Re: Constant space infinite itteration ... solution?
[EMAIL PROTECTED] wrote: Also ... I've been using the graphics libs with HUGS, but I can't find the equivalent in GHC ... what is the recomended library for writing GUIs in GHC Haskell? And where do I get it? My current favourite way to make GUI's is to use the GUI painter Glade and have it generate C code with calls to the Gtk GUI library, after that the only thing you have to implement are the callbacks. One big advantage Gtk offers over other GUI libraries is that it clearly separates the GUI construction code from the callback code, this makes it very suitable for GUI painters, i.e. the code the GUI painter generates doesn't have to be changed manually. Another advantage of Gtk, from the Haskell viewpoint, is that it was designed to work with other languages. Quoted from the Gtk page: GTK+ has been designed from the ground up to support a range of languages, not only C/C++. Using GTK+ from languages such as Perl and Python (especially in combination with the Glade GUI builder) provides an effective method of rapid application development. Although I haven't looked at the Haskell side in a while, everything to go from Glade to a Haskell GUI should all be there: http://www.gtk.org/ http://glade.gnome.org/ http://www.cse.unsw.edu.au/~chak/haskell/gtk/ The most recent version of Gtk+HS has support for libglade, this means you can directly read the Glade XML file that describes the GUI construction and the names of the callback functions. To be honest I don't know how the callbacks are actually connected to Haskell code, but there is probably an example in the Gtk+HS distribution to show how it works. Jan ___ Haskell mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/haskell
Re: Floats and Doubles
Juan Ignacio Garcia Garcia wrote: *P2 (fromRational ((toRational 4) - ( toRational 5.2 ))) -1.2002 I can't explain this one, how would fromRational know that it has to create a Double ? Jan ___ Glasgow-haskell-users mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
Re: Using the Parsec library
Jose Romildo Malaquias wrote: Any sugestion on how to implement the 'not starting with A: or B:' problem? Hi, I'm not familiar with Parsec, but in ParseLib it's fairly straightforward, first you add a new combinator: ifnot :: Parser a - Parser b - Parser b ifnot (P p1) (P p2) = P ( \s - if parseOk (p1 s) then [] else p2 s ) where parseOk :: [(a,String)] - Bool parseOk [x] = True parseOk _ = False then you can say: ifnot parse_a_or_b parse_foo You will need the sources of ParseLib to try it out though, because P isn't exported. Jan ___ Haskell mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/haskell
Re: Beginner's question: Memo functions in Haskell
Janis Voigtlaender wrote: It would also seem that one needs to write fast = memo slow instead, because otherwise a new memo-version of slow might be created for every call with some n (subject to let-floating?). However, the version: module Fib where import Memo slow 0 = 0 slow 1 = 1 slow n = fast (n-1) + fast (n-2) fast = memo slow is not particularly fast easier. Quite in contrary... strange. Janis. It works under hugs98, but not under ghc5.02.2, I CC'd the ghc bugs list. Actually, I was also expecting fast n = memo slow n to work ? Jan ___ Haskell mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/haskell
Re: sharing datatypes : best practice ?
Taesch, Luc wrote: do u isolate just the datatype, or a few related with, in a very small file (header like, i would say) or some basic accessor function with it ? isnt it leading to massiv quantities of small files ? Asuming you have some typed AST with many mutually recursive datatypes, I would keep them in one big file. This should be fine if the datatypes are simple (no deriving Read and Show etc.). For an AST you don't want accessor functions: the datatypes are the interface. For some datatypes you want to hide the datatype and provide a function based interface, this should be in the same file as the datatype. Usually there is also some kind of asumed hierarchy in datatypes, e.g. Int List FiniteMap, to determine where functions operating on multiple datatypes should be placed, but that's the same in OO. Jan ___ Haskell-Cafe mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: failed when trying to compile FranTk with ghc
luc wrote: I tried FranTk with ghc 4.08 and got : (this is the "fixed" FrankTk, alledged working with 4.06, if im not wrong) are there any difference with 4.06 and 4.08 ? Below is a list of fixes to get FranTk working with ghc4.08. Jan Run configure like normal, i.e.: ./configure --prefix=/scratch/kort/pkgs/frantk Change the file FranTk/src/FranSrc/Compatibility.ghc.hs to -- this module is for GHC; GSL module Compatibility ( double2Float , yield , debugMsgLn , setDebug , mkWeakIORef , toInt , fromInt ) where import GlaExts import NumExts(doubleToFloat) import PrelRead(readDec) import IOExts import Concurrent import Exception double2Float = doubleToFloat setDebug :: Bool - IO () setDebug _ = return () debugMsgLn :: String - IO () debugMsgLn s = assert (trace s `seq` True) $ return () In the file FranTk/src/FranTkSrc/Makefile change the line: HC_OPTS+= -i../FranSrc:../TclHaskellSrc -fallow-overlapping-instances -fallow-undecidable-instances -syslib misc to: HC_OPTS+= -i../FranSrc:../TclHaskellSrc -fallow-overlapping-instances -fallow-undecidable-instances -syslib data In the file FranTk/demos/Makefile change the line: HC_OPTS+= -i../src/FranTkSrc:../src/FranSrc:../src/TclHaskellSrc -fallow-overlapping-instances -fallow-undecidable-instances -syslib misc to: HC_OPTS+= -i../src/FranTkSrc:../src/FranSrc:../src/TclHaskellSrc -fallow-overlapping-instances -fallow-undecidable-instances -syslib data If the compiler complains about not finding the file "tcl.h", in the file FranTk/src/TclHaskellSrc/Makefile change the rule: tclhaskell.o: tclhaskell.c tclhaskell.h $(HC) -c tclhaskell.c to: tclhaskell.o: tclhaskell.c tclhaskell.h $(HC) -I/opt/arch/lib/tk8.3/include -c tclhaskell.c where /opt/arch... is the place "tcl.h" and "tk.h" are. In the directory FranTk/src/TclHaskellSrc do: rm *.hi *.o In the file FranTk/src/TclHaskellSrc/Makefile add TclPrim.hs to the HS_SRCS list. In the directory FranTk/src/FranSrc do: rm *.hi In the directory FranTk/demos do: rm *.hi And then continue the normal installation with: gmake boot gmake all ___ Glasgow-haskell-users mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
Re: Bagley shootout. Was: Lightningspeed haskell
I implemented the programs for hash1 and hash2, but I had to
make a lot of changes to FiniteMap to get them to work:
- I changed foldl to foldl' (as defined in Hugs).
- I made foldFM strict (like foldl').
- I made the datatype FiniteMap strict (put !'s everywhere).
I put the programs below, maybe there is a better way to do this.
Jan
-- Hash1
module Main(main) where
import FiniteMap
import System
writeHex' :: Int - String
writeHex' 0 = ""
writeHex' n
= c:writeHex' y
where
x = n `mod` 16
y = n `div` 16
c | x = 9= toEnum ((fromEnum '0') + x)
| otherwise = toEnum ((fromEnum 'A') + x - 10)
writeHex :: Int - String
writeHex n
| n 0 = reverse (writeHex' n)
| n == 0= "0"
| otherwise = '-':(reverse (writeHex' (-n)))
main :: IO()
main = do {
[arg] - getArgs;
let
n = read arg
fm = listToFM [(writeHex i,i) | i - [1..n]]
c = sum [if elemFM (show i) fm then 1 else 0 | i - reverse [1..n]]
in
putStrLn (show c);
-- Hash2
module Main(main) where
import FiniteMap
import System
import Maybe
foldl' :: (a - b - a) - a - [b] - a
foldl' _ a [] = a
foldl' f a (x:xs) = (foldl' f $! f a x) xs
f :: FiniteMap Int Int - FiniteMap Int Int - Int - FiniteMap Int Int
f hash1 hash2 k
| isJust elt2 = addToFM hash2 k (fromJust elt1 + fromJust elt2)
| otherwise = addToFM hash2 k (fromJust elt1)
where
elt1 = lookupFM hash1 k
elt2 = lookupFM hash2 k
main :: IO()
main = do {
[arg] - getArgs;
let
n = read arg
hash1 = listToFM [(i,i) | i - [1..1]]
keys= concat (take n (repeat (keysFM hash1)))
hash2 = foldl' (f hash1) emptyFM keys;
h1_1= fromJust (lookupFM hash1 1)
h1_ = fromJust (lookupFM hash1 )
h2_1= fromJust (lookupFM hash2 1)
h2_ = fromJust (lookupFM hash2 )
in
putStrLn (show h1_1 ++ " " ++ show h1_ ++ " " ++
show h2_1 ++ " " ++ show h2_);
}
___
Glasgow-haskell-users mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
Re: Bagley shootout. Was: Lightningspeed haskell
Simon Peyton-Jones wrote: A String is a [Char] and a Char is a heap object. So a file represented as a string takes a massive 20 bytes/char (12 for the cons cell, 8 for the Char cell). Then it's all sucked through several functions. It's entirely possible, though, that the biggest performance hit is in the I/O itself. We'd be happy if anyone wanted to invesigate and improve. Unless ghc is extremely fast at filling a heap, it's the memory allocation. I get 11.8 seconds for ghc with a standard heap and 7.3 seconds when I give it enough heap not to do garbage collection. Since this is 200M, I don't think there is much time to do anything else. The input is 200 bytes. So, this would be 40M worth of [Char] data, I guess lines and unlines make ehm 12*2*2=48M, so that's about 100M total. I guess the other 100M is used for function applications. Jan ___ Glasgow-haskell-users mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
Re: framework for composing monads?
Andy Gill's Monad Template Library is good for that, but the link from the Haskell library page is broken: http://www.cse.ogi.edu/~andy/monads/doc.htm Jan ___ Haskell-Cafe mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/haskell-cafe
Re: Floating point performance
"Julian Seward (Intl Vendor)" wrote: I tried this, with ghc-4.08.1 -O both with and without profiling, on a Sparc box of I believe around 300 MHz, and I can't reproduce it at all. Without profiling, it allocates about 505 k of heap and runs in 0.02 seconds. Ummm ? J I didn't use any optimizations, but I am sure that passing -O to ghc will make it see that 1*1*... is a constant expression. I guess turning all optimizations off is not the right way to test stuff either. Jan ___ Glasgow-haskell-users mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
Floating point performance
Hi, I noticed ghc (version 4.08.1) floating point performance is really slow on my computer: a 270Mhz sun ultra5. The program below does 1 milion floating point multiplications and takes 2 seconds to run. I made a profile and it says most of the time (93%) is spent in the function bar. Any idea what is going on ? Regards, Jan module Main where foo :: Double - [Double] foo 0 = [1] foo n = bar n:foo (n-1) bar :: Double - Double bar n = n*1*1*1*1*1*1*1*1*1*1 *1*1*1*1*1*1*1*1*1*1 *1*1*1*1*1*1*1*1*1*1 *1*1*1*1*1*1*1*1*1*1 *1*1*1*1*1*1*1*1*1*1 *1*1*1*1*1*1*1*1*1*1 *1*1*1*1*1*1*1*1*1*1 *1*1*1*1*1*1*1*1*1*1 *1*1*1*1*1*1*1*1*1*1 *1*1*1*1*1*1*1*1*1*1 foldl' :: (a - b - a) - a - [b] - a foldl' _ a [] = a foldl' f a (x:xs) = (foldl' f $! f a x) xs main :: IO() main = putStrLn (show (foldl' (+) 0 (foo 1))) --- Profile: total time =2.60 secs (130 ticks @ 20 ms) total alloc = 24,845,828 bytes (excludes profiling overheads) COST CENTRE MODULE %time %alloc bar Main93.1 96.1 foo Main 3.12.4 foldl' Main 3.11.4 individual inherited COST CENTRE MODULE entries %time %alloc %time %alloc MAIN MAIN 00.0 0.0100.0 100.0 mainMain 00.0 0.0 0.0 0.0 CAF PrelShow 10.0 0.0 0.0 0.0 CAF PrelFloat30.0 0.0 0.0 0.0 CAF PrelHandle 30.0 0.0 0.0 0.0 CAF Main 90.0 0.0100.0 100.0 main Main 10.8 0.0100.0 100.0 foo Main 100013.1 2.4 96.2 98.5 bar Main 1 93.1 96.1 93.1 96.1 foldl'Main 100023.1 1.4 3.1 1.4 ___ Glasgow-haskell-users mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
Re: C stack
Simon Marlow wrote: I don't forsee any problems with the C stack - the stack pointer stays static during execution of Haskell code, and only moves when we do a C call or return to the RTS. But how do I get the address of the bottom of the stack ? It seems I can define a C main so I can get the bottom of the stack, but how do I make it call Haskell's main to start the application ? Don't forget to tell the GC to traverse the Haskell heap too - GHC keeps this in mmapped address space. See fptools/ghc/rts/MBlock.c for the details. Ok, I will do that, it looks like a better idea than working with ForeignObj and finalizers. Jan ___ Glasgow-haskell-users mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
C stack
Hi,
I'm trying to write an interface for a C library that
uses a Boehm type garbage collector. So, I need to get
the address of the bottom of the C stack. In a C
application this would look something like:
main()
{
int bottomOfStack;
}
Where "bottomOfStack" would be the thing I need. So,
I'd want some function like:
bottomOfStack :: IO Addr
Would it be possible at all to make an interface to
a C library that uses Boehm's garbage collection ?
What Boehm's algorithm does is traverse the C stack
at each garbage collect marking all accessible
nodes and then sweeping all unmarked nodes. Would
this cause problems with the C code generated by
ghc ?
Jan
___
Glasgow-haskell-users mailing list
[EMAIL PROTECTED]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
Re: openFile, process file table full
Simon Marlow wrote:
Arguably, readFile is too strict: why is the file opened, if there
aren't any characters needed. What is the motivation for opening the
file eagerly?
Once we've started reading lazilly, it's hard to generate errors. The file
is opened eagerly so that any errors generated (such as if the file doesn't
exist) can be propogated immediately to the caller.
I checked the Haskell Report, which doesn't seem to specify the correct
behaviour - I guess this should be clarified. The report also contains the
sentence "The exceptions raised by the I/O functions in the Prelude are
defined in the Library Report", but I can't find any reference to
readFile/writeFile in the library report.
Cheers,
Simon
I can't seem to work around this behaviour, is it possible to
force readFile to just open a file, read it entirely and close it
before doing the next open ? I tried a few things with seq,
but this doesn't work:
xs - mapM (\x - let y = readFile x
in seq y y) (take 1000 (repeat "tmp"));
and neither does this:
let
f = mapM (\x - let y = readFile x
in seq y y) (take 1000 (repeat "tmp"));
in do {
xs - seq f f;
return ();
}
Jan
openFile, process file table full
Hi,
The following program:
module Main(main) where
main:: IO()
main = do {
xs - mapM readFile (take 1000 (repeat "tmp"));
return ();
}
Results in the error:
Fail: resource exhausted
Action: openFile
Reason: process file table full tmp
This doesn't work either:
myreadFile :: String - IO String
myreadFile filename = do {
f - openFile filename ReadMode;
x - hGetContents f;
hFlush f;
hClose f;
return x;
}
main:: IO()
main = do {
xs - mapM (\x - let y = myreadFile x in seq y y)
(take 1000 (repeat "tmp"));
return ();
}
Regards,
Jan
P.S. My system:
(~ 270) uname -sr
SunOS 5.6
(~ 271) dmesg | grep SUNW
cpu0: SUNW,UltraSPARC-IIi (upaid 0 impl 0x12 ver 0x12 clock 270 MHz)
SUNW,m64B0 is /pci@1f,0/pci@1,1/SUNW,m64B@2
stdout is /pci@1f,0/pci@1,1/SUNW,m64B@2 major 35 minor 0
SUNW,hme0: CheerIO 2.0 (Rev Id = c1) Found
SUNW,hme0 is /pci@1f,0/pci@1,1/network@1,1
SUNW,hme0: Using Internal Transceiver
SUNW,hme0: 100 Mbps full-duplex Link Up
(~ 272) cat /etc/release
Solaris 2.6 5/98 s297s_hw3smccDesktop_09 SPARC
Copyright 1998 Sun Microsystems, Inc. All Rights Reserved.
Assembled on 24 April 1998
(~ 273) ghc --version
The Glorious Glasgow Haskell Compilation System, version 4.08
(~ 274) gcc --version
2.8.1
(~ 275) which make
make:aliased to gmake
(~ 276) /home/kort/pkgs/make/bin/make --version
GNU Make version 3.78.1, by Richard Stallman and Roland McGrath.
Built for sparc-sun-solaris2.6
Copyright (C) 1988, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99
Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.
Report bugs to [EMAIL PROTECTED].
(~ 277) man X | grep 'X Version' | grep ' 1$'
Reformatting page. Wait... done
X Version 11 Last change: Release 6 1
compiled program loops
I compiled a program using ghc version 4.06 and it goes into a loop, I
tried
figuring out what goes on with "trace", but it looks as if this is a ghc
bug. Wether it loops or not depends on the length of the list: if you
remove
1 "Int 1" line in the code below it will finish immediatly.
I need a fix urgently, I can't find any way to work around this problem
in
the example program, let alone in the real application.
In the real application I also get the error:
bingraph: fatal error: evacuate: THUNK_SELECTOR: strange selectee 27
which I'm hoping is related, because I have no idea what it means and
I can't reproduce this in an example program. Any idea where I should
be looking for the cause of this ?
Jan
My system:
(~/src/haskell/bgtmp 54) uname -sr
SunOS 5.6
(~/src/haskell/bgtmp 55) dmesg | grep SUNW
cpu0: SUNW,UltraSPARC-IIi (upaid 0 impl 0x12 ver 0x91 clock 360 MHz)
SUNW,m64B0 is /pci@1f,0/pci@1,1/SUNW,m64B@2
stdout is /pci@1f,0/pci@1,1/SUNW,m64B@2 major 35 minor 0
SUNW,hme0: CheerIO 2.0 (Rev Id = c1) Found
SUNW,hme0 is /pci@1f,0/pci@1,1/network@1,1
SUNW,hme0: Using Internal Transceiver
SUNW,hme0: 100 Mbps full-duplex Link Up
(~/src/haskell/bgtmp 56) cat /etc/release
Solaris 2.6 5/98 s297s_hw3smccDesktop_09 SPARC
Copyright 1998 Sun Microsystems, Inc. All Rights Reserved.
Assembled on 24 April 1998
(~/src/haskell/bgtmp 57) ghc --version
The Glorious Glasgow Haskell Compilation System, version 4.06
(~/src/haskell/bgtmp 58) gcc --version
egcs-2.91.66
The program:
module Main where
import GlaExts
data ATerm = AList [ATerm]
| AInt Integer
-- deriving (Show,Read)
toNode (AList xs) = sum (map toNode xs)
toNode (AInt x) = x
-- Main.
--
main :: IO()
main = do {
let
aterm = trace "test1" (trace "test22" (AList (trace "test23" [
trace "test24" (AInt 1),
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt 1,
AInt
Re: records in Haskell
Simon Marlow wrote:
Jan Kort writes:
It seem that any record, no matter how trivial, can't be much
longer than about 200 lines in Haskell. If a try to compile a
300 line record containing just:
data X = X {
f1 :: String,
f2 :: String,
f3 :: String,
...
f300 :: String
}
It needs about 90M heap in ghc4.06. Whereas a 150 line record
requires less than 6M heap. After this big gap it levels off
to a somewhat more decent exponential increase: a 450 line
record requires about 180M heap.
I could file a bug report, but it seems that all compilers
(ghc4.06, nhc98, hbc0.9994 and hugs) have this problem. So,
is this a fundamental problem ?
Actually, the 150-line record needs about 20M, and the 300-line record needs
about 75M. These figures are roughly double the actual residency, because
GHC's underlying collector is a copying, not compacting, one.
GHC automatically increases the heap size up to a maximum of 64M unless you
tell it not to (with -optCrts-M32m, for example). I'll bet this is the
source of the confusion.
The heap requirement is still non-linear, but I'm guessing that this is
because for each line you add to the record the compiler has to not only
generate a new selector function, but also add a field to the record being
pattern matched against in all the existing selectors.
Cheers,
Simon
Thanks for the answers and sorry for the late reaction.
I worked out an example to understand what you wrote.
GHC will probably generate something like this:
data R = R String Integer
deriving (Read,Show)
selectA (R s _) = s
selectB (R _ i) = i
updateA (R _ b) a = (R a b)
updateB (R a _) b = (R a b)
emptyR = R undefined undefined
Which you can then use like this:
updateR = updateB (updateA emptyR "a") 2
testA = selectA updateR
testB = selectB updateR
I agree that the select and update pattern matchings would
get big for a 300 line record, but 75M is a lot of memory.
Especialy because the pattern matches and the right hand
sides of both the selects and updates are trivial pieces
of code: no nesting, no currying etc. But maybe GHC
generates something extra ? Is special code generated
for updating multiple fields for example ?
I can probably work around this in a simple way: since I'm
generating the big record, I might as well generate
the selects, updates and emptyR instead and split them
over a couple of files.
Jan
deriving Read and Show
Hi, When I use deriving Read and Show on many datatypes, both ghc and hugs don't work. I've attached an example, to get the error, do: hugs Main.hs or: ghc Main.hs Hugs quits with the error message: ERROR "Main.hs": Type constructor storage space exhausted And ghc quits with: GHC's heap exhausted; while trying to allocate 0 bytes in a 67108864-byte heap; use the `-Hsize' option to increase the total heap size. It's not urgent though, I can work around it. I use Hugs98 and my setup for ghc is: (~/project/genast 76) uname -sr SunOS 5.6 (~/project/genast 77) dmesg | grep SUNW cpu0: SUNW,UltraSPARC-IIi (upaid 0 impl 0x12 ver 0x12 clock 270 MHz) SUNW,m64B0 is /pci@1f,0/pci@1,1/SUNW,m64B@2 stdout is /pci@1f,0/pci@1,1/SUNW,m64B@2 major 35 minor 0 SUNW,hme0: CheerIO 2.0 (Rev Id = c1) Found SUNW,hme0 is /pci@1f,0/pci@1,1/network@1,1 SUNW,hme0: Using Internal Transceiver SUNW,hme0: 100 Mbps full-duplex Link Up (~/project/genast 78) cat /etc/release Solaris 2.6 5/98 s297s_hw3smccDesktop_09 SPARC Copyright 1998 Sun Microsystems, Inc. All Rights Reserved. Assembled on 24 April 1998 (~/project/genast 79) ghc --version The Glorious Glasgow Haskell Compilation System, version 4.06 (~/project/genast 80) gcc --version egcs-2.91.66 (~/project/genast 81) Jan module Main(main) where data X1 = X1 String deriving (Read,Show) data X2 = X2 String deriving (Read,Show) data X3 = X3 String deriving (Read,Show) data X4 = X4 String deriving (Read,Show) data X5 = X5 String deriving (Read,Show) data X6 = X6 String deriving (Read,Show) data X7 = X7 String deriving (Read,Show) data X8 = X8 String deriving (Read,Show) data X9 = X9 String deriving (Read,Show) data X10 = X10 String deriving (Read,Show) data X11 = X11 String deriving (Read,Show) data X12 = X12 String deriving (Read,Show) data X13 = X13 String deriving (Read,Show) data X14 = X14 String deriving (Read,Show) data X15 = X15 String deriving (Read,Show) data X16 = X16 String deriving (Read,Show) data X17 = X17 String deriving (Read,Show) data X18 = X18 String deriving (Read,Show) data X19 = X19 String deriving (Read,Show) data X20 = X20 String deriving (Read,Show) data X21 = X21 String deriving (Read,Show) data X22 = X22 String deriving (Read,Show) data X23 = X23 String deriving (Read,Show) data X24 = X24 String deriving (Read,Show) data X25 = X25 String deriving (Read,Show) data X26 = X26 String deriving (Read,Show) data X27 = X27 String deriving (Read,Show) data X28 = X28 String deriving (Read,Show) data X29 = X29 String deriving (Read,Show) data X30 = X30 String deriving (Read,Show) data X31 = X31 String deriving (Read,Show) data X32 = X32 String deriving (Read,Show) data X33 = X33 String deriving (Read,Show) data X34 = X34 String deriving (Read,Show) data X35 = X35 String deriving (Read,Show) data X36 = X36 String deriving (Read,Show) data X37 = X37 String deriving (Read,Show) data X38 = X38 String deriving (Read,Show) data X39 = X39 String deriving (Read,Show) data X40 = X40 String deriving (Read,Show) data X41 = X41 String deriving (Read,Show) data X42 = X42 String deriving (Read,Show) data X43 = X43 String deriving (Read,Show) data X44 = X44 String deriving (Read,Show) data X45 = X45 String deriving (Read,Show) data X46 = X46 String deriving (Read,Show) data X47 = X47 String deriving (Read,Show) data X48 = X48 String deriving (Read,Show) data X49 = X49 String deriving (Read,Show) data X50 = X50 String deriving (Read,Show) data X51 = X51 String deriving (Read,Show) data X52 = X52 String deriving (Read,Show) data X53 = X53 String deriving (Read,Show) data X54 = X54 String deriving (Read,Show) data X55 = X55 String deriving (Read,Show) data X56 = X56 String deriving (Read,Show) data X57 = X57 String deriving (Read,Show) data X58 = X58 String deriving (Read,Show) data X59 = X59 String deriving (Read,Show) data X60 = X60 String deriving (Read,Show) data X61 = X61 String deriving (Read,Show) data X62 = X62 String deriving (Read,Show) data X63 = X63 String deriving (Read,Show) data X64 = X64 String deriving (Read,Show) data X65 = X65 String deriving (Read,Show) data X66 = X66 String deriving (Read,Show) data X67 = X67 String deriving (Read,Show) data X68 = X68 String deriving (Read,Show) data X69 = X69 String deriving (Read,Show) data X70 = X70 String deriving (Read,Show) data X71 = X71 String deriving (Read,Show) data X72 = X72 String deriving (Read,Show) data X73 = X73 String deriving (Read,Show) data X74 = X74 String deriving (Read,Show) data X75 = X75 String deriving (Read,Show) data X76 = X76 String deriving (Read,Show) data X77 = X77 String deriving (Read,Show) data X78 = X78 String deriving (Read,Show) data X79 = X79 String deriving (Read,Show) data X80 = X80 String deriving (Read,Show) data X81 = X81 String deriving (Read,Show) data X82 = X82 String deriving (Read,Show) data X83 = X83 String deriving (Read,Show) data X84 = X84 String deriving (Read,Show) data X85 = X85 String
records in Haskell
Hi,
It seem that any record, no matter how trivial, can't be much
longer than about 200 lines in Haskell. If a try to compile a
300 line record containing just:
data X = X {
f1 :: String,
f2 :: String,
f3 :: String,
...
f300 :: String
}
It needs about 90M heap in ghc4.06. Whereas a 150 line record
requires less than 6M heap. After this big gap it levels off
to a somewhat more decent exponential increase: a 450 line
record requires about 180M heap.
I could file a bug report, but it seems that all compilers
(ghc4.06, nhc98, hbc0.9994 and hugs) have this problem. So,
is this a fundamental problem ?
Jan
FFI bug ?
Hi, It seems that FFI no longer accepts "ByteArray Int" in ghc 4.06 (it does in ghc 4.04), it looks intentional ? If so, is there another way to pass a String to a C function ? If not, I have attached a small example, to see the result do: tar -xf abc.tar make The result should be: --- gcc -O2 -Wall-c -o at.o at.c rm -f Main.o ghc -O -fglasgow-exts -recomp -dcore-lint -fvia-C -fallow-overlapping-instances -fallow-undecidable-instances -syslib posix -c Main.hs Main.hs:5: Unacceptable argument type in foreign declaration: ByteArray Int When checking declaration: foreign import _ccall "c_test" h_test :: ByteArray Int - IO () Compilation had errors make: *** [Main.o] Error 1 --- My configuration is: (~/project/bug 88) uname -sr SunOS 5.6 (~/project/bug 89) dmesg | grep SUNW cpu0: SUNW,UltraSPARC-IIi (upaid 0 impl 0x12 ver 0x12 clock 270 MHz) SUNW,m64B0 is /pci@1f,0/pci@1,1/SUNW,m64B@2 stdout is /pci@1f,0/pci@1,1/SUNW,m64B@2 major 35 minor 0 SUNW,hme0: CheerIO 2.0 (Rev Id = c1) Found SUNW,hme0 is /pci@1f,0/pci@1,1/network@1,1 SUNW,hme0: Using Internal Transceiver SUNW,hme0: 100 Mbps full-duplex Link Up (~/project/bug 90) cat /etc/release Solaris 2.6 5/98 s297s_hw3smccDesktop_09 SPARC Copyright 1998 Sun Microsystems, Inc. All Rights Reserved. Assembled on 24 April 1998 (~/project/bug 91) ghc --version The Glorious Glasgow Haskell Compilation System, version 4.06 (~/project/bug 92) gcc --version egcs-2.91.66 --- If this is a bug, there is no urgency in fixing it, ghc4.04 works very good too. Regards, Jan abc.tar
Re: drop take [was: fixing typos in Haskell-98]
Marcin 'Qrczak' Kowalczyk wrote: My preference is still (B). (A) is not *very* bad, but should really replicate (-7) "foo" be []? Mine too. Actually after writing my own version of "drop" it turns out that in my case n 0 is a programmer error and n length xs a user error. So what you end up with (if (B) is choosen) is something like: bafDrop :: Int - String - String bafDrop i xs | i length xs = bafError "corrupt BAF file" | otherwise = drop i xs Of course n 0 isn't always a programmer error and you might want to overwrite it. So that would suggest (A), but right now (B) is my favourite, because I can't think of a practical example where n 0 would be a user error. Actually, (A) might be better, the extra check for i 0 is not time consuming anyway. I guess flipping a coin to choose between (A) and (B) would work just as well (provided this doesn't lead to a discussion about what coin to use). Jan P.S. Now that I see the "bafDrop" outside the code, it looks odd. I have no right claiming that the BAF -file- is corrupt at that point. I'll have to rewrite this using some exception Monad. Since this would make the code even bigger I might as well put the entire code of drop in it. So much for code reuse...
FFI on sparc
Hi, I'm using the FFI on a sparc. Everything works fine except when I return a float or double from C to Haskell. My config is: SunOS 5.6 cpu0: SUNW,UltraSPARC-IIi (upaid 0 impl 0x12 ver 0x12 clock 270 MHz) SUNW,m64B0 is /pci@1f,0/pci@1,1/SUNW,m64B@2 stdout is /pci@1f,0/pci@1,1/SUNW,m64B@2 major 35 minor 0 SUNW,hme0: CheerIO 2.0 (Rev Id = c1) Found SUNW,hme0 is /pci@1f,0/pci@1,1/network@1,1 SUNW,hme0: Using Internal Transceiver SUNW,hme0: 100 Mbps full-duplex Link Up Solaris 2.6 5/98 s297s_hw3smccDesktop_09 SPARC Copyright 1998 Sun Microsystems, Inc. All Rights Reserved. Assembled on 24 April 1998 The Glorious Glasgow Haskell Compilation System, version 4.04, patchlevel 1 egcs-2.91.66 The output of my program is: check1 1.23 1.342972116e9 check2 1.23 1.07293442e9 check3 1.23 1.3429812e9 I've attached the tarred sources. Regards, Jan abc.tar
prelude drop and take
Hi, This minor inconsistency has been bothering me for some time: Prelude drop 10 "test" "" Prelude drop (-1) "" "" Prelude drop (-1) "a" " Program error: Prelude.drop: negative argument I got these results from Hugs, but the code is identical to that specified in the Haskell98 standard. Why not an error for the negative argument on an empty list ? Secondly, why not an error for an argument that's too big ? What I would like is this: drop 0 xs = xs drop n (_:xs) | n0 = drop (n-1) xs drop n _ | n0 = error "Prelude.drop: argument too big" | otherwise = error "Prelude.drop: negative argument" Or is there a good reason for drop being able to drop more elements than the list is long ? In which case I would like: drop 0 xs= xs drop n [] | n0 = [] drop n (_:xs) | n0 = drop (n-1) xs drop _ _ = error "Prelude.drop: negative argument" I doubt the second solution would be much less efficient. The first solution should be just as efficient as the one in Haskell98. The same goes for take and splitAt. Regards, Jan
Re: bootstrapping ghc on AIX
Sigbjorn Finne (Intl Vendor) wrote:
Jan Kort [EMAIL PROTECTED] writes:
Thanks, I can get a lot further now, everything apart from
compiling the assembler files and linking the hsc works, but
that will have to wait till tommorow :)
Jan
Great, if you should get stuck while doing this, let us know,
and we'll try to help out.
--Sigbjorn
It looks like I'm either missing more code, or I'm on the wrong
track. I can't find the powerpc assembly code for "JMP_" (should
be in ghc/includes/TailCalls.h), "StgRun" (should be in
ghc/rts/StgCRun.c) and for "StgReturn" (should be in ghc/rts/StgRun.S).
I also tried to compile with USE_MINIINTERPRETER set to 1, but this
led to having to set INTERPRETER_ONLY to 1 as well,
otherwise Hugs_CONSTR_entry would be undefined. I asume that
"INTERPRETER_ONLY" means no compiler ?
In case I'm not missing code and I have to write it myself, if
I don't use native code generation, can I do:
#ifdef powerpc_TARGET_ARCH
StgThreadReturnCode
StgRun(StgFunPtr f) {
f();
return (StgThreadReturnCode)R1.i;
}
#endif
And for the "JMP_":
#if powerpc_TARGET_ARCH
#define JMP_(cont) \
{ \
goto (void *)(cont); \
}
#endif powerpc_TARGET_ARCH
I don't know what to do about StgReturn in ghc/rts/StgRun.S,
judging from the ifdef, the StgReturn is only valid for 386
architectures, so where is it defined on a sparc or an alpha ?
--
Jan Kort
Re: bootstrapping ghc on AIX
Hi, In ghc/Makefile it says: 17 # Order is important! driver/ has to come before includes/ which 18 # again has to come before the rest. 19 # 20 # If we're booting from .hc files, swap the order 21 # we descend into compiler/ and lib/ 22 # 23 ifeq "$(GhcWithHscBuiltViaC)" "NO" 24 SUBDIRS = utils driver includes rts docs compiler lib 25 else 26 SUBDIRS = utils driver includes rts docs lib compiler 27 endif Shouldn't this be 'ifeq "$(GhcWithHscBuiltViaC)" "YES"' ? Jan -- Jan Kort
bootstrapping ghc on AIX
Hi, I tried to bootstrap ghc4.00 on AIX with the "--enable-hc-boot" flag, but I got an error: RTS -K2m -H10m -RTS-g rename/ParseIface.y make[2]: RTS: Command not found make[2]: *** [rename/ParseIface.hs] Error 127 make[1]: *** [boot] Error 1 make: *** [boot] Error 1 It looks like the variable HAPPY in ghc/compiler/Makefile is not defined, but since I'm bootstrapping I don't see why it would have to be defined ? Here are the commands I typed to produce the bug report, which I have gzipped and attached (unfortunately netscape attachments don't get through, so I'll send it as a seperate mail after this one): script myscript uname -srv gcc --version make --version tar -xf ghc-4.00-src.tar cd fptools ./configure --enable-hc-boot make boot exit I also tried to compile ghc2.10, but this gives the same error. I can get "make boot" through by commenting out 6 lines related to Happy and some dependencies in "ghc/compiler/Makefile". But the "make all" does not get very far, the first thing it complains about is "etext" being undefined in "ghc/includes/ClosureMacros.h", I changed to "_etext" to "etext" in "ghc/includes/config.h". Then I added a line to "ghc/rts/MBlock.c": #define ASK_FOR_MEM_AT 0x5000 Then I copied the aix aix_TARGET_OS ifdef from "ghc/rts/PrimOps.hc" to "ghc/includes/PrimOps.h". The I got the following error: make[4]: Leaving directory `/pluto/home/jan/fptools/ghc/rts/gmp/mpz' ../../ghc/driver/ghc -I../includes -I. -Igum -optc-Wall -optc-W -optc-Wstrict-prototypes -optc-Wmissing-proto Assembler: /tmp/ccHbwaUa.s: line 61: 1252-044 The specified source character 0xc does not have meaning in the command context used. /tmp/ccHbwaUa.s: line 61: 1252-142 Syntax error. make[2]: *** [StgRun.o] Error 1 make[1]: *** [all] Error 1 make: *** [all] Error 1 I removed the assembler files from "ghc/rts/Makefile" (probably not a good idea). Then I got the error: ==fptools== make all --no-print-directory -r; in /pluto/home/jan/ghc4/fptools/ghc/lib/std make[3]: *** No rule to make target `Array.o', needed by `libHS.a'. Stop. make[2]: *** [all] Error 1 make[1]: *** [all] Error 1 make: *** [all] Error 1 Which I don't understand, am I missing some hc files ? Regards, Jan -- Jan Kort
