- Haskell-Cafe - Haskell.org

Re: [Haskell-cafe] Question about data
by David Barbour 22 Aug '11

22 Aug '11

In addition to fixing 'Float', you should define the 'Fractional' instance for MathExpression. This would let you use: let pi = 3.14 :: MathExpression So your instance of Fractional would look like: instance Num MathExpression where (+) a b = Add a b (*) a b = Multiply a b (-) a b = Subtract a b negate b = Subtract 0 b fromInteger = MathFloat . fromInteger abs = undefined signum = undefined instance Fractional MathExpression where fromRational = MathFloat . fromRational (/) a b = Divide a b recip b = Divide 1 b That would give 'fromRational' and 'fromInteger' which are the main numeric conversion functions. Regards, Dave On Fri, Aug 19, 2011 at 1:40 PM, Paul Reiners <paul.reiners(a)gmail.com>wrote: > I've created a simple type declaration: > > data MathExpression = Float > | Add MathExpression MathExpression > | Subtract MathExpression MathExpression > | Multiply MathExpression MathExpression > | Divide MathExpression MathExpression > deriving (Show) > > Now how do I create an instance of MathExpression which is just a Float? >

1 0

ANN: Netwire 1.1.0 and instinct 0.1.0
by Ertugrul Soeylemez 22 Aug '11

22 Aug '11

Hello there, today I have released one new library 'instinct' for neural networks as well as a new revision of my AFRP library 'netwire', which comes along with many new features, but also with largely incompatible changes. Instinct is a library for neural networks, which currently features: * High performance, * arbitrary network models, as long as they don't loop, * reasonably compact memory representation, * backpropagation learning. Netwire is an arrowized functional reactive programming library written with networking applications in mind. Changes include: * Inhibiting signals now contain a SomeException value, * the Wire type has been generalized over the underlying monad. The second change brings up two new features: * Wires over arbitrary monads, hence * pure wires. I appreciate constructive feedback. Enjoy! Greets, Ertugrul -- nightmare = unsafePerformIO (getWrongWife >>= sex) http://ertes.de/

1 0

Is it written down somewhere what sort of things belong in Data.*?
by David Fox 22 Aug '11

22 Aug '11

I'm never quite sure what the distinction is that defines the modules under Data.*. Can anyone explain? -david

2 1

Problem with installing cabal-install on NetBSD.
by Loïc Maury 22 Aug '11

22 Aug '11

Hello, I try to install "cabal-install" on NetBSD. when I launch the installation (by booststrap.sh) and it come to Building cabal-install, it say : Distribution/Compat/FilePerms.hs:33:4 - Not in scope: `throwErrnoPathIfMinus1_' Do you know how I can fix this issue ? Thank you Loïc

1 0

QuasiQuotation with existing parsers
by Ivan Lazar Miljenovic 22 Aug '11

22 Aug '11

For graphviz [1], I'm thinking of adding in some quasi-quoting support for things like record labels [2], to make them easier to use and less verbose to define (i.e. by using actual Dot syntax [3] for them). If I wanted anti-quotation support in those though, what would be the preferable/recommended option? * Extend the current data types to also have the anti-quotation types (but don't export those constructors), adapt the parser accordingly and use that (though for quasi-quotation I'd need to convert the input String into a Text value, as that's what the existing parsers use). However, this means there's a chance that the anti-quotation types will leak through (and possibly cause problems when parsing real Dot code when the anti-quotation notation is found). This second complication could be relaxed with a Bool flag to the parser indicating whether or not anti-quotation should be supported, but it still feels a bit awkward to me. * Define new data types with anti-quotation support with new parsers for them, and after anti-quotation is evaluated convert them to the real data types. Requires duplicate work, etc. [1]: http://projects.haskell.org/graphviz/ [2]: http://hackage.haskell.org/packages/archive/graphviz/2999.12.0.3/doc/html/D… [3]: http://www.graphviz.org/doc/info/shapes.html#record -- Ivan Lazar Miljenovic Ivan.Miljenovic(a)gmail.com IvanMiljenovic.wordpress.com

1 0

custom SQL-to-Haskell type conversion in HDBC
by Henry House 22 Aug '11

22 Aug '11

Does there exist any sample code or other resources on writing a custom SQL-to-Haskell datatype converter instance for use with HDBC that would be accessible to someone just starting with Haskell? The reason I need this is because of this problem (using PostgreSQL): Prelude Database.HDBC.PostgreSQL Database.HDBC> res <- (quickQuery db "select 1::numeric(5,4);" []) Prelude Database.HDBC.PostgreSQL Database.HDBC> res [[SqlRational (1 % 1)]] Prelude Database.HDBC.PostgreSQL Database.HDBC> res <- (quickQuery db "select 1::numeric(5,0);" []) [[SqlRational (1 % 1)]] where db is a database connection. The SQL values 1::numeric(5,4) and 1::numeric(5,0) are supposed to be fixed-precision numbers having 4 and zero significant decimal figures after the decimal point, respectively. Both are offered by HDBC as the same SqlValue, SqlRational (1 % 1) but they are not really the same at all. The precision information has been lost. The native outputs of PostgreSQL, before HDBC's type conversion, are 1.0000 and 1 for 'select 1::numeric(5,4);' and 'select 1::numeric(5,0);', respectively. -- Henry House

4 14

Need help Very strange space behavior
by bob zhang 22 Aug '11

22 Aug '11

Hi all, I thought that Cont Monad is just equivalent to CPS Transformation, so if I have a monadic sum, if I run in Identity Monad, it will suck due to stackoverflow, and if I run it in Cont Monad, it will okay due to tail recursion. So I write a simple program to verify my idea. But to my surprise, the result is unreasonable due to my limited knowledge. All programs are compiled "ghc --make Test.hs -o test && ./test" Thank you in advance, the clearer the better!! (I am really confused) in the comments, suck means stackoverflow. sum0 n = if n==0 then 0 else n + sum0 (n-1) sum1 n = if n==0 then return 0 else sum1 (n-1) >>= \ v -> seq v (return (n+v)) sum2 n k = if n == 0 then k 0 else sum2 n (\v -> k (n + v)) sum3 n k = if n == 0 then k 0 else sum3 n (\ !v -> k (n + v)) sum4 n k = if n == 0 then k 0 else sum4 n (\ v -> seq v ( k (n + v))) sum5 n = if n==0 then return 0 else sum5 (n-1) >>= \ v -> (return (n+v)) -- main = print (sum0 3000000) -- suck reasonable -- main = print (flip runCont id (sum1 3000000)) -- rock 180M memory reasonable, but I am not clear why seq needed here, since its continuation is not applied until n goes to 0 -- main = print (flip runCont id (sum5 3000000)) -- suck -- why? -- main = print (flip runCont (const 0) (sum1 3000000)) -- rock 130M memory -- reasonable -- main = print (flip runCont (const 0) (sum5 3000000)) -- rock 118M memory -- reasonable -- main = print (sum2 3000000 (const 0)) -- a lot of memory (more than 1G) -- I thought sum2 is equivalent to sum5 (when sum5 is in Cont Monad), why? -- main = print (sum3 3000000 (const 0)) -- a lot of memory -- I thought sum3 is equivalent to sum1(Cont Monad), why? -- main = print (runIdentity (sum1 3000000)) -- suck -- exactly what I want -- main = print (sum3 3000000 id) -- a lot of memory -- equivalent to sum1 why? -- main = print (sum4 3000000 id) - - a lot of memory -- equivalent to sum1 why? -- main = print (sum [1 .. 3000000]) -- suck -- src sum = foldl (+) 0 -- reasonable -- main = print (foldl' (+) 0 [1 .. 3000000]) -- rock 1.5M memory -- reasonable -- Best, bob

1 0

Re: [Haskell-cafe] Failed link to mixed-language shared object
by David Banas 21 Aug '11

21 Aug '11

> ghc -dynamic -o ami_test -L. -lami ami_test.o > ./libami.so: undefined reference to `__stginit_haskell98_MarshalArray_' > ./libami.so: undefined reference to `__stginit_haskell98_MarshalError_' > collect2: ld returned 1 exit status I was able to solve this by changing these lines: import MarshalArray import MarshalError to read: import Foreign.Marshal.Array import Foreign.Marshal.Error I don't understand why this worked, since the available symbols in the 2 respective libraries appear totally similar: dbanas@dbanas-eeepc:/usr/lib/ghc-6.12.3$ nm -g haskell98-1.0.1.1/libHShaskell98-1.0.1.1.a | grep 'MarshalArray' MarshalArray__1.o: 00000030 T __stginit_haskell98_MarshalArray 00000000 T __stginit_haskell98_MarshalArray_ dbanas@dbanas-eeepc:/usr/lib/ghc-6.12.3$ nm -g -D haskell98-1.0.1.1/libHShaskell98-1.0.1.1-ghc6.12.3.so | grep 'MarshalArray' 00002b68 T __stginit_haskell98_MarshalArray 00002b24 T __stginit_haskell98_MarshalArray_dyn dbanas@dbanas-eeepc:/usr/lib/ghc-6.12.3$ nm -g base-4.2.0.2/libHSbase-4.2.0.2.a | grep '__stginit_.*Foreign.*Marshal.*Array' 0000006c T __stginit_base_ForeignziMarshalziArray 00000000 T __stginit_base_ForeignziMarshalziArray_ {Some output omitted.} dbanas@dbanas-eeepc:/usr/lib/ghc-6.12.3$ nm -g -D base-4.2.0.2/libHSbase-4.2.0.2-ghc6.12.3.so | grep '__stginit_.*Foreign.*Marshal.*Array' 003dc8cc T __stginit_base_ForeignziMarshalziArray 003dc834 T __stginit_base_ForeignziMarshalziArray_dyn Can anyone shed some light on this? Thanks!

1 0

Embedding Perl RegEx in Haskell
by C K Kashyap 21 Aug '11

21 Aug '11

Dear Haskell folks, I was attempting to do an EDSL that would allow me to describe regular expressions in Hakell and generate Perl as target - https://github.com/ckkashyap/LearningPrograms/blob/master/Haskell/edsl/rege… $ ghci regex.hs GHCi, version 7.0.3: http://www.haskell.org/ghc/ :? for help Loading package ghc-prim ... linking ... done. Loading package integer-gmp ... linking ... done. Loading package base ... linking ... done. Loading package ffi-1.0 ... linking ... done. [1 of 1] Compiling Main ( regex.hs, interpreted ) Ok, modules loaded: Main. *Main> let hello = stringToRegularExpression "hello" *Main> let world = stringToRegularExpression "world" *Main> :t hello hello :: RegularExpression *Main> :t world world :: RegularExpression *Main> let re = listToSequence [ hello, oneOrMore world, hello ] *Main> re hello(world)+hello *Main> :t re re :: RegularExpression I am looking for suggestions on how I could encode the anchors - ^ and $ and also how I could refer to captures. My goal is to be able to extend the idea to create an EDSL that could be used to build Perl tasks - such as filters, report generators etc Arguably this could be seen as a pointless exercise because I could chose to do the scripting in Haskell itself but I have a practical problem that I have to deal with clusters that have a little dated version of Linux where I could not really build the Haskell and where Haskell generated executables don't seem to run. When I started out with this idea, I was thinking of embedding Perl in Haskell but later I thought that perhaps that will sort of beat the point of "DS" bit in EDSL. So, I have started on this idea of trying to encode very precise "tasks". Regards, Kashyap

2 5

Existential question
by Tom Schouten 20 Aug '11

20 Aug '11

{-# LANGUAGE ExistentialQuantification #-} -- Dear Cafe, this one works. data Kl' s i o = Kl' (i -> s -> (s, o)) iso' :: (i -> Kl' s () o) -> Kl' s i o iso' f = Kl' $ \i s -> (\(Kl' kl') -> kl' () s) (f i) -- Is there a way to make this one work also? data Kl i o = forall s. Kl (i -> s -> (s, o)) iso :: (i -> Kl () o) -> Kl i o iso f = Kl $ \i s -> (\(Kl kl) -> kl () s) (f i) {- Couldn't match type `s0' with `s' because type variable `s' would escape its scope This (rigid, skolem) type variable is bound by a pattern with constructor Kl :: forall i o s. (i -> s -> (s, o)) -> Kl i o, in a lambda abstraction The following variables have types that mention s0 s :: s0 (bound at /home/tom/meta/haskell/iso.hs:11:17) In the pattern: Kl kl In the expression: \ (Kl kl) -> kl () s In the expression: (\ (Kl kl) -> kl () s) (f i) -}

2 2