- Haskell-Cafe - Haskell.org

XMLParser
by Ron de Bruijn 29 Jul '03

29 Jul '03

Hi there, I used the PARSEC parser combinator library to create an following XML-Parser. It's not exactly XML, but it's a usefull format for saving various data. The big difference there is, is that I don't have attributes, but those could easily be added, although I don't need them. Below is the complete code (20! lines of code for the core). I was excited when I saw it also worked like I had imagined, but then I tested it sometimes with increasingly big data. Using my test function below I noticed that there's some exponentional behaviour. I don't see a reason for this. Although I use the try combinator, it should at max double the time, because the parser uses the other parser when there occurs a parse error in the one. My definition of xmlParser looks also a bit odd, but without at least one of the functions of(subEnTop,topEnSub) it will fail parsing. I also don't understand why. Then I question myself: why does my program work with the downloaded version, and not with the PARSEC library that ships with GHC 5.0.4.2, I atleast thought that Parsec hasn't changed for a long time? Does anyone have an explanation for the above problems? Greets Ron module XMLCreator where --this module can parse XML import Parsec import GHC.Show import System.Time --datatype om XML te representeren: XML bestaat altijd uit een topelement en een aantal nestingen data XML a = TopElement String [XML a] | SubElement String a deriving Show type Level = Int --not finished Pretty Printer ppXML::XML a->Level->String ppXML (TopElement name (xml:xmls)) level = concat(take level (repeat " "))++openTag name ppXML _ _ = "hallo" --ppXML (NestedElement name info) level = openTag::String->String openTag s = "<"++s++">" --run :: Show a => Parser a -> String -> IO () run p input= do printTijd case (parse p "" input) of Left err -> do{ putStr "parse error at "; print err } Right x -> do print x printTijd --prints time(used above for printing the start time and the endtime of the algoritm printTijd = do time<-getClockTime (putStrLn.show) time test n = run xmlParser (concat(replicate n str)) --test n = parseTest xmlParser (concat(replicate n str)) str::String str = "<TopicList><TopicItem><Readed>1</Readed><ForumTitle>forumTitle</ForumTitle><TopicTitle>topicTitle</TopicTitle><TopicID>100100</TopicID><TopicStarter>someOne</TopicStarter><LastReplier>someOtherOne</LastReplier><ReplierID>1234</ReplierID><MessageID>messageID</MessageID><SubForumTitle>PW</SubForumTitle><SubForumID>14</SubForumID><Replies>2</Replies><ReplyTime>1050767736</ReplyTime><StartTime>1050762676</StartTime></TopicItem><FileInfo><Version>5</Version><TimeCreated>1059155053</TimeCreated><Application>SomeProgram</Application><About>aboutText</About></FileInfo></TopicList> "; --determine which sign can not be in the elementname or in the informationpiece signs = noneOf ['<','>'] --parses a piece of xml xmlParser::Parser ([XML String]) xmlParser = choice [try (many1 sub),try topEnSub,try subEnTop, many1 top] --xmlParser = choice[try(many1 sub),many1 top] getName::Parser String getName = do teken '<' name<-many1 signs teken '>' return name where teken t = do skipMany space char t skipMany space end::String->Parser String end name = do skipMany space string ("</"++name++">") sub::Parser (XML String) sub = do name<-getName info<-many1 signs end name return (SubElement name info) top::Parser (XML String) top = do name<-getName moreXml<-xmlParser end name return (TopElement name moreXml) subEnTop::Parser [(XML String)] subEnTop = do x<-many1 sub y<-many1 top return (x++y) topEnSub::Parser [(XML String)] topEnSub = do x<-many1 top y<-many1 sub return (x++y) __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com

1 0

Re: Can't compile hmake
by Thomas Davie 29 Jul '03

29 Jul '03

On Tuesday, Jul 29, 2003, at 11:16 Europe/London, Keith Wansbrough wrote: >> I was wondering if idiocy was my problem there, I don't see the >> directory though: >> >> [nimbus2:project/hmake-3.08/targets] tatd100% ls -la >> total 0 >> drwxr-xr-x 3 tatd100 staff 102 Jul 29 10:42 . >> drwxr-xr-x 14 tatd100 staff 476 Jul 29 10:43 .. >> drwxr-xr-x 7 tatd100 staff 238 Jul 29 10:44 powerpc-Darwin6 >> >> Am I being incredibly stupid here? > > 'Fraid so ;-/ `harch` means "powerpc-Darwin6" in your case. The > current host architecture name. I suspected as much. In which case I have the problem that the file doesn't exist :/ [nimbus2:hmake-3.08/targets/powerpc-Darwin6] tatd100% ls -l total 32 -rw-r--r-- 1 tatd100 staff 4 Jul 29 10:43 buildwith -rw-r--r-- 1 tatd100 staff 428 Jul 29 10:54 config.cache -rw-r--r-- 1 tatd100 staff 3 Jul 29 10:43 ghcsym -rw-r--r-- 1 tatd100 staff 721 Jul 29 10:43 hmake3.config drwxr-xr-x 3 tatd100 staff 102 Jul 29 10:44 obj Thanks for putting up with my idiocy, I'm not used to the unix build tools and where it puts things. Tom Davie

1 0

Re: Can't compile hmake
by Thomas Davie 29 Jul '03

29 Jul '03

On Tuesday, Jul 29, 2003, at 11:03 Europe/London, Keith Wansbrough wrote: >>> Yes, after configuring, try editing the generated config file as >>> follows: >>> In targets/`harch`/config, change the line >>> READLINE="" >>> to >>> READLINE="-DUSE_READLINE=1 -L/sw/lib/ -lreadline" >>> then proceed to 'make' as usual. > [..] > >> to configure, then editing the config.cache file - I have tried the >> exact text you specified and also this: >> READLINE="-DUSE_READLINE=1 -L/sw/lib/ -lreadline -I/sw/include/ >> -include readline/readline.h" >> But still with no joy. I'm currently getting this error: > > He said to edit targets/`harch`/config, not config.cache. > config.cache is just information for the next run of ./configure; it > doesn't affect the build process. I was wondering if idiocy was my problem there, I don't see the directory though: [nimbus2:project/hmake-3.08/targets] tatd100% ls -la total 0 drwxr-xr-x 3 tatd100 staff 102 Jul 29 10:42 . drwxr-xr-x 14 tatd100 staff 476 Jul 29 10:43 .. drwxr-xr-x 7 tatd100 staff 238 Jul 29 10:44 powerpc-Darwin6 Am I being incredibly stupid here? Thanks Tom Davie

1 0

Can't compile hmake
by Thomas Davie 29 Jul '03

29 Jul '03

Hi, I'm just trying to get started with haskell and have been having some problems getting hmake to compile. For some reason it doesn't detect libreadline. Lib readline is installed as evidenced by this: [computer:~/Desktop/hmake-3.08] username% fink list | grep readline i readline 4.3-5 Comfortable terminal input library i readline-shlibs 4.3-5 Comfortable terminal input library I've been attempting to force it to link against the library by using ./configure --buildopts="-L/sw/lib/ -lreadline -I/sw/include/ -include readline/readline.h" but with no success, I'm getting the error "gcc: readline/readline.h: No such file or directory", despite the file /sw/includes/readline/readline.h very definitely existing. Has anyone got any ideas for where I could go from here? FYI, I'm running Mac OS X.2.6 Thanks Thomas Davie

2 2

lazy Printing question
by AJ 26 Jul '03

26 Jul '03

Hi everyone, I have written the following program to find "magic numbers" i.e. integers 'n' such that both (n+1) and (n/2+1) are perfect squares. -- Program to find magic numbers Import IO main :: IO () main = do print (filter magicP sqs) sqs :: [Int] sqs = [x*x | x <- [1,3..mAX]] magicP :: Int -> Bool magicP x | ((x-1) `mod` /= 0 = False | otherwise = (((x-1) `div` 2) + 1) `elem` sqs) mAX :: Int mAX = 20000 -- End of listing If I try to run the program (compiled using GHC 6), it calculates all members of the list and then prints the whole list in the end. Since Haskell is 'lazy' I was expecting behaviour similar to HUGS where it prints the numbers as it finds them. Does this behaviour have something to do with the monadic IO in Haskell? I am a Haskell newbie and can't even claim having understood the "gentle introduction" properly. Now my question is - How do I reproduce such "lazy printing" behaviour in GHC? Wait... I thought of another one :-), how do I speed up this program? - AJ

5 6

Type design question
by Konrad Hinsen 25 Jul '03

25 Jul '03

I am a Haskell newbie working on my first serious test case, and I would like some feedback from the experts to make sure I am not doing anything stupid ;-) My applications are numerical (one goal of my current tests being to check how much of a performance penalty I will pay for using Haskell as opposed to C/C++) and involve a lot of geometry. Therefore I started by defining some appropriate types, in particular vectors. So here we go: -- Most general: element of a vector space. -- Operations: addition, subtraction, and multiplication by a scalar class Vect v where (<+>) :: Floating a => v a -> v a -> v a (<->) :: Floating a => v a -> v a -> v a (<*>) :: Floating a => a -> v a -> v a -- Vector space with a scalar product -- Adds dot product and norm class Vect v => VectSProd v where dot :: Floating a => v a -> v a -> a norm :: Floating a => v a -> a norm x = sqrt (dot x x) data Floating a => Vector a = Vector !a !a !a deriving (Eq, Show) -- Standard vector in 3D space instance Vect Vector where (<+>) (Vector x1 y1 z1) (Vector x2 y2 z2) = Vector (x1+x2) (y1+y2) (z1+z2) (<->) (Vector x1 y1 z1) (Vector x2 y2 z2) = Vector (x1-x2) (y1-y2) (z1-z2) (<*>) s (Vector x y z) = Vector (s*x) (s*y) (s*z) instance VectSProd Vector where dot (Vector x1 y1 z1) (Vector x2 y2 z2) = x1*x2+y1*y2+z1*z2 cross :: Floating a => Vector a -> Vector a -> Vector a cross (Vector x1 y1 z1) (Vector x2 y2 z2) = Vector (y1*z2-z1*y2) (z1*x2-x1*z2) (x1*y2-y1*x2) I would like to keep the numerical representation as general as possible, in particular to postpone decisions about precision (Float, Double, eventually arbitrary precision arithmetic) as much as possible. Therefore the "Floating a" constraint. Which leads to my first question: doing it the way I did, I find myself having to add a "Floating a" constraint to almost every function specification with a vector argument. I would prefer to specify once and for all that vector elements need to be "Floating", everywhere. Is there a way of doing that? Next, I would like to define some stuff for simulations of atomic systems. There are two kinds of environments for such simulations, infinite (no boundaries) and periodic (topologically a torus surface). The essential difference is the way that distances are calculated. -- General simulation universe properties class Universe u where distanceVector :: Floating a => u -> (Vector a) -> (Vector a) -> (Vector a) distance :: Floating a => u -> (Vector a) -> (Vector a) -> a distance u v1 v2 = norm (distanceVector u v1 v2) -- Infinite universe data InfiniteUniverse = InfiniteUniverse instance Universe InfiniteUniverse where distanceVector u v1 v2 = v2 <-> v1 -- Periodic universe data Floating a => OrthorhombicUniverse a = OrthorhombicUniverse a a a instance Floating a => Universe (OrthorhombicUniverse a) where distanceVector (OrthorhombicUniverse a b c) (Vector x1 y1 z1) (Vector x2 y2 z2) = Vector (fmod (x2-x1) a) (fmod (y2-y1) b) (fmod (z2-z1) c) where fmod x y = x - y*truncate (x/y) And here I run into a problem. The "Floating a" constraint on OrthorhombicUniverse has the same purpose as for vectors, the three arguments are the dimensions of the box that is periodically replicated. The compiler complains about the definition of distanceVector: ERROR "/home/hinsen/haskell/md/Universe.hs":24 - Inferred type is not general enough *** Expression : distanceVector *** Expected type : (Universe (OrthorhombicUniverse a), Floating b) => OrthorhombicUniverse a -> Vector b -> Vector b -> Vector b *** Inferred type : (Universe (OrthorhombicUniverse a), Floating a) => OrthorhombicUniverse a -> Vector a -> Vector a -> Vector a I do, in fact, want "Floating a" and "Floating b" to enforce the same type, but how can I do that? BTW, in case anyone wonders why I used a class for "Universe" instead of simply defining data Universe = InfiniteUniverse | OrthorhombicUniverse a a a the reason is that I want to leave the option of adding differently shaped universes later on. Konrad.

1 0

RE: segfault in haskell!
by Simon Marlow 25 Jul '03

25 Jul '03

> I've managed to get a segfault in haskell! And without even using the > FFI... actually my code uses the FFI, but the changes that > triggered the > segfault don't involve that, they just use Text.Regex. > > The code that triggers the segfault is the function produced by: > > filetype_function :: IO (FilePath -> FileType) > filetype_function = do > binsfile <- def_prefval "binariesfile" "_darcs/prefs/binaries" > bins <- (liftM lines $ readFile binsfile) `catch` > (\e-> if isDoesNotExistError e then return [] > else ioError e) > let rs = map (mkRegex.fixregex) bins > isbin f = or $ map (\r-> matchRegex r f /= Nothing) rs > ftf f = if isbin f then BinaryFile else TextFile > in > return ftf > > and the valgrind report on the segfault is: > > ==8705== Invalid free() / delete / delete[] > ==8705== at 0x4015D6A4: free (vg_clientfuncs.c:185) > ==8705== by 0x40399A34: (within /lib/libc-2.3.1.so) > ==8705== by 0x40399C13: regfree (in /lib/libc-2.3.1.so) > ==8705== by 0x8117683: s67U_entry (in /home/droundy/darcs/darcs) > ==8705== Address 0x1 is not stack'd, malloc'd or free'd > ==8705== > ==8705== Invalid read of size 4 > ==8705== at 0x40399A58: (within /lib/libc-2.3.1.so) > ==8705== by 0x40399C13: regfree (in /lib/libc-2.3.1.so) > ==8705== by 0x8117683: s67U_entry (in /home/droundy/darcs/darcs) > ==8705== Address 0x22 is not stack'd, malloc'd or free'd I've found one possible cause of this: we were calling regfree on the regular expression structure even if regcomp failed. This will be fixed in GHC 6.0.1. Workaround: don't pass any bogus regular expressions to regcomp. Cheers, Simon

1 0

ENQUIRY
by MR. DAVID LARSON 25 Jul '03

25 Jul '03

FROM: MR. DAVID LARSON ANTHONY EZE & CO. SIR, SUBJECT: TRUST NEED YOUR KIND RESPONSE TO THE FOLLOWING REQUEST: I HAVE A MANDATE TO SOURCE FOR A FINANCIAL MANAGEMENT PERSONNEL ASIDE AFRICA, BE IT YOUR PROFFESSION OR NOT, ITS NATURE DOES NOT COUNT, BUT ONLY HONESTY IS NEEDED AT ITS PEAK AND WANTED AS MOST PRIORTY TO ACTUALIZE THIS EXPECTED GOAL. HAS CLOSE LINK WITH THE PERSONNELS IN THE CORRIDORS OF POWER IN MY COUNTRY HAS BEEN DELEGATED TO SOURCE FOR A COMPETENT ASSOCIATE OUTSIDE AFRICA IF YOU ARE WORTHY OF BEEN TRUSTED WITH HUGE CAPITAL FOR ESTABLISHMENT IN YOUR COUNTRY CONTACT ME FOR FULL PERSONAL SECRETIVE INFORMATION OR DETIALS FOR COMMENCEMENT AND TERMS OF REWARD FOR YOUR EXPERTED INTEREST AND CONTRIBUTION. THE VOLUME OF CAPITAL WILL BE DISCLOSED TO YOU ON MY RECEIPT OF YOUR RETURN CORRESPONDENCE INDICATING YOUR WILLINGNESS TO PER-TAKE. KINDLY RESPOND TO THIS MAIL INCLUDING YOUR PERSONAL CONTACT DATA'S FOR FULL DISCLOSURE OF PROCEEDURES. JUST WAITING FOR YOUR REVERT THANKS MR. DAVID LARSON

1 0

RE: How do I create dynamic exceptions
by Bayley, Alistair 23 Jul '03

23 Jul '03

> From: Ross Paterson [mailto:ross@soi.city.ac.uk] > > On Tue, Jul 22, 2003 at 02:59:12PM +0100, Bayley, Alistair wrote: > > > handler :: Dynamic -> IO () > > > handler e = do > > > case (fromDynamic e) of > > > Nothing -> putStrLn ("dynamic cast failed") > > > Just (MkExc n s) -> putStrLn ("exception: " ++ (show > n) ++ " " ++ s) > > > > > main :: IO () > > > main = catchDyn temp handler > > You're catching a Dynamic wrapped as a Dynamic (catchDyn already does > fromDynamic, as indicated by its type). Change the handler to > > > handler :: MyException -> IO () > > handler (MkExc n s) = putStrLn ("exception: " ++ (show n) > ++ " " ++ s) > > any other dynamic exception fails through to the top level. Thanks. I was applying some advice from Hal Daume: > so, what you want is probably something like: > > > if rc < 0 > > then throwIO (DynException (toDyn ("Err", rc))) > > else ... > > and then when you catch, use fromDyn(amic) to get the value out if you can > predict its type... ... but in this case what I wanted was (just one more layer of indirection): > handler :: Exception -> IO () > handler e = do > case e of > DynException err -> case (fromDynamic err) of > Nothing -> putStrLn "dynamic cast failure" > Just (MkExc n s) -> putStrLn ("exception: " ++ (show n) ++ " " ++ s) > _ -> putStrLn "unhandled exception" > main :: IO () > main = catch temp handler -- use catch rather than catchDyn ***************************************************************** The information in this email and in any attachments is confidential and intended solely for the attention and use of the named addressee(s). This information may be subject to legal professional or other privilege or may otherwise be protected by work product immunity or other legal rules. It must not be disclosed to any person without our authority. If you are not the intended recipient, or a person responsible for delivering it to the intended recipient, you are not authorised to and must not disclose, copy, distribute, or retain this message or any part of it. *****************************************************************

1 0

RE: instance Typeable (Re: How do I create dynamic exceptions)
by Simon Peyton-Jones 22 Jul '03

22 Jul '03

| The above should work, but I wonder when it is possible to simply derive | Typeable? GHC 6.0 supports deriving( Typeable ), but for some strange reason it's not in the manual. It will be. Simon

2 1