inp) <- lex inp , (ab , inp) <- readsPrec 10 inp , > > ("}" , rest) <- lex inp]) > > input

1 0

[Linux-basic-compiling]GHC-3.02 build on GHC-2.10
by Rafael Martínez Torres 21 Nov '01

21 Nov '01

I apologize if this is not the correct moment-place to deal with this old version. I'm trying to compile the GHC-3.02 version using ghc-2.10 binaries, happy-1.6 binaries on a i386-unknown-linux. In the future I would like to hack with it. I had to modify the file ghc/run_time/main/Signals.lc because it used a non-existing structure, named "struct sigcontext_struct" instead of "struct sigcontext". Main problem arise when linking. If I'm not wrong, the output from Haskell code are "C" files. They are compiled and linked by the C compiler. At the final linking step, the linker shows me unresolved references to symbols: "_fxstat","_lxstat" or "__setmjmp". After inspecting the code, on directory ghc/lib/cbits/openFile.c [...], I realized this belong to C functions fstat(),lstat() and setjmp() . These libraries are implicitly included by libc, I think. Any C-program links without problem,when compiled by invoking explictly "gcc", but I seems the ghc-2.10 compiler not to find the libc (and so his sub-components ). Does any body suggest me what I can do to perform a good compiling ? What parameters should I modify and where ? In directory mk ? Where lasgow-haskell-users(a)haskell.org else ?

1 0

RE: [Linux-basic-compiling]GHC-3.02 build on GHC-2.10
by Simon Marlow 21 Nov '01

21 Nov '01

> I apologize if this is not the correct moment-place to deal with this > old version. > > I'm trying to compile the GHC-3.02 version using ghc-2.10 binaries, > happy-1.6 binaries on a i386-unknown-linux. > In the future I would like to hack with it. [ I presume you have a good reason for needing GHC 3.02 - I'd be interested to know what it is! ] It looks like that the ghc-2.10 binaries you are using don't work well with the version of Linux (in particular the version of the C library) that you're using. Probably the only way to get going is to bootstrap from the .hc files for 3.02. Cheers, Simon

1 0

RE: Binary library
by Simon Marlow 21 Nov '01

21 Nov '01

> Does Malcolm's Binary library exist for ghc? If not, is there > a standard > way (ie using Haskell that works on any compiler) to dump a data > structure into a file that another program or part of the same program > can then pick up and read (cf "serialize" in Java)? I'm > reluctant to go > with a solution that's compiler-specific. Using Show & Read is the only really portable way to do this, unfortunately. Sven Panne did a port of nhc's Binary library a while back, which you can find on haskell.org, but it will certainly need some work to get it going on recent GHC's. We'd really like a good serialisation library, BTW, if anyone out there feels like donating one! Cheers, Simon

1 0

GMP
by Marcin 'Qrczak' Kowalczyk 20 Nov '01

20 Nov '01

I think there exists a problem. I haven't encountered it because I haven't tried to provoke it, but if I'm not mistaken it can cause big trouble. Programs compiled by ghc set gmp's allocation functions to ones which allocate byte arrays on the Haskell heap, instead of the default malloc/free. Now imagine a program or library which uses gmp for its own purposes and is linked with a program or library compiled with ghc. GMP's hooks are global, gmp is shared. Neither of allocation methods is suitable for the other side. So it won't work. Is there a solution? -- __("< Marcin Kowalczyk * qrczak(a)knm.org.pl http://qrczak.ids.net.pl/ \__/ ^^ QRCZAK

2 1