March 2002 - Glasgow-haskell-users

Re: GHC on MacOS X
by Wolfgang Thaller 03 Mar '02

03 Mar '02

> I realize that you would rather clean things up first, but could I talk > you > into passing on your patches? We could use a build on MacOS X today (and > I > literally mean today). It's not easy... when I say "unclean", I _mean_ unclean. However, here it is... I even added a new file in a place where it definitely doesn't belong - the attached archive contains a "cvs diff" plus the extra file, sigaddsetwtt.h (wtt are my initials), which belongs in libraries/base/. One more thing: the build system passed --split-objs to the compiler for one or two library modules (which doesn't work for unregisterd builds). Not knowing much about the build system, I simply copied the failed command, removed the --split-objs and ran it by hand. Then just ran make again, and everthing went fine. > Interestingly enough, [...] > checked in the changes today. Perhaps the problems that you noticed with > your own modules are related to the problems that Sigbjorn fixed. Thanks for the hint, I'll "check that out" (pun intended). Wolfgang -- GMX - Die Kommunikationsplattform im Internet. http://www.gmx.net -- GMX - Die Kommunikationsplattform im Internet. http://www.gmx.net -- GMX - Die Kommunikationsplattform im Internet. http://www.gmx.net

1 0

Re: GHC on MacOS X
by Ashley Yakeley 02 Mar '02

02 Mar '02

At 2002-03-02 14:21, Wolfgang Thaller wrote: >The results: another unregistered build, a PowerPC implementation >of createAdjustor (foreign export dynamic now works!). Excellent! I will be porting JVM-Bridge to it as soon as it's ready. -- Ashley Yakeley, Seattle WA

1 0

GHC on MacOS X
by Wolfgang Thaller 02 Mar '02

02 Mar '02

Hello all, I've spent a day compiling ghc (5.03, checked out from CVS a few days ago) on MacOS X 10.1. The results: another unregistered build, a PowerPC implementation of createAdjustor (foreign export dynamic now works!). I needed to modify a total of six files: File: Adjustor.c Status: Locally Modified File: ClosureFlags.c Status: Locally Modified File: Main.c Status: Locally Modified File: Storage.h Status: Locally Modified File: Posix.hs Status: Locally Modified File: Socket.hsc Status: Locally Modified Most of my modifications are ugly hacks, I still need to clean them up. I would appreciate some advice on how to do that and how to submit them for inclusion. Yes, I can read the man page for diff, but are there other things I need to know when sending in a patch? Socket.hsc refused to compile because on MacOS X, sigaddset() is a macro that chokes if its parameter is a void*. I added a wrapper function in a small C header file, but that's really ugly. What's the "proper" way? And for some reason, Apple's version of gcc refused to accept the non-standard syntax used in ClosureFlags.c. My modifications in Main.c and Storage.h are closely based on those in the MacOS X compile of GHC 5.00.2. I also got a compiler panic on one of my own modules I tried to compile. I'll post that to ghc-bugs, as I somehow don't think it has to do with the PPC port. What I really need, of course, is a registered compiler... I'm even prepared to learn some Perl to do that, but for now I don't really have a clue what to do - I don't know enough about Intel/Sparc assembly language and runtime conventions to find out from the code without a few hints - can anyone give me a few pointers what the mangler is supposed to do _exactly_? CU, Wolfgang Thaller -- GMX - Die Kommunikationsplattform im Internet. http://www.gmx.net

1 0

client/server library
by Hal Daume III 01 Mar '02

01 Mar '02

I'm wondering if anyone has written anything akin to a client/server library which is based on sockets, something along the lines of: > mkServer :: Int -> (String -> String) -> IO () so I could then say: > module MyServer where > > main = mkServer portNum myServerFn > > myServerFn "getname" = "This server's name is Bob" > myServerFn _ = "Unknown command" and > data Server = ... > serverConnect :: Int -> IO Server > queryServer :: Server -> String -> IO String > serverDisconnect :: Server -> IO () so I could write: > module MyClient where > > main = > do s <- serverConnect portNum > res <- queryServer "getname" > putStrLn res > serverDisconnect s and this would print "This server's name is Bob". I don't imagine such a "library" would be difficult to write; it would mostly be wrapping around the Socket library (the assumption would be that the server and client would both be running on the same machine). But it would need to be robust (i.e., the server should my multithreaded among other things), etc. I could probably figure out how to write some such thing, but if someone else who knows more about the net and concurrent packages already has (I know nothing), I'd love to steal their code :) - Hal -- Hal Daume III "Computer science is no more about computers | hdaume(a)isi.edu than astronomy is about telescopes." -Dijkstra | www.isi.edu/~hdaume

1 0

RE: Weird symbol prevents GHCi from loading my package.
by Simon Marlow 01 Mar '02

01 Mar '02

> I tried to build the head of GHC and it works fine the first > time around. > Compiling it with itself fails to build GHCi due to the new library > system: > > /home/as49/local/bin/ghc-5.03 -DGHCI -cpp -fglasgow-exts > -Rghc-timing -I. > -IcodeGen -InativeGen -Iparser > -iutils:basicTypes:types:hsSyn:prelude:rename:typecheck:deSuga > r:coreSyn:specialise:simplCore:stranal:stgSyn:simplStg:codeGen > :absCSyn:main:profiling:parser:usageSP:cprAnalysis:compMan:ndp > Flatten:nativeGen:ghci > -package concurrent -package posix -package util -recomp -Rghc-timing > -H16M -O -fvia-C -monly-3-regs -c utils/PrimPacked.lhs -o > utils/PrimPacked.o > > utils/PrimPacked.lhs:55: > Ambiguous occurrence `unsafePerformIO' > It could refer to either `PrelIOBase.unsafePerformIO', > imported from > Foreign at utils/PrimPacked.lhs:40 > or `GHC.IOBase.unsafePerformIO', > imported from > GlaExts at utils/PrimPacked.lhs:33 This is very odd - it looks like you've got the .hi files for two versions of GHC mixed up somehow. There shouldn't be a PrelIOBase at all in the latest GHC, and yet it seems to be picking one up from somewhere. When doing a 2-stage bootstrap you can run the first stage directly from the build tree without installing it - just point the second stage at ghc/compiler/ghc-inplace in the first stage (i.e. ./configure --with-ghc=/some/where/ghc/compiler/ghc-inplace). Cheers, Simon

1 0

Windows: calling a process from Haskell
by George Russell 01 Mar '02

01 Mar '02

How do I start up a new process in Windows without stopping the world (as I presume System.system will do)?

1 0

RE: Weird symbol prevents GHCi from loading my package.
by Simon Marlow 01 Mar '02

01 Mar '02

> I built quite a big library and turned it into a package. It > works fine if > use ghc -package gtk2 Blah.hs to compile things. But when I > convert all > the libraries to object files as described in the documentation, GHCi > refuses to load the whole machinery: > > ~/source/mogul/fpDemo:$ ghci -package gtk2 Arabic.hs > ___ ___ _ > / _ \ /\ /\/ __(_) > / /_\// /_/ / / | | GHC Interactive, version 5.02.2, > for Haskell > 98. > / /_\\/ __ / /___| | http://www.haskell.org/ghc/ > \____/\/ /_/\____/|_| Type :? for help. > > Loading package std ... linking ... done. > Loading package lang ... linking ... done. > Loading package gtk2 ... > > GHCi runtime linker: fatal error: I found a duplicate definition for > symbol > gcc2_compiled. > whilst processing object file > /home/as49/local/lib/gtk2hs/gtk2hs.o I think this is due to a bug in GHCi's linker, which has been fixed post 5.02.2. I tripped over it recently while trying to use Gtk+HS in GHCi on my RedHat 7.2 box. There are two bugs tickled by gcc 2.96: our linker wasn't dealing with sections with names other than the standard ones (.text, .data, .rodata) properly, and there was also some infelicities in the local-symbol handling (Julian S. can tell you more). The workaround I used is to compile up any .c files in your package *without* -O, and without -g if you were also using that. Cheers, Simon

2 1