- Haskell-Cafe - Haskell.org

HaskellWiki XHTML not well-formed or valid
by Isaac Dupree 20 May '07

20 May '07

I mentioned this a few months ago in a place no-one noticed, it seems... http://www.haskell.org/haskellwiki/Talk:Haskell#Well-formedness_.2F_validity Please fix, or tell me who to contact (and how), or how to fix it myself? :-) Isaac

1 0

Random idea
by Andrew Coppin 20 May '07

20 May '07

Greetings. I was thinking... we already have Lambdabot sitting in an IRC channel. How hard would it be to mangle Lambdabot to the point where it works over HTTP? You know - so you could type some Haskell into a form on a web page, hit [submit], and get the result sent back to you? (Again, assuming it can be computed in a sane amount of time/space, and truncated to some reasonable textual length.) I think it might be kinda neat to have such a thing linked from the Haskell.org homepage. You know, sort of "hey, wanna try Haskell without installing any stuff? Click here!" ...on the other hand, maybe I'm just strange... :-$

5 6

Debunking tail recursion
by Jules Bean 20 May '07

20 May '07

A conversation on #haskell just showed that it's quite hard to explain (at least it is for me) to people attached to tail recursion why that is a red herring in Haskell. I had a poke around the wiki and couldn't see a page which explains it clearly. In fact http://www.haskell.org/haskellwiki/Performance/Accumulating_parameter is actuall slightly misleading since it suggests that the performance gain is from the tail recursion whereas really it's from the change from a function which accumulates an enormous thunk to an function which can do the work as it goes along. If someone has the energy and the clarity to explain this clearly and make a wiki page about it, that would be great. If someone finds an existing wiki page I'm too unperceptive to notice, even better! Jules

7 9

Profiling, measuring time
by Steffen Mazanek 20 May '07

20 May '07

Hello, I have written a function f, that performs a quite complex operation on its argument. Furthermore I have another function genInput that takes a number and constructs an argument for f of this size. What I want now is a list [(n,time)] that gives me for every size of input n the time, f needs to compute this input. How can I do this? I have found the module Time and tried diffClockTimes, but I ever got zero as a delta. Help is greatly appreciated. There is probably a way to do this via ghcs profiling but I hope that there is also a "normal" Haskell way of doing this. Best regards, Steffen -- Dipl.-Inform. Steffen Mazanek Institut für Softwaretechnologie Fakultät Informatik Universität der Bundeswehr München 85577 Neubiberg Tel: +49 (0)89 6004-2505 Fax: +49 (0)89 6004-4447 E-Mail: steffen.mazanek(a)unibw.de

2 2

Multiple DB drivers in HDBC
by kpierre 20 May '07

20 May '07

Hello. I have a problem with migration to HDBC 1.1. My application had transparent support for multiple database types, because of HDBC's Connection object. How can I preserve this feature with minimal code changes? I tried to use ConnWrapper, as migration guide suggest, but it seems there's no way to convert arbitrary IConnection object into ConnWrapper. Thanks in advance, pierre -- View this message in context: http://www.nabble.com/Multiple-DB-drivers-in-HDBC-tf3784996.html#a10703798 Sent from the Haskell - Haskell-Cafe mailing list archive at Nabble.com.

1 0

New YAHT maintainer
by Eric Y. Kow 20 May '07

20 May '07

Hi, I just wanted to announce that Sam Hughes will be taking over as Keeper of the YAHT. YAHT is the venerable Yet Another Haskell Tutorial. It is available both as a wikibook and as a PDF (*). The Keeper of the YAHT has the grave duty of synchronising the YAHT wikibook and tex source, and releasing an updated PDF from time to time. You can send modifications to YAHT either by (a) updating the wikibook or (b) darcs send'ing a patch to the tex source (patches will go to the Haskell Café mailing list for community review). Many thanks to Sam for stepping up! * http://darcs.haskell.org/yaht/yaht.pdf -- Eric Kow http://www.loria.fr/~kow PGP Key ID: 08AC04F9 Merci de corriger mon français.

1 0

idlelog
by Leandro Lisboa Penz 20 May '07

20 May '07

Hi I made a program that detects user presence in a linux box by checking for keyboard interruptions in /proc/interrupts. Problem is, it is supposed to run for a long time, and it keeps about 40MB for itself. Yeah, this is one more "help me with this memory problem" mails... The code can be found in http://hpaste.org/1907, and in "darcs get http://lpenz.50webs.com/idlelog I guess I'm having some conceptual problem. Cheers Leandro Penz

2 1

Re: [Haskell-cafe] Pesky monads...
by Matthew Cox 20 May '07

20 May '07

It occurred to me that the predicate will generally be a monadic function itself, so here's a refined version: :: Monad m => (a -> m Bool) -> (a -> m a) -> a -> m a untilM pred f x = do c <- pred x if c then return x else f x >>= untilM pred f Think of a computation in the State monad, a predicate will probably want to consult that state. Matthew Cox ----- Original Message ----- From: Matthew Cox To: <haskell-cafe(a)haskell.org> Sent: Sat, 19 May 2007 13:10:40 -0600 Subject: Re: [Haskell-cafe] Pesky monads... You can define one: untilM :: Monad m => (a -> Bool) -> (a -> m a) -> a -> m a untilM pred f x | pred x = return x | otherwise = f x >>= untilM pred f Matthew Cox ----- Original Message ----- From: Andrew Coppin To: haskell-cafe(a)haskell.org Sent: Sat, 19 May 2007 19:49:40 +0100 Subject: [Haskell-cafe] Pesky monads... I've been getting some pretty weird complaints from the type checker. And I just figured out why! Grr... Is there any function that does the same thing as "until", but in a monad? _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe(a)haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

2 1

idlelog
by Leandro Lisboa Penz 19 May '07

19 May '07

Hi I made a program that detects user presence in a linux box by checking for keyboard interruptions in /proc/interrupts. Problem is, it is supposed to run for a long time, and it keeps about 40MB for itself. Yeah, this is one more "help me with this memory problem" mails... The code can be found in http://hpaste.org/1907, and in "darcs get http://lpenz.50webs.com/idlelog I guess I'm having some conceptual problem. Cheers Leandro Penz

2 1

Who lives in a heap like this?
by Andrew Coppin 19 May '07

19 May '07

Greetings. I have just implemented a heap. But... um... I can't acutally figure out *which kind* of heap it is! LOL. Any ideas? (Seems to work really well, whatever it is. Oh, and I discovered that you can sort data just by shoving it all into a heap, and then taking it all out again. Apparently this is a standard algorithm, and it's known as a "heap sort", unsurprisingly. You learn something every day...) module Heap where import Data.List (intersperse, unfoldr) data Heap t = Node !t !(Heap t) !(Heap t) | Null instance (Show t) => Show (Heap t) where show = work "" where work p (Null) = p ++ "-" work p (Node v t0 t1) = concat $ intersperse "\n" $ [work (' ':p) t0, p ++ show v, work (' ':p) t1] empty = Null is_empty Null = True is_empty _ = False insert v (Null) = Node v Null Null insert v (Node v0 t0 t1) = let lo = min v v0 hi = max v v0 in Node hi (insert lo t1) t0 get_max (Null) = error "heap is empty" get_max (Node v _ _) = v delete_max (Null) = error "heap is empty" delete_max (Node _ Null Null) = Null delete_max (Node _ Null t1) = t1 delete_max (Node _ t0 Null) = t0 delete_max (Node _ t0 t1) | get_max t0 > get_max t1 = Node (get_max t0) (delete_max t0) t1 | otherwise = Node (get_max t1) t0 (delete_max t1) size (Null) = 0 size (Node _ t0 t1) = 1 + (size t0) + (size t1) from_list :: (Ord t) => [t] -> Heap t from_list = foldr insert empty to_list :: (Ord t) => Heap t -> [t] to_list = unfoldr (\h -> if is_empty h then Nothing else Just (get_max h, delete_max h)) heap_sort :: (Ord t) => [t] -> [t] heap_sort = to_list . from_list

2 1