- Nhc-users - Haskell.org

Native code generation
by Arunkumar S Jadhav 25 Nov '04

25 Nov '04

Hello All, As I understand it, nhc98 generates byte-code out of a program that is interpreted by mutator. I've developed a native code generator (for i386) that takes as input G-code which is described in SPJ's book. Now I would like to do the same for G-code produced by nhc98. I would like to know. 1) Has any one tried this (nhc G-code to native code generation) ? If yes then it would be great if we can share some of the ideas. 2) Is it possible to generate native code fully *just* by looking at G-code and symbol table generated by nhc98 ? Any pointers / suggestions would be really helpful. Regards, -- Arunkumar S Jadhav, Masters Student, KReSIT, IIT-Bombay, India Ph: +91-22-25764967 http://www.it.iitb.ac.in/~arunk

2 10

hmake 3.09
by Ian Lynagh 24 Nov '04

24 Nov '04

Hi Malcolm, I happened to notice a new release of hmake was out - thanks for doing this! However, unfortunately it seems that the patch I sent here: http://www.haskell.org//pipermail/nhc-users/2003-September/000099.html didn't make it into CVS :-( I've attached the patch for 3.09 (it has a few other small tweaks I made too). In src/hmake/Imports.hs does "stringgap (c:cs) = stringgap cs" deliberately not check isSpace c? Hmm, there's no [] case either. There is a docs/hmake/.#index.html.1.30 in the tarball - it looks like you have lost the change of hi copyright from 2000 to 2000-2003. Thanks Ian

2 3

NHC98 bindist for AIX 5.1L available
by John Goerzen 05 Oct '04

05 Oct '04

Hello, Thanks to help from Malcolm Wallace, a build of the nhc98 1.17 snapshot is available for AIX5.1L PowerPC architectures at: gopher://quux.org/1/devel/AIX/nhc98 http://quux.org/devel/AIX/nhc98 I am also uploading a built source tree tarball in case that is useful for anyone. (Wait about an hour for that one to appear) -- John

1 0

Trouble Building NHC98 for AIX5.1L
by John Goerzen 02 Oct '04

02 Oct '04

Hello, I am trying to build NHC98 for AIX5.1L. I have tried 1.16 and the 1.17 snapshot and get similar results either way: /home/jgoerzen/programs/nhc98/nhc98-1.17/script/nhc98 -c +CTS -lib -redefine -C TS +RTS -H32M -RTS -o /home/jgoerzen/programs/nhc98/nhc98-1.17/targets/powerpc -AIX-5.1.0.0/obj/prelude/DErrNo/DErrNo.o DErrNo.hs /home/jgoerzen/programs/nhc98/nhc98-1.17/script/nhc98[452]: 54306 Segmentation fault(coredump) Trying to get tag from unevaluated node in TABLESWITCH at 200a32fd! Node is: 20053057 at 200c653c The program that coredumped is hmake-PRAGMA and the unevaluated node message is coming from nhc98comp. I have tried expanding the --heap argument to configure, but it didn't help. Here's some info about the environment: OS: AIX5.1L CPU: PowerPC64 (but running and compiling in 32-bit mode!) Compiler: GNU GCC 3.3.2 Bootstrapping with: gcc (I have no other Haskell compiler on AIX) Any hints would be greatly appreciated. Thanks! -- John Goerzen

1 0

nhc98 garbage collection core dumps and build problems
by Olaf Chitil 29 Jul '04

29 Jul '04

Malcolm, every non-trivial Haskell program compiled with nhc98 (current cvs) core dumps on Solaris. The garbage collector seems to be the cause of the fault. Small computations not requiring garbage collection work fine. But take: import List plus x = x + 2 primes :: [Int] primes = 2 : filter prime [3..] prime :: Int -> Bool prime n = not (any (`divides` n) (takeWhile (< n) [2..])) divides :: Int -> Int -> Bool divides n m = m `rem` n == 0 main = print (take 33 primes) Then you observe: myrtle$ Test +RTS -B [2,3,5,7,11,13,17,19,23,29,31,37,41,43,47,53,59,61,67,71,73,79,83,89,97,101,103,107,109,113,127,131<GC 1:Start>Segmentation Fault (core dumped) Any idea? With respect to building nhc98 (all). After a long time the process stops with: > /nfs/myrtle/d23/part1/home/cur/oc/nhc98/targets/sparc-solaris2/objz/prelude/tpltmp.c gcc -c -o /nfs/myrtle/d23/part1/home/cur/oc/nhc98/targets/sparc-solaris2/objz/prelude/tpl3Observe.o /nfs/myrtle/d23/part1/home/cur/oc/nhc98/targets/sparc-solaris2/objz/prelude/tpltmp.c rm /nfs/myrtle/d23/part1/home/cur/oc/nhc98/targets/sparc-solaris2/objz/prelude/tpltmp.c /nfs/myrtle/d23/part1/home/cur/oc/nhc98/script/tprofprel 1 \ PackedString System CString Driver NonStdEval NonStdProfile NonStdTrace NonStdUnsafeCoerce NonStdGetPID Warning UnboxedArray BinArray Prelude Array CPUTime Char DErrNo FFI IO IOExtras Ix List Maybe Numeric Random Ratio Bit Binary Locale Observe Complex Directory GreenCard Monad Time > /nfs/myrtle/d23/part1/home/cur/oc/nhc98/targets/sparc-solaris2/objz/prelude/tpltmp.c $ "$" is not a proper shell prompt. It does accept some input like "gmake install" (without doing anything until interrupted), but it doesn't do anything on input like "ls". I'm confused. In the end I only build "basic". A minor point: When make tries to build the package haskell-src it calls happy. Because I don't have happy installed this fails. No big problem, because building continues. However, shouldn't configure test for happy, warn about it and possibly exclude haskell-src from building... Ciao, Olaf

2 2

Disappearing instances?
by Martin Sjögren 29 Jul '04

29 Jul '04

Hello list. Since NHC doesn't distribute Control.Monad.Error that, among other things, defines the Monad (Either e) instance, I wrote the following module for Cabal: http://cvs.haskell.org/darcs/cabal/Compat/H98.hs However, when I compile it, the Monad instance seems to disappear! With a small test program > module Main where > import Monad > import Compat.H98 > foo :: Either Char [()] > foo = sequence [Right (), Left 'a', Right (), Left 'b'] > main = print foo nhc complains: Fail: The class Prelude.Monad has no instance for the type Prelude.Either. and indeed if I look at Compat/H98.hi, it mentions instance Error Prelude.Char; instance (Error a) => Error [a]; but nothing about Error e => Monad (Either e) What should I do? Is this a bug? /Martin

2 1

/usr/lib/hmake/$arch-$os/
by Ian Lynagh 14 Jun '04

14 Jun '04

Hi, hmake uses /usr/lib/hmake/$arch-$os/ to put various files in, but I've had a couple of issues with this. First, as /usr can be mounted read-only I've moved hmakerc to /var/lib/hmake/$arch-$os/hmakerc. I can filter out the bits of the diff that do this (with a --confdir argument to configure) if you are interested. Second, the $arch component as detected by harch can vary between machines using a single Debian arch, e.g. s390 and s390x both use s390. This of course means hmake can't find its files and so thinks it isn't installed. I imagine similar issues will face binary packagers of other distributions in time. I have two ways of fixing this. One is to make a replacement harch that just echos what "dpkg-architecture -qDEB_HOST_ARCH" says instead of what it would currently detect (although I don't have a good replacement for $os currently). The other is to just drop the $arch-$os/ path component completely; the FHS says (well, strongly implies at least) that /usr/lib is architecture-dependent anyway. I imagine the reason you put this in is that there are some systems where /usr/lib is shared between machines of different arches? However, would it not be simpler for the admins of those machines to just give --libdir accordingly? Presumably they already have to do so for other OS software anyway. I think the nicer solution for Debian is the latter, so the reason for my bringing it up is that it will make my life easier if you adopt it too :-) The other use of harch I've seen is per-arch/os .hmakerc files, which seems fine as it is to me. relevant links: http://www.pathname.com/fhs/pub/fhs-2.3.html#THEUSRHIERARCHY http://www.pathname.com/fhs/pub/fhs-2.3.html#USRLIBLIBRARIESFORPROGRAMMINGA… http://www.pathname.com/fhs/pub/fhs-2.3.html#VARLIBVARIABLESTATEINFORMATION Thanks Ian

3 2

nhc98-1.16 for OpenBSD
by dons＠cse.unsw.edu.au 17 May '04

17 May '04

nhc98-1.16 has been imported into the OpenBSD ports tree. It is known to build on i386, sparc and powerpc OpenBSD. I'm attempting to build on the other 32 bit platforms (arm, hppa, m68k, m88k, vax), and we'll see how that goes. It is built via GHC on i386 and sparc, and via gcc everywhere else. I've also stripped down the ghc regress tests so that we could auto-regress test nhc as it gets built. This consisted in removing as many ghc language extension tests to their testsuite as I could, until around 550 tests were being built. This may be of interest to other nhc users. An ideal solution would be to add a 'no-nhc' flag to each test in the testsuite, so that we can have a unified testsuite tree... The port script that controls this is at: http://www.openbsd.org/cgi-bin/cvsweb/ports/lang/nhc98/Makefile?rev=1.1.1.1 nhc won't build on at least amd64, so I've disabled 64 bit platforms (this is a known issue?) Transcripts (via gcc and ghc) of the amd64 build are attached. It seems gcc-3.3.2 doesn't play well with nhc, at the moment, as this failure looks similar to a gcc3 failure I've had on linux/i386 too. Cheers, Don

2 2

by 李明 10 Mar '04

10 Mar '04

1 0

Re: [nhc-users] failure to compile on mips-sgi6.5
by Malcolm Wallace 18 Dec '03

18 Dec '03

"William S." <wstan(a)xs4all.nl> writes: > OK, got past the first problem. Have this now: > > gcc -S -DHIGH_BYTE_FIRST -O3 -I/usr/people/wstan/nhc98-1.16/include > -DVERSION="\"1.16\"" -o - newtables.c | sed '/.align 32/s/32/4/' |\ > gcc -c -x assembler -DHIGH_BYTE_FIRST -O3 > -I/usr/people/wstan/nhc98-1.16/include -DVERSION="\"1.16\"" -o > /usr/people/wstan/nhc98-1.16/targets/mips-IRIX6/obj/runtime/Kernel > /newtables.o - > as ERROR: -E or specified language required when input is from standard input Strange. The second call to gcc does indeed get its input from stdin, but it also explicitly specifies the language with '-x assembler'. Because of the mention of the -E option (indicating the C pre-processor), it is possible that on mips-IRIX6 you actually require the language to be '-x assembler-with-cpp'. You can change this option in src/runtime/Kernel/Makefile, in the rule for ${SOBJS}. > Easy fix or should I try the Cygwin branch? (not sure how to select > for that). No, Cygwin's quirks are not relevant here. Regards, Malcolm

2 15