- Haskell-Cafe - Haskell.org

Functions that return functions
by michael rice 12 Apr '09

12 Apr '09

The following are exercises 5.10, and 5.11 from the Scheme text "Concrete Abstractions" by Max Hailperin, et al. The text at that point is about writing verifiers to check ID numbers such as ISBNs, credit card numbers, UPCs, etc. ====== Exercise 5.10 Write a predicate that takes a number and determines whether the sum of its digits is divisible by 17. Exercise 5.11 Write a procedure make-verifier, which takes f and m as its two arguments and returns a procedure capable of checking a number. The argument f is itself a procedure of course. Here is a particularly simple example of a verifier being made and used. (define check-isbn (make-verifier * 11)) (check-isbn 0262010771) #t The value #t is the "true" value; it indicates that the number is a valid ISBN. As we just saw, for ISBN numbers the divisor is 11 and the function is simply f(i,d(i)) = i * d(i). Other kinds of numbers use slightly more complicated functions, but you will still be able to use make-verifier to make a verifier much more easily than if you had to start from scratch. ======= Here's the Scheme check-verifier function I wrote, followed by my humble attempt at a Haskell function that does the same thing. Below that are some verifier functions created with the Scheme make-verifier. Admittedly, functions that return functions are Lispy, but perhaps there a Haskelly way to accomplish the same thing? Michael =============== (define (make-verifier f m) ;f is f(i,d) & m is divisor (lambda (n) (let* ((d (digits n)) (i (index (length d)))) ;(index 3) => (1 2 3) (divides? m (reduce + 0 (map f i d)))))) #f makeVerifier :: (Int -> Int -> Int) -> Int -> (Int -> Bool) makeVerifier f m = \n -> let d = digits n i = [1..(length d)] in \n -> divides m (foldl (+) 0 (map2 f i d)) -- Note: Reduce is just foldl f 0 lst, but map2 is map2 :: (Int -> Int -> Int) -> [Int] -> [Int] -> [Int] map2 f m n = [ f i d | (i,d) <- zip m n] -- And here's my digits function digits :: Int -> [Int] digits 0 = [] digits n = rem n 10 : digits (quot n 10) -- And divides function divides :: Int -> Int -> Bool divides divisor n = 0 == rem n divisor ===== (define check-isbn ;book number (make-verifier * 11)) (define check-upc ;universal product code (make-verifier (lambda (i d) (if (odd? i) d (* 3 d))) 10)) (define check-cc ;credit card (make-verifier (lambda (i d) (if (odd? i) d (if (< d 5) (* 2 d) (+ (* 2 d) 1)))) 10)) (define check-usps ;postal money order (make-verifier (lambda (i d) (if (= i 1) (- d) d)) 9))

4 3

GHC RTS bug: segmentation fault
by Manlio Perillo 12 Apr '09

12 Apr '09

Hi. In a program I have experienced a segmentation fault; not only that but also: * tpp.c:63: __pthread_tpp_change_priority: Assertion `new_prio == -1 || (new_prio >= __sched_fifo_min_prio && new_prio <= __sched_fifo_max_prio)' failed. * internal error: removeThreadFromQueue: not found (GHC version 6.8.2 for i386_unknown_linux) Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug I have managed to write a program that reproduces the problem: http://hpaste.org/fastcgi/hpaste.fcgi/view?id=3714 Compile with: $ghc --make -threaded -O2 main.hs Execute with: $./main +RTS -A128M -s -c -N4 -RTS 5000000 500 This should produce a segmentation fault (or one of the previous errors). I strongly suspect that the cause is an uncaught stack overflow. Executing the code with only 1 threads, there are no more problems. Increasing the thread stack size (-k1M), there are no more problems. If the problem is really an uncaught stack overflow, I'm rather sure that a more simple test can be written, to reproduce the problem. I'm on Linux Debian Etch i386; GHC 6.8.2. Some notes about the program: - when no segmentation fault occurs, the program requires about 1 GB of memory - there is no real parallelization, only 1 CPU is used - the mapReduce implementation is taken from Real World Haskell - the code is adapted from the code I'm using in a project Thanks Manlio Perillo

2 1

Re: [Haskell-cafe] Functions that return functions
by michael rice 12 Apr '09

12 Apr '09

Cool. I expected as much. My ISP seems to be swamped (Easter?) so maybe more later after I look at your remarks. Thanks. Michael --- On Sun, 4/12/09, Daniel Fischer <daniel.is.fischer(a)web.de> wrote: From: Daniel Fischer <daniel.is.fischer(a)web.de> Subject: Re: [Haskell-cafe] Functions that return functions To: haskell-cafe(a)haskell.org Cc: "michael rice" <nowgate(a)yahoo.com> Date: Sunday, April 12, 2009, 12:15 PM Am Sonntag 12 April 2009 17:38:53 schrieb michael rice: > The following are exercises 5.10, and 5.11 from the Scheme text "Concrete > Abstractions" by Max Hailperin, et al. The text at that point is about > writing verifiers to check ID numbers such as ISBNs, credit card numbers, > UPCs, etc. > > ====== > > Exercise 5.10 > Write a predicate that takes a number and determines whether the sum of its > digits is divisible by 17. > > Exercise 5.11 > Write a procedure make-verifier, which takes f and m as its two arguments > and returns a procedure capable of checking a number. The argument f is > itself a procedure of course. Here is a particularly simple example of a > verifier being made and used. > > (define check-isbn (make-verifier * 11)) > > (check-isbn 0262010771) > #t > > The value #t is the "true" value; it indicates that the number is a valid > ISBN. > > As we just saw, for ISBN numbers the divisor is 11 and the function is > simply f(i,d(i)) = i * d(i). Other kinds of numbers use slightly more > complicated functions, but you will still be able to use make-verifier to > make a verifier much more easily than if you had to start from scratch. > > ======= > > Here's the Scheme check-verifier function I wrote, followed by my humble > attempt at a Haskell function that does the same thing. Below that are some > verifier functions created with the Scheme make-verifier. Admittedly, > functions that return functions are Lispy, but perhaps there a Haskelly way > to accomplish the same thing? Functions returning functions are quite natural in Haskell. Since usually functions are written in curried form, a function returng a function corresponds to a function of multiple arguments applied to only some of them. > > Michael > > =============== > > > (define (make-verifier f m) ;f is f(i,d) & m is divisor > (lambda (n) > (let* ((d (digits n)) > (i (index (length d)))) ;(index 3) => (1 2 3) > (divides? m (reduce + 0 (map f i d)))))) #f > > makeVerifier :: (Int -> Int -> Int) -> Int -> (Int -> Bool) > > makeVerifier f m = \n -> let d = digits n > > i = [1..(length d)] > > in \n -> divides m (foldl (+) 0 (map2 f i d)) > > makeVerifier f m n = divides m . foldl (+) 0 $ zipWith f [1 .. ] (digits n) just call makeVerifier f m to get your verification function :) If you don't want to name the last argument: makeVerifier f m = divides m . foldl (+) 0 . zipWith f [1 .. ] . digits more point-freeing would be obfuscation. Instead of foldl (+) 0, you could also just write sum. > > -- Note: Reduce is just foldl f 0 lst, but map2 is > map2 :: (Int -> Int -> Int) -> [Int] -> [Int] -> [Int] > map2 f m n = [ f i d | (i,d) <- zip m n] map2 is zipWith, already in the Prelude. > > -- And here's my digits function > digits :: Int -> [Int] > digits 0 = [] > digits n = rem n 10 : digits (quot n 10) Unless you are desperate for speed and sure you deal only with positive numbers (or know that you really want quot and rem), better use div and mod. Those give the commonly expected (unless your expectation has been ruined by the behaviour of % in C, Java...) results. > > -- And divides function > divides :: Int -> Int -> Bool > divides divisor n = 0 == rem n divisor > > ===== > > (define check-isbn ;book number > (make-verifier > * > 11)) > > (define check-upc ;universal product code > (make-verifier > (lambda (i d) (if (odd? i) d (* 3 d))) > 10)) > > (define check-cc ;credit card > (make-verifier > (lambda (i d) (if (odd? i) d (if (< d 5) (* 2 d) (+ (* 2 d) 1)))) > 10)) > > (define check-usps ;postal money order > (make-verifier > (lambda (i d) (if (= i 1) (- d) d)) > 9))

1 0

Broken beyond repair: Control.Concurrent.SampleVar
by ChrisK 12 Apr '09

12 Apr '09

Hello all, The SampleVar module in base is not exception safe. I believe that there is no way to fix this module to be exception safe while retaining the current behavior. The problem with the current behavior is that the writeSampleVar pretends to know how many blocked reader threads are waiting on a value. In reality these blocked threads may have been killed. When writeSampleVar sees blocked threads it does a blocking putMVar and then decrements the blocked reader thread count. If any two of the blocked reader threads ever die then eventually you get a blocked writer. I can find no efficient way to retain the current behavior in the presence of exceptions. The logic of the current behavior is that the blocked readers mean that the previously written value should not be discarded, the previous value belongs to the next blocked reader. I propose that SampleVar be either removed, or replaced with a slightly different exception safe version. I propose not considering the previously written value to belong to a blocked reader, and to replace it with the new value. I will open a ticket about this when I get more time. -- Chris

1 0

Bug in Data.Stream?
by Peter Verswyvelen 12 Apr '09

12 Apr '09

Since I realized my code was always using infinite lists, I replaced it by Data.Stream. However, my code stopped working. The problem is with these functions: scan :: (a -> b -> a) -> a -> Stream b -> Stream ascan f z (Cons x xs) = z <:> scan f (f z x) xs-- | @scan'@ is a strict scan.scan' :: (a -> b -> a) -> a -> Stream b -> Stream ascan' f z (Cons x xs) = z <:> (scan' f $! (f z x)) xs They are too strict I think. My code works again when I add a lazy pattern match: scan f z ~(Cons x xs) = z <:> scan f (f z x) xs scan' f z ~(Cons x xs) = z <:> (scan' f $! (f z x)) xs This is justified since they then behave like scanl on lists. However it seems this package is used a lot, so maybe some code depends on this strictness. What to do? PS: Why does scanl' not exist in Data.List?

2 2

How to get glyph outline from ttf font.
by Dmitry V'yal 11 Apr '09

11 Apr '09

Greetings. I'm trying to render some glyphs from ttf font to svg image using gtk2hs cairo binding. Since I can find nothing appropriate in gtk2hs API, I decided to draw outlines with bezier curves myself. But how to get them out of font? As far as I know, freetype library is capable of extracting outlines, but is there any haskell binding for it which supports this functionality? Here [1] one such lib was mentioned, but it wasn't availible online that time. Have situation changed today? Thanks in advance [1] http://www.nabble.com/Re%3A-Rendering-TTF-fonts-in-Haskell-and-OpenGL-p1563…

4 3

what does the hidden package error mean?
by Anatoly Yakovenko 11 Apr '09

11 Apr '09

I am trying to build ParseP on the latest ghc, and i am getting this error: Text/ParserCombinators/ParseP/Interface.hs:26:17: Could not find module `Data.ByteString.Char8': it is a member of package bytestring-0.9.1.4, which is hidden what does "hidden" mean? why does this work in ghci > import Data.ByteString > import Data.ByteString.Char8 > without any problems

5 5

How to define a common return and bind?
by Bas van Dijk 11 Apr '09

11 Apr '09

Hello, Suppose you have defined a monad transformer such as: > newtype T1 m a = T1 { unT1 :: A1 m a } Where 'A1 m' is an arbitrary monad of your choosing. For this discussion we just take the identity: > type A1 m a = m a -- (can be any monad) If you want to define a Monad instance for 'T1 m' you generally do this: instance Monad m => Monad (T1 m) where return = T1 . return m >>= f = T1 $ unT1 m >>= unT1 . f (I know I can use the 'GeneralizedNewtypeDeriving' language extension to automatically derive a Monad but suppose that isn't available) Now when I define a new monad transformer: > newtype T2 m a = T2 { unT2 :: A2 m a } Where 'A2 m' is again an arbitrary monad of your choosing but for now just the identity: > type A2 m a = m a -- (can be any monad) The Monad instance for it is almost completely identical to the former: instance Monad m => Monad (T2 m) where return = T2 . return m >>= f = T2 $ unT2 m >>= unT2 . f Note that the only differences are: * a function to convert from the outer monad _to_ the inner monad: 'unT1' and 'unT2' * a function to convert _from_ the inner monad to the outer monad: 'T1' and 'T2' The common parts seem to be: liftReturn from = from . return liftBind from to m f = from $ to m >>= to . f My question is: can these be given suitable and general enough types so that they can be used to define Monad instances for monad transformers? In other words can I use them to write: instance Monad m => Monad (T1 m) where return = liftReturn T1 (>>=) = liftBind T1 unT1 and: instance Monad m => Monad (T2 m) where return = liftReturn T2 (>>=) = liftBind T2 unT2 Thanks, Bas

3 6

cabal haddock and Paths_Xxx
by Martijn van Steenbergen 11 Apr '09

11 Apr '09

Hello, cabal generates a Paths_Xxx file for me which I import and use, but cabal haddock doesn't seem to like it much. If I don't specify the generated module at all in my cabal file, cabal haddock generates visible documentation for the module, which is not what I want: the module should stay internal. If I specify it in either Exposed-Modules or Other-Modules, 'cabal configure; cabal haddock' fails with: > cabal: can't find source for Paths_Xxx in ., dist/build/autogen While 'cabal configure; cabal build; cabal haddock' fails with: > cabal: can't find source for module Paths_Xxx Can I have the best of both, i.e. generate docs without problems and use the generated module? Thanks, Martijn.

2 2

ANN: MSem replacement for QSem
by ChrisK 11 Apr '09

11 Apr '09

"Grasshopper, try to take the stone from my hand" Hello all, After looking at the code for QSem [1], I realize it is not exception safe. And it does more work than needed. I have a proposed replacement module MSem at http://haskell.org/haskellwiki/SafeConcurrent on the wiki. I claim this is a safer, more correct, and more efficient replacement for QSem. But I may be tragically wrong which is where you come in. Please try and find some strange concurrency bug that I have missed. Once there is reviewed replacement code I will propose changing the base package. Exception safe replacements for QSemN and SampleVar are also needed. The MSem code I have can be generalized into an "MSemN", but the semantics are be different from "QSemN". This MSemN only considers the needs of the first blocked waiter instead of all of them. The code MSemN is on the same wiki page, perhaps someone will find this variant useful. Cheers, Chris [1] http://www.haskell.org/ghc/docs/latest/html/libraries/base/src/Control-Conc…

1 0