- Haskell-Cafe - Haskell.org

Control.Exception try and catch
by Mads Lindstrøm 28 Apr '10

28 Apr '10

Hi From http://haskell.org/ghc/docs/6.12.1/html/libraries/base-4.2.0.0/Control-Exce… "... The difference between using try and catch for recovery is that in catch the handler is inside an implicit block (see "Asynchronous Exceptions") which is important when catching asynchronous exceptions ..." However, 'try' is implemented by calling catch http://haskell.org/ghc/docs/6.12.1/html/libraries/base-4.2.0.0/src/Control-…: try :: Exception e => IO a -> IO (Either e a) try a = catch (a >>= \ v -> return (Right v)) (\e -> return (Left e)) Thus, I wonder, why do 'try' not "inherit" the implicit block mentioned above? Looking at catch: catch :: Exception e => IO a -- ^ The computation to run -> (e -> IO a) -- ^ Handler to invoke if an exception is raised -> IO a catch io h = H'98.catch io (h . fromJust . fromException . toException) I see no call to 'block'. But maybe it is hidden in H'98.catch? And is H'98.catch == Prelude.catch ? /Mads

2 1

Broken ghc documentation links
by haskell＠kudling.de 28 Apr '10

28 Apr '10

To reproduce: 1) Look for "data IO" on Hoogle, e.g. http://haskell.org/ghc/docs/6.12.1/html/libraries/base-4.2.0.0/Prelude.html… 2) Click "source", which redirects to http://haskell.org/ghc/docs/6.12.1/html/libraries/base-4.2.0.0/src/GHC-Type… 3) Observe a "Not Found" page Is there anyone interested in fixing this? Bye, Lenny

4 5

Seeking the correct quote
by Jacques Carette 28 Apr '10

28 Apr '10

I have heard generic programming described tongue-in-cheek as "the kind of polymorphism that a language does not (yet) have". I find this description rather apt, and it matches fairly what I see called 'generic' in various communities. But who said this, where and when? Jacques

3 5

Is XHT a good tool for parsing web pages?
by John Creighton 28 Apr '10

28 Apr '10

> Subject: Is XHT a good tool for parsing web pages? > I looked a little bit at XHT and it seems very elegant for writing > concise definitions of parsers by forms but I read that it fails if > the XML isn't strict and I know a lot of web pages don't use strict > XHTML. Therefore I wonder if it is an appropriate tool for web pages. > >

5 5

Getting used and available memory
by Mads Lindstrøm 28 Apr '10

28 Apr '10

Hi I have tried haskell.org, Google and Hoolge, but I cannot find any function to give me the available and/or used memory of a Haskell program. Is it just not there? Or am I missing it somehow? /Mads

4 5

happstack/SOAP
by Johannes Waldmann 28 Apr '10

28 Apr '10

Hi - I'm looking for an example/demo happstack server that handles SOAP requests. - Thanks, J.W.

2 1

A newbie question ?
by zaxis 28 Apr '10

28 Apr '10

newtype TypeMap = TypeMap (Map.Map TypeRep Dynamic) lookup :: Typeable a => TypeMap -> Maybe a lookup (TypeMap mp) = res where res = liftM (fromJust . fromDynamic) $ Map.lookup (typeOf $ fromJust res) mp It seems that the `res` in `fromJust res` has not been defined ? Sincerely! ----- fac n = let { f = foldr (*) 1 [1..n] } in f -- View this message in context: http://old.nabble.com/A-newbie-question---tp28383563p28383563.html Sent from the Haskell - Haskell-Cafe mailing list archive at Nabble.com.

3 3

mtl design (was ``What do _you_ want to see in FGL?'')
by Bradford Larsen 28 Apr '10

28 Apr '10

On Mon, Apr 26, 2010 at 7:35 AM, Ivan Lazar Miljenovic <ivan.miljenovic(a)gmail.com> wrote: > Whilst freezing it is an option, I feel that this will lead to the same > problems that we already face with mtl: most people agree/know that the > approach/design is bad, but we keep using it because there's no (one) > viable alternative to be used (and thus mtl stays in the Platform, which > means more people use it, and thus we have this vicious cycle). Ivan, Could you elaborate on what you mean regarding the mtl approach & design? What makes them bad? Sincerely, Brad

2 1

Number of CPUs/cores?
by Dimitry Golubovsky 27 Apr '10

27 Apr '10

Hi, Does there exist a Haskell library function that returns the number of CPUs/cores (in portable way) on a computer where the program calling it runs? Thanks. -- Dimitry Golubovsky Anywhere on the Web

3 4

Run haskell program in emacs without typing "main" in the ghci buffer.
by Zura_ 27 Apr '10

27 Apr '10

Hello, Is it possible to run haskell program in emacs without typing "main" in the ghci buffer? Assuming "main" function exists of course. Or, maybe automate sending "main\n" string to ghci buffer input. In other words, I want edit/run/see result style session. Thanks in advance, Zura -- View this message in context: http://old.nabble.com/Run-haskell-program-in-emacs-without-typing-%22main%2… Sent from the Haskell - Haskell-Cafe mailing list archive at Nabble.com.

4 4

Trouble with the FFI, Cabal, and GHCI -- installed libraries have invalid FFI pointers
by John Millikin 27 Apr '10

27 Apr '10

I have two modules. One is a library, installed with Cabal: ------------------------------------------------------------------------- -- Data/Text/IDN/StringPrep.hs {-# LANGUAGE ForeignFunctionInterface #-} module Data.Text.IDN.StringPrep where import Foreign foreign import ccall "stringprep.h &stringprep_nameprep" profileNameprep :: Ptr () ------------------------------------------------------------------------- The other is the main program, either interpreted directly or compiled and run: ------------------------------------------------------------------------- -- TestIDN.hs {-# LANGUAGE ForeignFunctionInterface #-} module Main where import Foreign import Data.Text.IDN.StringPrep foreign import ccall "stringprep.h &stringprep_nameprep" stringprep_nameprep :: Ptr () main = do putStrLn $ "stringprep_nameprep = " ++ show stringprep_nameprep putStrLn $ "profileNameprep = " ++ show profileNameprep ------------------------------------------------------------------------- Now, given that both modules import the exact same C symbol, I would expect that the pointers would be the same when running. This is the case for compiled code: ------------------------------------------------------------------------- $ ghc --make TestIDN [1 of 1] Compiling Main ( TestIDN.hs, TestIDN.o ) Linking TestIDN ... $ ./TestIDN stringprep_nameprep = 0x000000000067e3e0 profileNameprep = 0x000000000067e3e0 ------------------------------------------------------------------------- However, if I use ghci / runhaskell, the pointers are different: ------------------------------------------------------------------------- $ rm TestIDN TestIDN.o TestIDN.hi $ runhaskell TestIDN.hs stringprep_nameprep = 0x00007feebe5fe4e0 profileNameprep = 0x0000000040a908bc ------------------------------------------------------------------------- Specifically, the pointer loaded from the library ('profileNameprep') is invalid -- attempting to pass it to libidn will result in errors, and trying to poke around in it yields segmentation faults. I'm using GHC 6.12.1, on Linux. This problem occurs with both dynamic and static libraries. Does anybody have suggestions about how to solve or work around it?

1 0

Re: Sorting Types
by John Creighton 27 Apr '10

27 Apr '10

> I was wondering if it is possible to sort types in hakell and if so what > language extension I should use. Not sure if > this is possible but here is my attempt: > > (I'm aware I don't need so many pragmas > > {-# LANGUAGE GADTs #-} > {-# LANGUAGE TypeFamilies #-} > {-# LANGUAGE UndecidableInstances #-} > {-# LANGUAGE MultiParamTypeClasses #-} > {-# LANGUAGE FlexibleInstances #-} > {-# LANGUAGE TypeSynonymInstances #-} > data Z=Z deriving (Show) > data S i=S i deriving (Show) > data family N a > type family Add n m > type instance Add Z m = m > type instance Add m Z = m > type instance Add (S n) (S m) = S (S (Add n m)) > --14 > type family LT a b > data Cat=Cat > data Dog=Dog > data Fish=Fish > type family Sort a --19 > data And a b=And a b > > type instance LT Dog Z = Cat > type instance LT Fish Z = Dog > type instance LT a (S i) = LT (LT a Z) i > type instance Sort (And a (LT a i))=And (LT a i) a > > I get the following error: > > Illegal type synonym family application in instance: And a (LT a i) > In the type synonym instance declaration for 'Sort' > Failed, modules loaded: none, > > >

3 2

The instability of Haskell libraries
by John Goerzen 27 Apr '10

27 Apr '10

It is somewhat of a surprise to me that I'm making this post, given that there was a day when I thought Haskell was moving too slow ;-) My problem here is that it has become rather difficult to write software in Haskell that will still work with newer compiler and library versions in future years. I have Python code of fairly significant complexity that only rarely requires any change due to language or library changes. This is not so with Haskel. Here is a prime example. (Name hidden because my point here isn't to single out one person.) This is a patch to old-locale: Wed Sep 24 14:37:55 CDT 2008 xxxxx(a)xxxxx.xxxx * Adding 'T' to conform better to standard This is based on information found at http://en.wikipedia.org/wiki/ISO_8601#Combined_date_and_time_representations diff -rN -u old-old-locale/System/Locale.hs new-old-locale/System/Locale.hs --- old-old-locale/System/Locale.hs 2010-04-23 13:21:31.381619614 -0500 +++ new-old-locale/System/Locale.hs 2010-04-23 13:21:31.381619614 -0500 @@ -79,7 +79,7 @@ iso8601DateFormat mTimeFmt = "%Y-%m-%d" ++ case mTimeFmt of Nothing -> "" - Just fmt -> ' ' : fmt + Just fmt -> 'T' : fmt A one-character change. Harmless? No. It entirely changes what the function does. Virtually any existing user of that function will be entirely broken. Of particular note, it caused significant breakage in the date/time handling functions in HDBC. Now, one might argue that the function was incorrectly specified to begin with. But a change like this demands a new function; the original one ought to be commented with the situation. My second example was the addition of instances to time. This broke code where the omitted instances were defined locally. Worse, the version number was not bumped in a significant way to permit testing for the condition, and thus conditional compilation, via cabal. See http://bit.ly/cBDj3Q for more on that one. I could also cite the habit of Hackage to routinely get more and more pedantic, rejecting packages that uploaded fine previously; renaming the old exception model to OldException instead of introducing the new one with a different name (thus breaking much exception-using code), etc. My point is not that innovation in this community is bad. Innovation is absolutely good, and I don't seek to slow it down. But rather, my point is that stability has value too. If I can't take Haskell code written as little as 3 years ago and compile it on today's platform without errors, we have a problem. And there is a significant chunk of code that I work with that indeed wouldn't work in this way. I don't have a magic bullet to suggest here. But I would first say that this is a plea for people that commit to core libraries to please bear in mind the implications of what you're doing. If you change a time format string, you're going to break code. If you introduce new instances, you're going to break code. These are not changes that should be made lightly, and if they must be made (I'd say there's a stronger case for the time instances than the s/ /T/ change), then the version number must be bumped significantly enough to be Cabal-testable. I say this with a few hats. One, we use Haskell in business. Some of these are very long-term systems, that are set up once and they do their task for years. Finding that code has become uncompilable here is undesirable. Secondly, I'm a Haskell library developer myself. I try to maintain compatibility with GHC & platform versions dating back at least a few years with every release. Unfortunately, this has become nearly impossible due to the number of untestable API changes out there. That means that, despite my intent, I too am contributing to the problem. Thoughts? -- John

22 59

Final call for participation: Summer school "Applied Functional Programming", Aug 16-27, Utrecht NL, Application deadline: May 1
by Andres Loeh 27 Apr '10

27 Apr '10

[Feel free to advertise the school to others who might be interested.] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Utrecht Summer School in Computer Science: Applied Functional Programming August 16-27, 2010 Utrecht, The Netherlands ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The summer school is intended as a first or an advanced course on Haskell, consisting of lectures and projects to be solved by the participants, working in teams. We mainly target bachelor and master students, but others who are interested are also welcome to apply. The deadline for applications is *** Saturday, May 1, 2010 ***. More information: http://www.cs.uu.nl/wiki/USCS2010 If you have any questions, feel free to contact me. Cheers, Andres -- Andres Loeh, Universiteit Utrecht mailto:andres@cs.uu.nl mailto:mail@andres-loeh.de http://www.andres-loeh.de

1 0

ANN: iteratee-parsec 0.0.3
by Maciej Piechotka 27 Apr '10

27 Apr '10

iteratee-parsec is a package allowing parsec parser using iteratee. Changes in 0.0.3: - Update to transformers 0.2.* (tested with 0.2.1.0) - Add stricter checks for version of dependencies Regards

1 0

Haddock infix constructors in markup
by Ozgur Akgun 27 Apr '10

27 Apr '10

Hi all, If I have the following data type: data Expr = Num Int | Expr :+: Expr | Expr :-: Expr Haddock handles the infix constructors, and generates a very nice output (html in this case) However when I try to reference to one of the infix constructors in the markup of other functions it fails to hyperlink. -- | If the input is 'Num' does magic, if it is ':+:' does even more magic! someFunc :: Expr -> Expr In the output of this markup, the 'Num' is hyperlinked but the ':+:' is not. Is there a different syntax for infix constructors (and for infix functions presumably) or does haddock simple lack this feature? Best, -- Ozgur Akgun

4 8

Curl & UTF8
by Rickard Karlsson 27 Apr '10

27 Apr '10

Hi, I'm trying to download a file in UTF-8 with libcurl(1.3.5) and GHC 6.12: >import Network.Curl >u = "http://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-demo.txt" >main = curlGetString u [] >>= putStrLn . snd Which doesn't print the characters correctly. If i read the file from local storage with getFile it is displayed properly.

3 3

Abstracted File Library?
by Kazu Yamamoto 27 Apr '10

27 Apr '10

Hello, I'm implementing a command which manipulates files both on Unix/Mac and Windows. I was very surprised because there is not "getStatusChangeTime" function. So, I wrote it with CPP. http://github.com/kazu-yamamoto/Mew/blob/master/bin/hs/Stat.hs Another problem is that since '\\' is used as the file separator of Windows, I cannot use regular expression naturally for file path. In many programming languages, '/' is only file separator for programmers. And my friend, who is an expert of Windows API, says to me that Windows API allows '/' as a file separator. Why don't we use '/' on Windows, too? Are there any abstracted file library to solve these problems? Or should I start to write such a library? --Kazu

2 3

Proper Handling of Exceptional IEEE Floating Point Numbers
by Barak A. Pearlmutter 27 Apr '10

27 Apr '10

Comparison of exceptional IEEE floating point numbers, like Nan, seems to have some bugs in ghci (version 6.12.1). These are correct, according to the IEEE floating point standards: Prelude> 0 < (0/0) False Prelude> 0 > (0/0) False Prelude> 0 == (0/0) False But these are inconsistent with the above, and arguably incorrect: Prelude> compare (0/0) (0/0) GT Prelude> compare (0/0) 0 GT Prelude> compare 0 (0/0) GT Prelude> compare (0/0) (1/0) GT Prelude> compare (1/0) (0/0) GT I'd suggest that compare involving a NaN should yield error "violation of the law of the excluded middle" The min and max functions have strange behaviour with regard to NaN, especially when mixed with Infinity: Prelude> max (0/0) (1/0) NaN Prelude> max (1/0) (0/0) Infinity Prelude> min (0/0) (1/0) Infinity Prelude> max (0/0) 0 NaN Prelude> max 0 (0/0) 0.0 Hugs (Version: September 2006) has similar issues: Hugs> compare (0/0) (0/0) EQ Hugs> compare (0/0) 1 EQ Hugs> (0/0) == (0/0) False Hugs> min (0/0) 1 nan Hugs> min 1 (0/0) 1.0 Hugs> max (0/0) 1 1.0 Discuss? -- Barak A. Pearlmutter <barak(a)cs.nuim.ie> Hamilton Institute & Dept Comp Sci, NUI Maynooth, Co. Kildare, Ireland http://www.bcl.hamilton.ie/~barak/

7 18

Re: What do _you_ want to see in FGL?
by Louis Wasserman 26 Apr '10

26 Apr '10

I'm a fan of making FGL more record-based, but definitely keeping the inductive graph style. My own biggest gripe with previous versions of FGL was that the graph implementations were severely underoptimized. In particular, PatriciaTree left open a *lot* of optimization. Ivan, would you like to set up a darcs repo for FGL-Prime or something at code.haskell.org, and start cracking? I'm really excited for this project. Louis Wasserman wasserman.louis(a)gmail.com http://profiles.google.com/wasserman.louis

2 1

Haskell statefull computations and co-algebras?
by Vasili I. Galchin 26 Apr '10

26 Apr '10

Hello, I am fairly conversant in category. In any case, I am re-reading Bart Jacobs co-algebra tutorial citeseer.ist.psu.edu/jacobs97tutorial.html and draft book www.cs.ru.nl/B.Jacobs/CLG/JacobsCoalgebraIntro.pdf. Have others read these? What role do co-algebras play in Haskell stateful computations (as opposed to or complementing monads)? Kind regards, Vasili

1 0

and [] = True; or [] = False
by Bjorn Buckwalter 26 Apr '10

26 Apr '10

Dear all, Does it make good sense that 'and []' returns 'True' and 'or []' returns 'False'? The Haskell Road to Logic, Maths and Programming says so: "The function or takes a list of truth values and returns True if at least one member of the list equals True, while and takes a list of truth values and returns True if all members of the list equal True." "Should the conjunction of all elements of [] count as true or false? As true, for it is indeed (trivially) the case that all elements of [] are true. So the identity element for conjunction is True. Should the disjunction of all elements of [] count as true or false? As false, for it is false that [] contains an element which is true. Therefore, the identity element for disjunction is False." While the above reasoning is fine, and allows straight-forward implementations, it isn't extremely convincing. In particular, it isn't clear that, while simple, the definitions of the first paragraph are the most sensible. Perhaps one of the more mathematically versed readers on the Cafe could enlighten me? What got me thinking about this was the apparently incorrect intuition that 'and xs' would imply 'or xs'. Thanks, Bjorn

6 7

Re: FRP for game programming / artifical life simulation
by Patai Gergely 26 Apr '10

26 Apr '10

> (in my own FRP experiments I have an update thread and a render thread). I wonder how to nicely deal with state that requires communication with the outer world, even though it is functional at heart. For instance, if you want to change a vertex buffer or a texture or whatever during the update, how do you organise your code? Do you maintain separate pure and impure state information blocks? > However, it is not clear if you want to "pull" or "push" such a network >From what I've seen so far, I think you really need both at the same time. Pull for continuous signals and push for events. The tricky part is that events can be derived from signals (cf. Fran's predicate events), and you really don't want the signals depending on these events to be reevaluated even when there are no occurrences. The best you can do with pull is to explicitly keep the previous output when there's no incoming event. However, if you fmap over this output, the function will be evaluated in every step regardless of the lack of change, since functions are not memoised. I believe this is a problem Grapefruit aims to solve with its notion of segmented signals [1]. Deciding between push and pull according to profiling results is a nice idea. :) It might be too expensive to do it adaptively during runtime (the overhead might easily distort the results and thus render them invalid), but treating it as just a flag could give us an interesting architecture to play with. Gergely [1] http://hackage.haskell.org/packages/archive/grapefruit-frp/0.0.0.0/doc/html… -- http://www.fastmail.fm - A fast, anti-spam email service.

2 1

The Poor Man's PVP-Checking Tool
by Neil Brown 26 Apr '10

26 Apr '10

Hi, The issue of a tool to help with checking packages against the Package Versioning Policy (PVP: http://www.haskell.org/haskellwiki/Package_versioning_policy) has come up several times on this list, and it seems to be a generally wanted tool. One of the things desired in such a tool is the ability to check what has been added/changed/removed in the latest version of a package, to help see what kind of version bump is needed according to the PVP. This could also be used to check if past releases have obeyed the PVP. A proper way to build this tool is to parse the types of everything involved (functions, data types, and instances), and check them for being the same (given that the type parameters may have been renamed, or a definition substituted for a synonym and so on). I have instead hacked together a tool in an hour or so that roughly does the job. It uses Perl, which I haven't written for a long time, because that was quickest -- if this bothers you, pretend its a crazy Haskell EDSL or something :-). I've attached the Perl script. It combines the command-line tools diff and sort with a tiny bit of manual processing along with the Haskell tool cabal-install, and brings them to bear on the output of "haddock --hoogle", which already nicely spits out the types of all functions, data types and instances in your library. It looks for differences in the output between two package versions, and tells you what version bump it thinks you need. It errs on the conservative side, I believe -- it should never miss a bump when it's needed, but you may sometimes get a false alarm if you've made a harmless change. As an example, let's take the recent stm release. You run the Perl script with the name and version of both package versions (if you omit the version, it should use the latest): ==== neil@beast ~: perl cmp.perl stm-2.1.2.0 stm-2.1.1.2 Warning: The documentation for the following packages are not installed. No links will be generated to these packages: ffi-1.0, rts-1.0 Warning: The documentation for the following packages are not installed. No links will be generated to these packages: ffi-1.0, rts-1.0 25a26 > Control.Concurrent.STM.TVar:readTVarIO :: TVar a -> IO a 39c40 < @version 2.1.1.2 --- > @version 2.1.2.0 Given previous version number A.B.C.D: It seems you have added something. You must increase C (or A or B) ==== For those who can read diff, you get the diff. At the bottom, the tool works out the required bump; here something was added so C must be increased, and we can see that the package did just that. Let's take another example from my latest development version of my CHP library; here I run the command with one argument (chp, which means the latest version on Hackage): ==== neil@banshee ~/work/chp: perl cmp.perl chp Warning: The documentation for the following packages are not installed. No links will be generated to these packages: rts-1.0 Control/Concurrent/CHP/Alt.hs:116:9: Warning: orphan instance: instance Alternative CHP Warning: The documentation for the following packages are not installed. No links will be generated to these packages: rts-1.0 6a7 > Control.Concurrent.CHP.Alt:instance Alternative CHP 82d82 < Control.Concurrent.CHP.Channels.Creation:class ChannelTuple t 88,92d87 < Control.Concurrent.CHP.Channels.Creation:instance (Channel r w) => ChannelTuple (Chan r w a, Chan r w a) < Control.Concurrent.CHP.Channels.Creation:instance (Channel r w) => ChannelTuple (Chan r w a, Chan r w a, Chan r w a) < Control.Concurrent.CHP.Channels.Creation:instance (Channel r w) => ChannelTuple (Chan r w a, Chan r w a, Chan r w a, Chan r w a) < Control.Concurrent.CHP.Channels.Creation:instance (Channel r w) => ChannelTuple (Chan r w a, Chan r w a, Chan r w a, Chan r w a, Chan r w a) < Control.Concurrent.CHP.Channels.Creation:instance (Channel r w) => ChannelTuple (Chan r w a, Chan r w a, Chan r w a, Chan r w a, Chan r w a, Chan r w a) 102d96 < Control.Concurrent.CHP.Channels.Creation:newChannels :: (ChannelTuple t, MonadCHP m) => m t 149c143 < Control.Concurrent.CHP.Monad:class (MonadIO m) => MonadCHP m --- > Control.Concurrent.CHP.Monad:class (Monad m) => MonadCHP m 155a150 > Control.Concurrent.CHP.Monad:foreverP :: CHP a -> CHP b 156a152 > Control.Concurrent.CHP.Monad:liftIO_CHP :: IO a -> CHP a 160a157 > Control.Concurrent.CHP.Monad:process :: String -> a -> a 164a162 > Control.Concurrent.CHP.Monad:subProcess :: String -> a -> a 169c167 < Control.Concurrent.CHP.Parallel:data (Monad m, MonadCHP m) => ForkingT m a --- > Control.Concurrent.CHP.Parallel:data ForkingT m a 173d170 < Control.Concurrent.CHP.Parallel:instance (MonadIO m) => MonadIO (ForkingT m) 175c172 < Control.Concurrent.CHP.Parallel:instance MonadTrans ForkingT --- > Control.Concurrent.CHP.Parallel:liftForking :: (Monad m) => m a -> ForkingT m a 239c236 < @version 2.1.0.1 --- > @version 2.2.0 Given previous version number A.B.C.D: It seems you have added, removed or changed some instances. You must increase A or B It seems you have removed something, or changed its name or type. You must increase A or B It seems you have added something. You must increase C (or A or B) ==== As you can see, I've been changing a lot in CHP, and I'm going to need a major version bump (which I've already put in, as you can see: 2.1.0.1 to 2.2.0). It shouldn't be too hard to augment the tool to check the versions against the required bump. I think that covers everything. The script has been tested with GHC 6.10.4 (and tools of a similar vintage) and 6.12.1 (with near-enough the latest versions of most things), but it will probably go wrong if you sneeze at it. In particular, many newer Haskell features (e.g. GADTs) may throw it off, and I'm not sure if it will work if you use it on an OS that's too far from Linux. But hopefully it'll provide a useful stop-gap for some until a real tool is developed, and I think it's at least better than having nothing. Thanks, Neil.

1 0

Halp: wacky FFI problems
by Louis Wasserman 26 Apr '10

26 Apr '10

Hey, I'm trying to do some funky things with the FFI, and I'm clueless enough with C that I could use some help. In particular, I'm trying to bind to the glpk library in a multithreaded style. It's not normally thread-safe, but somebody suggested how to make it safe here <http://www.mail-archive.com/help-glpk@gnu.org/msg04021.html>. I have no idea how to integrate this with FFI-style bindings. A single-threaded set of bindings that I've written is here: http://hackage.haskell.org/package/glpk-hs I'm trying to figure out how to integrate FFI bindings with pthreads, and yadda yadda, and I'm pretty clueless. Any suggestions? Louis Wasserman wasserman.louis(a)gmail.com http://profiles.google.com/wasserman.louis

1 0

Fwd: Mancoosi International Solver Competition
by Johannes Waldmann 26 Apr '10

26 Apr '10

This looks interesting: an international competition of solvers for package/component installation and upgrade problems. http://www.mancoosi.org/misc-2010/ (deadline: June 10) Is the underlying package dependency model the same as cabal's? If (nearly) so, then cabal's solver could enter the competition - if it wins, then fine; if not, then it could be replaced by the winner :-)

1 0

Bulk Synchronous Parallel
by Gregory Crosswhite 26 Apr '10

26 Apr '10

Hey everyone, Has anyone done any work with bulk synchronous parallel computing in Haskell? The idea behind the model is that you divide your computation into a series of computation and communication phases, and it has recently occurred to me that this might be an ideal setup for parallelizing a pure language like Haskell because you can think of it as alternating between a stage that independently applies a bunch of functions to a bunch of independent chunks of data and a stage that applies a big function to all of the chunks that recombines them into new chunks for the next parallel phase, so that all stages are conceptually pure even if eventually the second stage is turned into something involving communication and hence side-effectful under the hood. Experiences? Thoughts? Cheers, Greg

10 13

Order of arguments to GCC when building SDL on Windows with Cabal
by JP Moresmau 26 Apr '10

26 Apr '10

Hello, I've been trying for a while to build the SDL haskell library on windows. People have told me it worked for them, but I can't get to work for the life of me. I have the latest Haskell Platform, I think more or less the latest MINGW and MSYS tools. I'm on Windows 7 64 bits. I installed the SDL library without hassle, and the samples work. It's only the haskell library that's bothering me. I've modified the .cabal file as indicated for Windows (extra-include-dirs, extra-lib-dirs), and runhaskell Setup configure. But then runhaskell Setup build fails with: Preprocessing library SDL-0.5.9... dist\build\Graphics\UI\SDL\General_hsc_make.o:General_hsc_make.c:(.text +0x29): m ultiple definition of `main' d:/dev/mingw/bin/../lib/gcc/mingw32/4.5.0/../../../ libmingw32.a(main.o):main.c:( .text+0x0): first defined here D:\dev\SDL-1.2.14\lib/libSDLmain.a(SDL_win32_main.o): In function `console_main' : /Users/hercules/trunk/SDL-1.2/./src/main/win32/SDL_win32_main.c:315: undefined r eference to `SDL_main' collect2: ld returned 1 exit status linking dist\build\Graphics\UI\SDL\General_hsc_make.o failed command was: d:\dev\MinGW\bin\gcc.exe -LD:\dev\SDL-1.2.14\lib -L/usr/ lib -lmingw 32 -lSDLmain -lSDL -mwindows -LD:\dev\haskell\HaskellPlatform \2010.1.0.0\base-4. 2.0.0 -lwsock32 -luser32 -lshell32 -LD:\dev\haskell\HaskellPlatform \2010.1.0.0\i nteger-gmp-0.2.0.0 -LD:\dev\haskell\HaskellPlatform\2010.1.0.0\ghc- prim-0.2.0.0 -LD:\dev\haskell\HaskellPlatform\2010.1.0.0 -LD:\dev\haskell \HaskellPlatform\201 0.1.0.0/gcc-lib -lm -lwsock32 -LD:\dev\haskell\HaskellPlatform \2010.1.0.0 dist\b uild\Graphics\UI\SDL\General_hsc_make.o -o dist\build\Graphics\UI\SDL \General_hs c_make.exe (I have the same error with the gcc inside the Haskell Platform, the same error running in a windows shell without MSYS on the path, in the MSYS bash with MSYS on the path). And I figured out what's wrong with that line after thinking for a while (as in, several days...): basically the line says something like gcc sdl-flags file.o -o file.exe. If I revert the order of the arguments and run gcc file.o -o file.exe sdl-flags, it works! The gcc documentation explains that the libraries are searched in the order they appear (http://gcc.gnu.org/onlinedocs/gcc/Link-Options.html) so to put the libraries you depend on after the code that has the dependencies. And I can see that building the SDL samples indeed puts the .c file before the sdl linker flags. So this seems to be my problem. What's the solution? 1. I've done something stupid or forgot some essential MINGW tool that would make my system not dependent on the order of the arguments to gcc? I've tried with gcc 3.4.5 and gcc 4.5.0, same result. 2. I can hack somewhere (where???) the command passed on to GCC when building? Maybe some voodoo in the cabal file? 3. I can get runhaskell Setup build to only dump the commands it's going to run, modify them in my favorite text editor and run them manually (as said above, I've managed to create the first exe required but build insists on rebuilding it even if present) 4. ...? The fact that some people managed seem to point with a problem with my config, but it's seems reasonnably clean to me, and I've even managed to install wxWidgets and wxHaskell... Any help appreciated! -- JP Moresmau http://jpmoresmau.blogspot.com/

1 0

ANNOUNCE: graphviz 2999.9.0.0
by Ivan Lazar Miljenovic 26 Apr '10

26 Apr '10

I'm pleased to announce the latest version of my graphviz library that provides Haskell bindings to the Graphviz graph visualisation suite. There are numerous changes in this release, the most important of which are: * graphviz now has an FAQ and an improved README as well as its own homepage: http://projects.haskell.org/graphviz/ (as prompted by Eric Kow). * Add support for record labels; values are automatically escaped/unescaped. The `Record` and `MRecord` shapes have been added for use with these labels. Requested by Minh Thu and Eric Kow. * Add support for HTML-like values (this replaces the wrong and completely broken URL datatype). Strings are automatically escaped/unescaped. * Various parsing improvements (including a slight parsing speed increase!). In particular, graphviz is now able to parse almost all Dot graphs found on my system (including samples shipped with Graphviz, Linux kernel documentation and various other package documentations). A list of the breakages and why: * /usr/share/sgml/docbook/xsl-stylesheets/roundtrip/template.dot seems to be a binary file and thus can't be read. * /usr/share/graphviz/graphs/directed/Latin1.gv uses Latin1 encoding; at the moment graphviz uses the system's locale encoding (or whatever GHC < 6.10 defaults to). * /usr/share/doc/boost-*/html/libs/graph/example/graphviz_test.dot (various boost versions) has subgraphs in edges; graphviz currently can't cope with these. * /usr/src/linux-2.6.33-gentoo-r1/Documentation/blockdev/drbd/drbd-connection-state-overview.dot uses incorrect syntax for the "minlen" attribute (it is meant to be an integer but actually contains a floating point value). The plans for the next release (which I don't plan on even starting for a while) are to focus on improving printing and parsing performance, using a state-based printer and parser (as part of Dot syntax is state-based) and force usage of UTF-8 (via text or utf8-string). -- Ivan Lazar Miljenovic Ivan.Miljenovic(a)gmail.com IvanMiljenovic.wordpress.com

2 1

Why does cabal select base-3.0.3.2 when base-4.2.0.0 is available?
by Bjorn Buckwalter 26 Apr '10

26 Apr '10

Dear all, Why does cabal seem to prefer base-3.0.3.2 over base-4.2.0.0 when installing packages with an unqualified base requirement? Example: $ cabal install -v fad --reinstall [snip] Resolving dependencies... selecting fad-1.0 (hackage) selecting base-3.0.3.2 (installed) and 4.2.0.0 (installed) and discarding syb-0.1.0.0, 0.1.0.1, 0.1.0.2, 0.1.0.3, 0.2 and 0.2.1 [snip] Configuring fad-1.0... Dependency base ==3.0.3.2: using base-3.0.3.2 [snip] [1 of 1] Compiling Numeric.FAD ( Numeric/FAD.hs, dist/build/Numeric/FAD.o ) Numeric/FAD.hs:1:0: Warning: Module `Prelude' is deprecated: You are using the old package `base' version 3.x. Future GHC versions will not support base version 3.x. You should update your code to use the new base version 4.x. [snip] Thanks, Bjorn

3 6

Re: [Haskell-cafe] Broken ghc documentation links
by Ivan Lazar Miljenovic 26 Apr '10

26 Apr '10

"haskell(a)kudling.de" <haskell(a)kudling.de> writes: > While the URL http://www.haskell.org/hoogle/ contains the word "Hoogle", the URL > http://haskell.org/ghc/docs/6.12.1/html/libraries/base-4.2.0.0/Prelude.html… > does not and it does no mention "Hoogle" on the page either. > > How did you determine that the latter is related to Hoogle? Oh, sorry, I misunderstood what you were doing; I thought the actual link you got from Hoogle was wrong (since a lot of people recently have been complaining that Hoogle should now point to the 6.12.2 documentation). In this case, you should talk to GHC maintainers. -- Ivan Lazar Miljenovic Ivan.Miljenovic(a)gmail.com IvanMiljenovic.wordpress.com

1 0

Fwd: Re: [Haskell-cafe] Broken ghc documentation links
by haskell＠kudling.de 26 Apr '10

26 Apr '10

FYI ---------- Ursprüngliche Nachricht ---------- Von: "haskell(a)kudling.de" <haskell(a)kudling.de> An: Ivan Lazar Miljenovic <ivan.miljenovic(a)gmail.com> Cc: Neil Mitchell <ndmitchell(a)gmail.com> Datum: 26. April 2010 um 13:45 Betreff: Re: [Haskell-cafe] Broken ghc documentation links Hi, > So, the problem is that there are broken links _in Hoogle_; have you > thought about contacting the author Neil Mitchell directly? While the URL http://www.haskell.org/hoogle/ contains the word "Hoogle", the URL http://haskell.org/ghc/docs/6.12.1/html/libraries/base-4.2.0.0/Prelude.html… does not and it does no mention "Hoogle" on the page either. How did you determine that the latter is related to Hoogle? Bye, Lnny

1 0

Ada-style ranges
by haskell＠kudling.de 26 Apr '10

26 Apr '10

Hi list, how would you describe Ada's ranges in Haskell's typesystem? http://en.wikibooks.org/wiki/Ada_Programming/Types/range I feel, you can not because this is a runtime property. Bye, Lenny

2 1

Unicode strings and runCommand / runProcess
by John Goerzen 26 Apr '10

26 Apr '10

Here is a very interesting little problem. ghci GHCi, version 6.12.1: http://www.haskell.org/ghc/ :? for help Loading package ghc-prim ... linking ... done. Loading package integer-gmp ... linking ... done. Loading package base ... linking ... done. Prelude> :m System.Process Prelude System.Process> runCommand "echo привет" ?@825B This is a minimal test case for a bug reported in HSH at http://github.com/jgoerzen/hsh/issues#issue/1 It is not entirely clear to me what the behavior here should be. It seems inconsistent with the default behavior of System.IO to, apparently, just strip the bits higher than 0xFF. On the other hand, when it's OS commands we're talking about, it's not entirely clear to me if the default should be to encode in UTF-8. There should almost certainly be an *option* controlling this, and perhaps a version of runProcess that accepts ByteStrings. Thoughts? -- John

4 7

London HUG domain expired
by Bayley, Alistair 26 Apr '10

26 Apr '10

> From: glasgow-haskell-users-bounces(a)haskell.org > [mailto:glasgow-haskell-users-bounces@haskell.org] On Behalf > Of Denys Rtveliashvili > > On the side note, is London HUG still active? The website > seems to be down... Looks like the London HUG domain (londonhug.net) registration has expired. Neil Bartlett was the registrant. Neil: do you plan to renew? Alistair ***************************************************************** Confidentiality Note: The information contained in this message, and any attachments, may contain confidential and/or privileged material. It is intended solely for the person(s) or entity to which it is addressed. Any review, retransmission, dissemination, or taking of any action in reliance upon this information by persons or entities other than the intended recipient(s) is prohibited. If you received this in error, please contact the sender and delete the material from any computer. *****************************************************************

2 2

singleton types
by Günther Schmidt 25 Apr '10

25 Apr '10

Hello, HaskellDB makes extensive use of Singleton Types, both in its original version and the more recent one where it's using HList instead of the legacy implementation. I wonder if it is possible, not considering feasibility for the moment, to implement HaskellDB *without* using Singleton Types. Luke Palmer demonstrates in one blog post replacing type classes with data structures, so I wonder if this could also be applied in this setting. Best regards Günther

3 3

Re: FGL up for adoption
by Louis Wasserman 25 Apr '10

25 Apr '10

I'd be really interested; I have a lot of improvements I'd personally like to make. Louis Wasserman wasserman.louis(a)gmail.com http://profiles.google.com/wasserman.louis

1 0

Announcement: Data.ISO3166_CountryCodes version 0.1
by Jon Fairbairn 25 Apr '10

25 Apr '10

Just uploaded to hackage. Package providing datatype generated from the ISO3166 country code list. This was originally part of my Typeful HTMLs stuff, but I thought it deserved a wider audience. I think anything that uses country codes should use this datatype! As this is my first attempt at using cabal and uploading to hackage, expect errors and omissions. In particular, as I didn't want it to depend on template haskell, the build process generates the module's source file, which means that that file shouldn't be in the tarball. Cabal complains that it's missing if it's not there, so I created an empty .hsc file to shut it up. I'm sure this is not the right way, so I'd be glad to hear what is. -- Jón Fairbairn Jon.Fairbairn(a)cl.cam.ac.uk http://www.chaos.org.uk/~jf/Stuff-I-dont-want.html (updated 2009-01-31)

2 3

I need help getting started
by mitchell＠kaplan2.com 25 Apr '10

25 Apr '10

Hi, I'm just starting to learn, or trying to learn Haskell. I want to write a function to tell me if a number's prime. This is what I've got: f x n y = if n>=y then True else if gcd x n == 1 then f x (n+1) y else False primeQ x = f x 2 y where y = floor(sqrt(x)) The compiler is very unhappy. Apparently there's some ambiguity in this, and I'm kind of running around in circles. I'd appreciate any help that will get me started.

6 11

Threads freezing
by Eitan Goldshtrom 25 Apr '10

25 Apr '10

Hello fellow Haskell programmers, I seem to be having problems with some threads of mine. I wrote an OpenGL program that employs some threads (forkIO) in order to separate any calculations from the OpenGL code. At first it all seemed to be working just fine. Then I added some code for keyboard input to reset the program when you press 'r'. Since then my threads have been giving me problems. Just a second after the program starts all the calculations stop. The program is still running the code to display graphics (I tested this by putting "putStrLn "still displaying" into the display function) but the threads that do all of the computation don't seem to return from their threadDelay, which is only a delay of 5000 microseconds. I tried getting rid of the code for keyboard input but the problem is still there. I tried to run the program on another computer and it ran just fine (without keyboard input). Does anyone have any idea what could be going on? Additionally, I just tried compiling it to an executable and it seems that when I run it from that it works fine, with the keyboard input too. So I'm guessing the problem is with the interpreter? I'm not really sure. Haskell is still pretty mysterious to me. -Eitan

4 3

Re: Confusions about the Haskell Platform (for Mac)
by Gregory Collins 25 Apr '10

25 Apr '10

Yair Chuchem <yairchu(a)gmail.com> writes: > I don't have a good working solution for this, so the situation will > have to remain status quo (i.e. clicking two installers) for now. > > afaik you can make an "mpkg" installer that invokes/contains other > installers: > see http://developer.apple.com/Mac/library/documentation/DeveloperTools/Concept… If you can figure out how to take the binary package the GHC guys make and fold it into a larger .mpkg, please tell me. G. -- Gregory Collins <greg(a)gregorycollins.net>

1 0

Cabal generating text files?
by Michał J Gajda 25 Apr '10

25 Apr '10

Hi, Does anyone know how to generate a text file in a Cabal package? E.g. I want cabal to install a text file into $share_dir/applications/ app.desktop, and fill in certain fragments of this file with paths depending of the $share_dir variable. Say: ==============Yi.desktop template================== [Desktop Entry] Name=Yi Text Editor Comment=Edit text files, and add configuration in Haskell Comment[en_CA]=Edit text files, and configure editor in Haskell Comment[en_GB]=Edit text files, and configure editor in Haskell Comment[pl]=Edytor plików tekstowych z konfiguracją w Haskellu TryExec=$yi_bindir/bin/ yi Exec=$yi_bindir/bin/yi -f %F Terminal=true Type=Application Icon=$yi_sharedir/share/art/yi +lambda.svg Categories=Utility;TextEditor; StartupNotify=false MimeType=text/plain; =============================================== Sorry I couldn't find it in Cabal documentation. -- Best regards Michal

2 1

FGL up for adoption
by Martin Erwig 24 Apr '10

24 Apr '10

Dear FGL users and contributors, The FGL seeks a new home. I don't really have the time anymore to maintain the FGL and add new components at a level that it deserves. I am therefore looking for someone (or a group of people) who would be interested in taking over the maintenance and future development. Anybody interested? Thanks, Martin

3 3

Second Ghent Functional Programming Group Meeting: Date and Call for Talks
by Jeroen Janssen 24 Apr '10

24 Apr '10

(apologies if you receive multiple copies) Dear all, The second Ghent Functional Programming Group (Ghent FPG) meeting takes place on Thursday, May 13 at 19:00 in the Technicum building of Ghent University (Sint-Pietersnieuwstraat 41, 9000 Gent). We have a number of 20-minute timeslots (excluding questions) available for talks, so if you are interested in giving a talk you can e-mail Jeroen Janssen (jejansse(a)gmail.com) The final program will be announced shortly. Hope to see you all then, Bart Coppens (Bart.Coppens(a)elis.ugent.be; Twitter: @BartCopp) Jasper Van der Jeugt (jaspervdj(a)gmail.com; Twitter: @jaspervdj) Jeroen Janssen (jejansse(a)gmail.com; Twitter: @jejansse) Note: if you do not want to receive any more e-mails from Ghent FPG, just send an e-mail to Jeroen Janssen (jejansse(a)gmail.com) and we will remove you from our mailinglist.

1 0

Re: [Haskell-cafe] Using a library in a executable within same cabal package
by Arnaud Bailly 24 Apr '10

24 Apr '10

You are right, I hit "reply" without checking the destination. Thanks for your answer. Arnaud On Sat, Apr 24, 2010 at 3:24 PM, Ivan Lazar Miljenovic <ivan.miljenovic(a)gmail.com> wrote: > First of all, did you mean to reply back to -cafe as well? > > Arnaud Bailly <arnaud.oqube(a)gmail.com> writes: > >> I was not precise enough: Yes, I found the examples to have both a >> library and an executable clear enough that I managed to have >> something workable. But I found somewhat surprisingly that the modules >> used in both the library and the executable were compiled twice. I was >> expecting the executable to link to the dynamic library produced. >> I understand from your answer this is a known issue. > > Right; with Cabal 1.8 this situation is "solved", but at the expense of > not being able to build documentation (I think I was the first person to > fully try this, which is why the bug went unnoticed until recently; the > darcs people are considering using it as well but they didn't test > building haddock docs). > > So yeah, at the moment you're stuck with building the library multiple > times :s > >> >> On Sat, Apr 24, 2010 at 12:14 PM, Ivan Lazar Miljenovic >> <ivan.miljenovic(a)gmail.com> wrote: >>> Arnaud Bailly <abailly(a)oqube.com> writes: >>>> Not sure this question is totally on-topic. >>> >>> Sure is. >>> >>>> Please point me in the right direction if it is not. I would like to >>>> make a cabal package containing both an executable and several >>>> libraries, >>> >>> Each cabal package can contain only one library (which is why gtk2hs is >>> only now starting to be cabalised). >>> >>>> and link the main executable to the libraries produced in >>>> the build. It is not clear how I can achieve that from the >>>> documentation. I tried Extra-libraries option but it does not work. >>> >>> By default, you don't need to do anything extra to have executables >>> using a library in the same package (assuming they have the same source >>> directory) as GHC will find all the modules it needs. >>> >>> Cabal-1.8 has a new feature that lets executables explicitly depend upon >>> the libraries, which means the library doesn't need to get rebuilt >>> several times. However, this is causing problems with building haddock >>> documentation: http://hackage.haskell.org/trac/hackage/ticket/656 >>> >>> -- >>> Ivan Lazar Miljenovic >>> Ivan.Miljenovic(a)gmail.com >>> IvanMiljenovic.wordpress.com >>> > > -- > Ivan Lazar Miljenovic > Ivan.Miljenovic(a)gmail.com > IvanMiljenovic.wordpress.com >

1 0

Compressing GHC tarballs with LZMA
by Leon Smith 24 Apr '10

24 Apr '10

Out of curiousity, I downloaded a binary distribution for GHC-6.12.2, and tried compressing and recompressing it with bzip2 and lzma compression, using no command line arguments (all default parameters) file: http://haskell.org/ghc/dist/6.12.2/ghc-6.12.2-x86_64-unknown-linux-n.tar.bz2 bzip2 lzma size 106M 72M compression time 2:01 9:51 decompression time 0:25 0:12 the time is reported as "minutes:seconds". At it default compression level, lzma was almost five times slower than bzip2, but it's approximately twice as fast when decompressing and produced a file nearly a third smaller. Given that many modern variants of the "tar" command support .tar.lzma files directly, and that both debian's dpkg and red hat's rpm now use lzma, maybe it's time to consider offering .tar.lzma as a download option for GHC's binary (and maybe source-code) bundles. Best, Leon

6 5

Using a library in a executable within same cabal package
by Arnaud Bailly 24 Apr '10

24 Apr '10

Hello, Not sure this question is totally on-topic. Please point me in the right direction if it is not. I would like to make a cabal package containing both an executable and several libraries, and link the main executable to the libraries produced in the build. It is not clear how I can achieve that from the documentation. I tried Extra-libraries option but it does not work. Thanks for your help -- Arnaud Bailly

3 2

Re: [Haskell-cafe] Proper Handling of Exceptional IEEE Floating Point Numbers
by John Lato 23 Apr '10

23 Apr '10

> From: "Barak A. Pearlmutter" <barak(a)cs.nuim.ie> > >> ... An invalid comparison evaluating to _|_ is arguably more >> correct, but I personally find the idea of introducing more bottoms >> rather distasteful. > > Too late! NaN is pretty much the _|_ of IEEE Floating Point. > > That was certainly the intent of the IEEE standard, and is why NaN is > so contagious. But they wanted to relax the usual strictness of their > languages (FORTRAN, C) when this particular _|_ is around. So NaN is > contagious through strict arithmetic (+, *, etc) like _|_. But it has > strange behaviour with respect to comparison. In the context of > Haskell, which does not have the issue of needing to relax strictness > just for NaN, I think the "right thing" would be to have compare give > _|_, and maybe also <, >, ==. After all, NaN is outside the carefully > defined total ordering of all other IEEE floating point values > including +/- Infinity. +1 John

1 0

Re: [Haskell-cafe] lazy-vs-strict, recursive IO (was: memory needed for SAX parsing XML)
by John Lato 23 Apr '10

23 Apr '10

(changed the thread title hoping to get some others to weigh in on the recursive I/O scheme presented here) > From: Daniil Elovkov <daniil.elovkov(a)googlemail.com> > John Lato wrote: >> >>> Another (additional) approach would be to encapsulate unsafeInterleaveIO >>> within some routine and not let it go out into the wild. >>> >>> lazilyDoWithIO :: IO a -> (a -> b) -> IO b >>> >>> It would use unsafeInterleave internally but catch all IO errors within >>> itself. >>> >>> I wonder if this is a reasonable idea? Maybe it's already done? >>> So the topic is shifting... >> >> doWithIO :: NFData b => IO a -> (a -> b) -> IO b >> doWithIO m f = liftM (\a -> let b = f a in b `deepseq` b) m >> >> It works (just stick it in a "try" block for error handling), but you >> need to write a lot of NFData instances. You also need to be careful >> that b is some sort of reduced structure, or you can end up forcing >> the whole file (or other data) into memory. > > I meant a different thing. In your example there is no unsafeInterleave > at all. I think you mean that 'm' argument is supposed to be an > unsafeInterleaved io action, like getContents, and deepseq'ing saves us > from it hanging somewhere for a long time. Ok Yes, that's what I intended. I understand now that it isn't what you meant. > > But I meant to have a routine that lets us use ordinary io actions in a > lazy way, and restricting that 'hanging' within bounds of this routine. > > But having thought about it a little more I understood that it's impossible. I think it's possible provided you have unsafeInterleaveIO and deepseq. As you discovered, to force evaluation through an unsafeInterleaveIO you need to fully evaluate data, and deepseq is the way to do so. > > Lazy io works now because unsafeInterleaveIO is sticked into getContents > itself, and is called repeatedly (via recursion). Or at least I can > think of this implementation, haven't looked into it. > > I realised that calling unsafeInterleaveIO for an ordinary io action > will not make it run lazily. It will, still, run all-at-once, just not > now, but later. Yes, exactly. unsafeInterleaveIO needs to be used recursively for getContents to function properly. > > So, to cope with it, I can think of exposing a little structure of an io > action. Normally it's completely opaque. But if we knew where its > recursion point lies, then we could control its course of execution. > > So, if we had something like > type RecIO a = IO a -> IO a > and io actions were like > getContents :: RecIO [Char] > getContents rec = do > c <- readOneChar > rest <- rec > return (c:rest) > > then we could either run them > normally :: RecIO a -> IO a > normally r = r (normally r) > or > lazily :: RecIO a -> IO a > lazily r = unsafeInterleavIO $ r (lazily r) > > And > lazilyDoWithIO :: RecIO a -> (a -> b) -> IO b > lazilyDoWithIO m f = do > a <- lazily m > return $ f a > > Hmm, but then, we would have to take special care to not let it out of > this function anyway... So here we come to deepSeq'ing you proposed. I think my understanding matches yours at this point, except I'm not sure I follow RecIO. Your getContents, for example, seems to only read one character. I think you mean something like this? recGetContents :: RecIO String recGetContents rec = do c <- getChar ise <- isEOF if ise then return [c] else liftM (c:) rec normally :: RecIO a -> IO a normally = fix lazily :: RecIO a -> IO a lazily rec = let f = unsafeInterleaveIO (rec f) in f This works, but I'm not sure that it accomplishes anything. (it doesn't work in ghci because EOF isn't sent through the terminal, but if you test for another character it works fine). > > And anyway, instead of re-writing pure functions to become iteratees we > will have to re-write io functions to adopt continuation passing style. > > Initially it looked better to me :) Re-writing pure functions to be iteratees doesn't have to be hard; many of the provided iteratees have the same names as standard functions so a lot of the work can be done by just changing imports. > > But with this approach we can run lazily any io action that has the form > of RecIO. Also, we can interleave normally and lazily based on the time > of day and other conditions :) > Originally I didn't see much use to this scheme, but I'm starting to change my mind. It would be nice for users to be able to control the evaluation of functions like getContents, hGetContents, etc., and this looks like it could be easier to maintain. > >> It also doesn't help with >> other IO effects, e.g. writing output. I consider this one of the >> nicest features of iteratee-based processing. > > Can you clarify what's the problem with writing? I think I just haven't > switched from the topic of gluing code. > > Because as for gluing code, type signature for io writing d -> IO () is > perfectly sufficient. The problem isn't with writing output per se, but with sequencing I/O effects with lazy data processing. If you wanted to write to a log file after processing each 100kbytes of a file, for example, doing so with bytestrings would require manually chunking the string, creating output, etc., which then hampers both modularity and readability. Adding this sort of monadic side-effect while data is only partially evaluated is pretty simple with iteratees. Actually, this recursive IO combinator idea you've sketched out would be somewhat helpful in this regard too. John

1 0

Re: [Haskell-cafe] memory needed for SAX parsing XML
by John Lato 23 Apr '10

23 Apr '10

> Message: 8 > Date: Tue, 20 Apr 2010 12:08:36 +0400 > From: Daniil Elovkov <daniil.elovkov(a)googlemail.com> > Subject: Re: [Haskell-cafe] memory needed for SAX parsing XML > To: Haskell-Cafe <haskell-cafe(a)haskell.org> > Message-ID: <4BCD6104.50508(a)googlemail.com> > Content-Type: text/plain; charset=ISO-8859-1; format=flowed > > Jason Dagit wrote: >> >> >> On Mon, Apr 19, 2010 at 3:01 AM, Daniil Elovkov >> <daniil.elovkov(a)googlemail.com <mailto:daniil.elovkov@googlemail.com>> >> wrote: > >> I think iteratees are slowly catching on as an alternative to lazy io. >> Basically, the iteratee approach uses a left fold style to stream the >> data and process it in chunks, including some exception handling. >> Unfortunately, I think it may also require a special sax parser that is >> specifically geared towards iteratee use. Having an iteratee based sax >> parser would make processing large xml streams very convenient in >> haskell. Hint, hint, if you want to write a library :) (Or, maybe it >> exists, I admit that I haven't checked.) > > > Iteratees seem like a natural thing if we want to completely avoid > unsafeInterleaveIO. But the presence of the latter is so good for > modularity. > > We can glue IO code and pure code of the type String -> a so seamlessly. > In case of iteratees, as I understand, pure functions of the type String > -> a would no longer be usable with IO String result. The signature (and > the code itself) would have to be changed to be left fold. To some extent, yes, although the amount of work required can vary. The general rule of thumb is that functions can be used directly with iteratees if they can work strictly. Functions that rely on laziness need some adaptation, although how much varies. If you've built your string handling functions out of a parser combinator library, e.g. parsec-3 or attoparsec, you can just lift the parser into an iteratee and use all your existing functions. Since you're using HaXml, this should work. The only missing part is a polyparse-iteratee converter. I haven't used polyparse, but it looks like the converter would be similar to the one used for attoparsec in the attoparsec-iteratee package. That said, it's not how I would do it. Since SAX is a stream processor, an iteratee-based SAX implementation would be a good fit. I would write a lexer and parser (using a parser combinator library) and then use those with Data.Iteratee.convStream. That's how I would write an iteratee-based SAX parser. HaXml already includes a suitable lexer and parser, but unfortunately they're not exposed. > > Another (additional) approach would be to encapsulate unsafeInterleaveIO > within some routine and not let it go out into the wild. > > lazilyDoWithIO :: IO a -> (a -> b) -> IO b > > It would use unsafeInterleave internally but catch all IO errors within > itself. > > I wonder if this is a reasonable idea? Maybe it's already done? > So the topic is shifting... doWithIO :: NFData b => IO a -> (a -> b) -> IO b doWithIO m f = liftM (\a -> let b = f a in b `deepseq` b) m It works (just stick it in a "try" block for error handling), but you need to write a lot of NFData instances. You also need to be careful that b is some sort of reduced structure, or you can end up forcing the whole file (or other data) into memory. It also doesn't help with other IO effects, e.g. writing output. I consider this one of the nicest features of iteratee-based processing. John

2 1