- Nhc-users - Haskell.org

Google Summer of Code 2006
by Paolo Martini 24 Apr '06

24 Apr '06

Hello, Haskell.org is going to partecipate as mentoring organization to the Google Summer of Code programme of this year. We have formed a group of volunteers taking care of the administrative work as well as the mentoring part. The group is listed here: <http://haskell.org/haskellwiki/Summer_of_Code:_People>. We are now looking for projects, and mentors. We've reached number 24 but we are happily going to add more, Haskell Open Source paid work for the summer! If you run some active open source program you might want to mentor some student who will be working on your code, for example. Either just add your ideas here: <http://haskell.org/haskellwiki/Summer_of_Code:_Project_suggestions> (as well as your informations on the People's page) or contact me for details. I am also looking for students, I'd like to ear from them if they are going to partecipate with Haskell projects, last year I was alone! (I did write the bindings to the Cairo vector graphics library and integrated them in Gtk2Hs.) Details about the programme itself can be found on the official webpage: http://code.google.com/summerofcode.html Hoping you like the idea as much as me, Paolo. --

1 0

Re: [nhc-users] Help Needed
by arunk＠it.iitb.ac.in 13 Apr '06

13 Apr '06

FP course of Prof Sanyal has started eh :) ? nhc adheres to Haskell-98 standards (Malcolm can correct me if I'm wrong), so most probably some linking/library issues. BTW it'll help if you can give us the exact error you got. Was it unable to compile ? or did you get some error while running the executable. - ASJ

1 0

Help Needed
by BHASKAR G. REDDY 13 Apr '06

13 Apr '06

Hi, Me, pretty much new to using NHC-98. Till now i have used HUGS only. Is the syntax of input haskell programs same? Its not compiling when i am giving the same programs which are successfully got exeuted with HUGS. What could be the probelm? If thr is any subtle difference in syntax, then plz let me know good link whr can i get enough info? Thx in adavance, -- Bhaskar Reddy.G, Graduate Student, Department of CSE, IIT Bombay

1 0

Problems in nhc98 compilation
by Monique Louise 21 Nov '05

21 Nov '05

Hi all, I receive the following message when I try to compile and install nhc98 at WinXP: timeUnix.c:7:23: sys/times.h: No such file or directory timeUnix.c: In function `timerRead': timeUnix.c:19: storage size of `buffer' isn't known timeUnix.c:20: `clock_t' undeclared (first use in this function) timeUnix.c:20: (Each undeclared identifier is reported only once timeUnix.c:20: for each function it appears in.) make[2]: ** [c:/nhc98-1.18/targets/ix86-CYGWIN_NT-5.1 /obj/runtime/Kernel/timeUni x.o] Erro 1 make[2]: Leaving directory `/cygdrive/c/nhc98-1.18/src/runtime/Kernel' make[1]: ** [ofiles] Erro 2 make[1]: Leaving directory `/cygdrive/c/nhc98-1.18/src/runtime' make: ** [targets/ix86-CYGWIN_NT-5.1/runtime] Erro 2 The error above appears when I type "make". Has anyone any idea about what's happening? Thanks in advance, Monique Monteiro

2 3

Newbie question
by Monique Louise 31 Aug '05

31 Aug '05

Hi all, I'm a newbie to nhc and I'm trying to compile a simple "hello world" program: module Main where main = putStr "hello world" with nhc98 hello.hs or nhc98 hello.hs -o hello.exe but the following messages appera: C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/main.o(.text+0xa2):main.c: unde fined reference to `setjmp' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/main.o(.text+0xdd):main.c: unde fined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/main.o(.text+0x12a):main.c: und efined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/mutator.o(.text+0x2d2):mutator. c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/mutator.o(.text+0x425):mutator. c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/mutator.o(.text+0x44d):mutator. c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/mutator.o(.text+0x2ac9):mutator .c: more undefined references to `_impure_ptr' follow C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(haskellInit.o)(.text+ 0x94):haskellInit.c: undefined reference to `_imp___ctype_' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(haskellInit.o)(.text+ 0x180):haskellInit.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(haskellInit.o)(.text+ 0x504):haskellInit.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(haskellInit.o)(.text+ 0x757):haskellInit.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(haskellInit.o)(.text+ 0x771):haskellInit.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(haskellInit.o)(.text+ 0x7ca):haskellInit.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(haskellInit.o)(.text+ 0x93f):haskellInit.c: more undefined references to `_impure_ptr' follow C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(cIOExtras.o)(.text+0x d7):cIOExtras.c: undefined reference to `__errno' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(mark.o)(.text+0x498): mark.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(mark.o)(.text+0x568): mark.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(mark.o)(.text+0x6f8): mark.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(mark.o)(.text+0x90a): mark.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(timeUnix.o)(.text+0x4 3):timeUnix.c: undefined reference to `times' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(timeUnix.o)(.text+0x6 5):timeUnix.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(timeUnix.o)(.text+0xb 3):timeUnix.c: undefined reference to `times' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(timeUnix.o)(.text+0xd 2):timeUnix.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(timeUnix.o)(.text+0x1 13):timeUnix.c: undefined reference to `times' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(timeUnix.o)(.text+0x1 32):timeUnix.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(stableptr.o)(.text+0x 199):stableptr.c: undefined reference to `_impure_ptr' C:/cygwin/usr/local/lib/nhc98/ix86-CYGWIN_NT-5.1/Runtime.a(stableptr.o)(.text+0x 33a):stableptr.c: undefined reference to `_impure_ptr' I'm running version 1.16 at Windows XP with Cygwin. Does anyone have any idea about what is happening? Thanks in advance, Monique

2 1

Maximum Heap Size
by Arunkumar S Jadhav 09 Jun '05

09 Jun '05

Hello All, Am I right in assuming that nhc98 uses only 29 (out of possible 32) bits for addressing. Because, the top bit is used for ZAP, and least two bits are used to differentiate between various data nodes and so on. If this is the case then, we can't allocate more than 2^29 bytes or 2^27 (i.e 128 Mega Words) words for the total heap size (i.e Sp + Hp). Even if we were to use this maximum limit, the addressing has to start from 0x0, because otherwise we'll need more than 29 bits. So the actual limit is much less than 128 M Words. Or am I missing something ? Regards, ------------------------------------ Arunkumar S Jadhav, Masters Student, KReSIT, IIT-Bombay, India Ph: +91-22-25764967 http://www.it.iitb.ac.in/~arunk ------------------------------------ I exist because I work.

2 1

No default for MutualRecursion.PP
by Christian Maeder 07 Jun '05

07 Jun '05

Hi, The program below is accepted by "ghc -fno-monomorphism-restriction" and "hugs +98". Why can't nhc98 (v1.18) handle it? Tossing in something like: instance PP () default (()) looks a bit stupid in order to become "portable", doesn't it? If I replace PP by Show (and pp by show and omit the class declaration) nhc98 has no problem. Also if I replace one occurrence of fA or fB directly with pp it goes through. Cheers Christian > hmake -nhc98 MutualRekursion.hs nhc98 -c -o MutualRekursion.o MutualRekursion.hs ====== Errors after type inference/checking: No default for MutualRecursion.PP at 23:1-24:18.(174,[(148,187)]) No default for MutualRecursion.PP at 19:1-20:18.(183,[(148,186)]) -- -------------------------------------------------------------------- module MutualRecursion where data A a = A a | BasA (B a) data B a = B a | AasB (A a) class PP a where pp :: a -> String instance PP a => PP (A a) where pp (A a) = pp a pp (BasA b) = pp b instance PP a => PP (B a) where pp (B a) = pp a pp (AasB b) = pp b fA :: PP a => A a -> String fA (A a) = pp a fA (BasA b) = fB b fB :: PP a => B a -> String fB (B b) = pp b fB (AasB a) = fA a

1 1

Debugger related
by Arunkumar S Jadhav 11 May '05

11 May '05

Hello All, I know that this is unrelated mailing list, but still, since it had to do with mutator.c wanted to get a doubt cleared here. Suppose I compile mutator.c (or any other C file for that matter) with debugging option on (-ggdb3) and get an object file as output. Now I want to view the assembly file produced corresponding to it, but I also want to see the correspondence between C statements and i386 instructions. So I run "objdump" command with appropriate switches on and I get the desired file. Now my doubt: Will the debugger option (-ggdb3) cause the compiler to emit any i386 instructions that wouldn't have been emitted otherwise. In otherwords, does gcc compiler "insert" extra machine (i386 in my case) instructions for the purpose of debugging ? As far as I understand the compiler design, at the level of machine code, it shouldn't make any difference whether I compile a file with debugger option on. Please share your opinion / knowledge regarding this. Regards, ------------------------------------ Arunkumar S Jadhav, Masters Student, KReSIT, IIT-Bombay, India Ph: +91-22-25764967 http://www.it.iitb.ac.in/~arunk ------------------------------------ I exist because I work.

2 1

Space Profiling
by Benjamin Fraser 10 May '05

10 May '05

Hello, I wanted to use nhc to profile a simple program but ran into a string of errors. The example I present here is one of four small programs I have played with. This example is from nhc98-1.18 compilied from source, I have played with the 1.18 and 1.16 binary versions also. I'm using debian 3. The program is: #####################strict.hs module Main where foldl' f e [] = e foldl' f e (x:xs) = (foldl' f $! f e x) xs main = print $ foldl' (+) 0 [1..1000] {-main = print $ foldr (+) 0 [1..1000]-} #####################strict.hs With the aim of showing strict.hs would run in constant space as compared to the commented out version. I compilied and ran the program thus (I know I can use hmake instead): $ nhc98 -p -o strict strict.hs $ ./strict +RTS -p -i25kb -RTS 500500 $ hp2graph strict.hp Illegal token JOB at 69 Aborted This is the most common error encountered. Try again: $ ./strict +RTS -p -i24kb -RTS 500500 $ hp2graph strict.hp Segmentation fault Segmentation faults are the second most common error. It is also possible for hp2graph to run without error but it produces invalid postscript that cannot be viewed. Here are the strict.hp files from each run: #####################strict 25kb interval JOB ./strict +RTS -p -i25kb -RTS ; DATE "Tue May 10 21:45:49 2005 " PROFILE_PRODUCER MARK 0.00 SAMPLE 0 0.00 344 <APPLY> 72 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 1 0.00 356 <binary_int_op> 8 <APPLY> 80 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 2 0.00 328 <binary_int_op> 8 <APPLY> 72 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 3 0.00 344 <binary_int_op> 8 <APPLY> 72 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 4 0.00 356 <binary_int_op> 8 <APPLY> 80 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 5 0.00 328 <binary_int_op> 8 <APPLY> 72 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 6 0.00 344 <binary_int_op> 8 <APPLY> 72 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 7 0.00 ; #####################strict 25kb interval #####################strict 24kb interval JOB ./strict +RTS -p -i24kb -RTS ; DATE "Tue May 10 21:46:40 2005 " PROFILE_PRODUCER MARK 0.00 SAMPLE 0 0.00 328 <APPLY> 72 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 1 0.00 328 <APPLY> 72 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 2 0.00 328 <binary_int_op> 8 <APPLY> 72 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 3 0.00 328 <binary_int_op> 8 <APPLY> 72 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 4 0.00 328 <binary_int_op> 8 <APPLY> 72 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 5 0.00 328 <binary_int_op> 8 <APPLY> 72 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 6 0.00 328 <binary_int_op> 8 <APPLY> 72 <Main> 8 Builtin.primIntFromInteger 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 7 0.00 544 IO.hPutChar:7:30-7:52 12 <APPLY> 56 <Main> 8 Prelude.primIntegerAdd 8 ; MARK 0.00 SAMPLE 8 0.00 ; #####################strict 24kb interval What am I doing wrong, some option perhaps? Thanks, Ben Fraser

2 1

Updates for Mac OS X users using Darwinports
by Gregory Wright 17 Mar '05

17 Mar '05

Hi, Users of Mac OS X/Darwinports will find new portfiles for ghc (ghc-6.4), nhc98 (nhc-1.18) and hugs98 (March 2005 release). The hmake port has been bumped to version 3.10 and a new port of cpphs (version 0.9) has been added. The darwinports system builds from source, although progress is being made on extending it to provide binary packages as well. A new ghc-devel port will be available soon for those who want to live on the bleeding edge. Thanks to Malcolm Wallace for ironing out some last minute issues with nhc98. Best Wishes, Greg Gregory Wright

1 0