October 2005 - Haskell-Cafe

Hi, I'm a Bolivian student working in my proyect of grade...
by Soujiro Seta 12 Oct '05

12 Oct '05

Hello, sorry for this bother, I'm a student of computer science of one university of Bolivia (UMSS, Cochabamba - Bolivia http://www.umss.edu.bo/) In my university the principal tool is Java and the OO paradigm, but for the area of programming languages we use the functional programming paradigm and Haskell. I´m working in my project of grade, and one teacher suggest me to use something new, in this case "Hugs98 for .NET" for develop something interesant. At this time I have learned about ASP.NET and C# and I think to develop a Web Service, but I'm not so happy with the idea of my project, please, can you suggest me something interesant to develop with Hugs98 for .NET?, in other words, I'm asking you about a something interesant thing to develop using the .NET framework and haskell, something interesant to apply the functional paradigm and a intelligent architecture to implement it. Thanks, and sorry for my bad english :( Best regards. _________________________________________________________________ Horóscopo, tarot, numerología... Escucha lo que te dicen los astros. http://astrocentro.msn.es/

1 0

Papers from the 2005 Haskell Workshop (Tallinn)?
by Dimitry Golubovsky 12 Oct '05

12 Oct '05

The papers presented at the Workshop are already available in the ACM library which requires membership/subscription to read full text PDFs. Are there any plans to make those papers available anywhere else on the Web without subscription? -- Dimitry Golubovsky Anywhere on the Web

4 3

puzzling memory leak? (GHC)
by Young Hyun 11 Oct '05

11 Oct '05

I'm seeing a very puzzling behavior that would be considered a memory leak, or unwanted retention of objects, in other languages. If I've made a newbie mistake, could someone point out what I am doing wrong? The process size (resident size and not just virtual size) of my program just grows and grows, passing a few hundred MB's in just 30 seconds, until I'm forced to kill the process. I get the same problem with GHC 6.4 and 6.4.1 on Solaris 2.8 and MacOS X 10.4.2, and GHC 6.2.2 on Solaris 2.8. Here's what the program does. It reads a binary file containing a large number of varying-length structured records. Some of these records are corrupt (e.g., truncated), and it is the job of my program to write out a new file with corrupted records filtered out. I decided to take advantage of Haskell's lazy evaluation to separate out the task of input parsing from the task of filtering as follows: import qualified Data.Map as Map main = do contents <- getContents fixArts $ parseArts contents [] (Map.singleton "offset" 0) The invoked functions have the following signatures: parseArts :: [Char] -> [Char] -> HeaderMap -> [Either ParseError Arts] fixArts :: [Either ParseError Arts] -> IO () where type HeaderMap = Map.Map String Word data Arts = Arts { ar_version::Int, ar_flags::Word, ar_data_length::Int, ar_timestamp::Word, ar_src_ip::Word, ar_dst_ip::Word, ar_list_id::Word, ar_cycle_id::Word, ar_rtt::Word, ar_rtt_mod::Word, ar_hop_distance::Int, ar_dst_replied::Bool, ar_halt_reason::Int, ar_halt_data::Int, ar_reply_ttl::Int, ar_hops::[ArtsHop], ar_offset::Int, ar_bytes:: [Char] } data ArtsHop = ArtsHop { ar_hop_num::Int, ar_hop_addr::Word, ar_irtt::Word, ar_num_tries::Int } data ParseError = ParseError { pe_message::String, pe_offset::Int, pe_bytes::[Char] } Here's the body of fixArts: fixArts ((Left x):xs) = do hPutChar stderr '\n' hPutStrLn stderr $ "Parse error: " ++ pe_message x hPutHexDump stderr bytes offset fixArts xs where bytes = pe_bytes x offset = pe_offset x {-- XXX normally: do putStr (ar_bytes x); fixArts xs --} fixArts ((Right x):xs) = fixArts xs fixArts [] = return () -------------------------------- The function hPutHexDump simply prints out a hex dump of a [Char] (and commenting out the call makes no difference). Unless I'm badly mistaken, nothing about fixArts suggests that it would leak memory. So, it would appear parseArts is somehow to blame for the memory leak, but this is where it gets weird. If I just slightly change fixArts (see below), then I no longer get a memory leak (the process size stays at just over 3MB): fixArts ((Right x):xs) = do hPutArts stderr x fixArts xs The function hPutArts :: Handle -> Arts -> IO () simply prints out the fields of the Arts object. So why should this change fix the memory leak if parseArts is truly at fault? The function parseArts has the following basic form (though the actual implementation is much more involved): parseArts (x:xs) ... = (createArts x) : parseArts xs So, even without seeing all of the code, does anyone have any clues about what may be wrong? --Young

4 4

RE: [Haskell-cafe] puzzling memory leak? (GHC)
by Simon Marlow 11 Oct '05

11 Oct '05

On 11 October 2005 07:34, Young Hyun wrote: > What intuition should be my guide about the "proper way" of employing > lazy evaluation? It's not yet clear to me what you mean by "the > system needs to hang onto all the input data ...". It seems > counterintuitive for the evaluator to hang onto partially evaluated > functions and input data when it can be proven (through static > analysis) that references to that data can no longer exist in the > remainder of the program execution; for example, consider > > case head $ drop 500000 $ parseArts ... of > Right x -> hPutArts stderr x This example should not have a space leak (try it). Space leaks arise when large unevaluated expressions build up, or there are unevaluated expressions that refer to large amounts of data. The solution is usually to insert some strictness, to force evaluation of the offending expressions, assuming the evaluated representation is smaller than the unevaluated one. I suggest you try out space profiling (with GHC or nhc98), and try various modifications to your code to get a handle on what's happening. It's hard to tell exactly where the space leak is in your original code, because it depends on the actual definition of parseArts. Ben's analysis is definitely plausible. Take a look at the code and ask yourself: does each element of the list refer to previous elements, or to an expression accumulated while traversing the list? If this is the case, and you never actually evaluate the head of the list, then the entire list will be retained. A good way to begin to understand space leaks is to study foldl (see many previous threads on this list). Cheers, Simon

1 0

NewBinary: What is the license?
by Joel Reymont 08 Oct '05

08 Oct '05

Folks, NewBinary is based on nhc98's Binary and nhc98 is GPL. Does this mean that I cannot use NewBinary in commercial software? I'm afraid I just wasted a few days of work then :-(. Thanks, Joel -- http://wagerlabs.com/idealab

2 2

Memoization
by Gerd M 08 Oct '05

08 Oct '05

Hello, I'm trying to use Data.Map to memoize computations. Unfortunately this didn't improve the runtime of f at all, so there must be something wrong with my implementation. Thanks in advance! f 1 s (HMM s0 _ sts) = s ??? sts s0 f t s hmm = memory hmm Map.! (t,s) memory hmm@(HMM s0 sss sts) = Map.fromList [ ((t,s),f' t s hmm) | t <- [1..100], s <- sss, s/=s0 ] f' 1 s (HMM s0 _ sts) = s ??? sts s0 f' t s hmm@(HMM s0 sss sts) = sum [ (memory hmm)Map.!(t-1,s') * (s ??? sts s') | s' <- sss, s' /= s0 ] _________________________________________________________________ Express yourself instantly with MSN Messenger! Download today it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/

6 9

runST
by Joel Reymont 08 Oct '05

08 Oct '05

Folks, Why use runST in the following example at http://www.n-heptane.com/nhlab/repos/NewBinary/Tests/BinDouble.hs ? doubleToInts d = runST ( do arr <- newDoubleArray (1,2) writeDoubleArray arr 1 d i1 <- readIntArray arr 1 i2 <- readIntArray arr 2 return (i1,i2)) Isn't it possible to do the same within the IO monad? Thanks, Joel -- http://wagerlabs.com/idealab

3 2

Re: [Haskell-cafe] MutableByteArray as a Ptr
by Joel Reymont 08 Oct '05

08 Oct '05

This is no longer an issue as I found how to read and write Float and Double at the link below http://www.n-heptane.com/nhlab/repos/NewBinary/Tests/BinDouble.hs Originally, I was looking to get a chunk of memory using getByteArray from NewBinary.Binary and read a Storable after converting the array to a pointer. getByteArray :: BinHandle -> Int -> IO ByteArray getByteArray bh (I# sz) = do (MBA arr) <- newByteArray sz let loop n | n ==# sz = return () | otherwise = do w <- getByte bh writeByteArray arr n w loop (n +# 1#) loop 0# freezeByteArray arr Thanks, Joel On Oct 8, 2005, at 4:49 AM, Donald Bruce Stewart wrote: > Try it and see.. I think I have indeed done this in an old version of > Yi. However, it sounds a bit iffy, and there's no reason not to use a > ForeignPtr-ified malloc'd chunk of memory. -- http://wagerlabs.com/idealab

1 0

Trouble understanding NewBinary
by Joel Reymont 08 Oct '05

08 Oct '05

Folks, In http://www.n-heptane.com/nhlab/repos/NewBinary/NewBinary/Binary.hs there's the following bit of code. What are the I#, S# and J#? instance Binary Integer where put_ bh (S# i#) = do putByte bh 0; put_ bh (I# i#) put_ bh (J# s# a#) = do p <- putByte bh 1; put_ bh (I# s#) let sz# = sizeofByteArray# a# -- in *bytes* put_ bh (I# sz#) -- in *bytes* putByteArray bh a# sz# get bh = do b <- getByte bh case b of 0 -> do (I# i#) <- get bh return (S# i#) _ -> do (I# s#) <- get bh sz <- get bh (BA a#) <- getByteArray bh sz return (J# s# a#) Thanks, Joel -- http://wagerlabs.com/idealab

2 2

MutableByteArray as a Ptr
by Joel Reymont 07 Oct '05

07 Oct '05

Folks, http://www.haskell.org/hdirect/ffi-6.html "GHC's ByteArray and MutableByteArray primitive types are (im)mutable chunks of memory allocated on the Haskell heap, and pointers to these can be passed to foreign imported external functions provided they are marked as unsafe." Does this mean I can pass a MutableByteArray as an argument to Storable's peek and poke? Thanks, Joel -- http://wagerlabs.com/idealab

1 0