- Libraries - Haskell.org

Template haskell helper
by Rene de Visser 15 Aug '06

15 Aug '06

Hello, I have hacked together some untested haskell code to do the following z8 = printTH [d| instance TH_Render Char1 where render x = HsLit $ HsChar $ 'a' |] results in something like $(sequence [(instanceD (cxt []) (appT (conT (mkName "TH_Render")) (conT (mkName "Char1"))) [(funD (mkName "render") [(clause [(varP (mkName "x"))] (normalB (infixE (Just (conE (mkName "HsLit"))) (varE (mkName "$")) (Just (infixE (Just (conE (mkName "HsChar"))) (varE (mkName "$")) (Just (litE (CharL 'a'))))))) [])])])]) (minus the "$(sequence )") which you can then be pasted into your program and modified. Hopefully this will people get started with template haskell (and save me a little bit of typing). Rene. begin 666 TH_render.hs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`` ` end begin 666 Base.hs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` end

1 0

Re: compile error (macosx intel), test failure, and few questions
by Chris Kuklewicz 11 Aug '06

11 Aug '06

[ I am copying my reply to the libraries mailing list ] Georgios Siganos wrote: > Hi Chris, > I am checking your library and i have found some problems with the head > version on my system (macosx intel with ghc 6.5) > > First, I get a compile error when I build regex-tre > Wrap.hsc:120:23: error: tre/regex.h: No such file or directory > I can build the remaining modules with no errors. The current README file suggest that you "edit list of BACKENDS in Makefile if you want to exclude regex-tre or regex-pcre" Since you do not have libtre, you can remove it from the BACKENDS near the top of the main Makefile. Theoretically, if you edited regex-tre/regex-tre.cabal to not define -DHAVE_TRE_H then regex-tre would compile and install, but using it would be an error. This is not going to work at the moment since I have not recently tested that part of the source. But it could be handy if you want the module to exist, but not depend on compiling against libtre. > Second, I can successfully run your examples, but when i run > TestTextRegexLazy I get a segmentation fault > ... > "StringPCRE" > Us (backrefs): > Cases: 15 Tried: 15 Errors: 0 Failures: 0 > Us (lazy patterns): > Cases: 31 Tried: 31 Errors: 0 Failures: 0 > Us (possessive patterns): > Cases: 27 Tried: 27 Errors: 0 Failures: 0 > Us (shared): > Cases: 344 Tried: 50 Errors: 0 Failures: 0Segmentation fault > I am not sure how helpful this can be, let me know if you want more > details to find the cause > of the segmentation fault. Note, that the stable version passes all the > tests. Hmmm....I just uploaded (about 10 minutes before your e-mail) version 0.70 to both http://evenmere.org/~chrisk/trl/head/ and http://evenmere.org/~chrisk/trl/stable. There was a pcre segfault that was fixed back in version 0.66 or 0.67. (The FunPtr to pcre_free was null, so it died randomly whenever the garbage collector ran.) > > Regarding the documentation, is there an easy way to produce it, or are > you still working on it? ./setup haddock in each of the regex-* subdirectories will build ./dist/doc/* but I don't have an automatic documentation installer. > > Finally, do you have any suggestion on how I can use sth similar to > splitRegex on a ByteString? I want to split a ByteString using "\n\n" as > a delimiter, and i am not sure what > is the best way to do it. I will eventually expand the compatibility API to provide that. If you needed the split to be lazy, you can copy what regex-compat/Text/Regex/New.hs does: > splitRegex :: Regex -> String -> [String] > splitRegex _ [] = [] > splitRegex delim str = > case matchM delim str of > Nothing -> [str] > Just (firstline, _, remainder, _) -> > if remainder == "" > then firstline : [] : [] > else firstline : splitRegex delim remainder But a more efficient way would be to get all the matches from PCRE in one pass and then return all the intermediate parts: > import Data.ByteString.Char8 > import Text.Regex.Base > import Text.Regex.PCRE > > split' :: Regex -> ByteString -> [ByteString] > split' regex source = > let matches :: [(MatchOffset,MatchLength)] > matches = match regex source > process pos _ | pos `seq` False = undefined > process pos [] = [after pos source] > process pos ((s,o):rest) = extract (pos,s-pos) source : process (s+o) rest > in process 0 matches > *Main> split' (makeRegex "l|d") (pack "Hello World") > ["He","","o Wor","",""] > *Main> split' (makeRegex "l|o.") (pack "Hello World") > ["He","","","W","","d"] > *Main> split' (makeRegex ".o|\?") (pack "Hello World, how are you?") > <interactive>:1:24: lexical error in string/character literal > *Main> split' (makeRegex ".o") (pack "Hello World, how are you?") > ["Hel"," ","rld, ","w are ","u?"] > *Main> split' (makeRegex "l|.o") (pack "Hello World, how are you?") > ["He","","o ","r","d, ","w are ","u?"] > *Main> split' (makeRegex "H") (pack "Hello World, how are you?") > ["","ello World, how are you?"] > *Main> split' (makeRegex "H|h") (pack "Hello World, how are you?") > ["","ello World, ","ow are you?"] > *Main> split' (makeRegex "\n\n") (pack "Hello \nWorld,\n\n how\n are you?") > ["Hello \nWorld,"," how\n are you?"] > Keep on the good work, > Georgos > :)

1 0

ANN: TextRegexLazy 0.70 (stable)
by Chris Kuklewicz 11 Aug '06

11 Aug '06

What: "TextRegexLazy" version 0.70 Where: darcs get --partial http://evenmere.org/~chrisk/trl/stable/ (and darcs get --partial http://evenmere.org/~chrisk/trl/head/ ) The conversion to separate regex-* packages seems stable now. More changes will be done as part of interfacing with GHC. What should I do to advance the plan of changing GHC's regex support? Note that the new Posix library backend is much higher performance than the old one when searching for all the (non-overlapping) matches. The regex-compat package provides Text.Regex.New that uses this better wrapping. The low level API from the Wrap modules always handles errors with Either types. The (Byte)String modules provide a medium level API that handles errors with Either types (compile/execute/regexec). By comparison, the "match" or "=~" high level API still calls error or fail. The build environment will have to be slightly edited when installing. Details are in the README file (and you can and should email me questions). I would appreciate someone with ghc on windows to build as many of the packages as possible and tell me how to do it or how to fix it. GHC includes a copy of the Posix regex library to compile on windows, but regex-posix does not do this (yet). regex-parsec may undergo some small changes to give it a more posix-compatible mode. Right now QuickCheck has found some sub-expression captures that are different for pcre and posix, and parsec agrees with pcre in its left-most strategy but does not always agree with posix in its longest strategy. -- Chris

1 0

additions to Data.Map
by Ross Paterson 11 Aug '06

11 Aug '06

I'd like to propose the following additions to Data.Map (and IntMap): -- | /O(n)/. Map values and collect the 'Just' results. mapMaybe :: Ord k => (a -> Maybe b) -> Map k a -> Map k b -- | /O(n)/. Map keys\/values and collect the 'Just' results. mapMaybeWithKey :: Ord k => (k -> a -> Maybe b) -> Map k a -> Map k b -- | /O(n)/. Map values and separate the 'Left' and 'Right' results. mapEither :: Ord k => (a -> Either b c) -> Map k a -> (Map k b, Map k c) -- | /O(n)/. Map keys\/values and separate the 'Left' and 'Right' results. mapEitherWithKey :: Ord k => (k -> a -> Either b c) -> Map k a -> (Map k b, Map k c)

2 3

Announcing regex-tre-0.66 and benchmarks
by Chris Kuklewicz 10 Aug '06

10 Aug '06

skaller wrote: > On Mon, 2006-08-07 at 20:38 +0100, Chris Kuklewicz wrote: >> skaller wrote: > >>> Wouldn't it be nice to use Ville Laurikari's TRE >>> package instead of PCRE? >>> >>> [It is also Posix compliant and drop in replacement for >>> gnu regex .. as well as supporting nice extensions] >>> >> It is possible to add support for more backends. >> The more the merrier, no need to replace anything. >> I have never heard of TRE before. > > TRE is actually based on mathematics. I have added TRE as a new backend. (I have libtre-0.7.4 on OS X, it is LGPL). And I have a benchmark, which I will now copy and paste. On "b(..?c)?d": PCRE is fast, TRE is almost as fast, the DFA (without subexpression capture) is a close 3rd. The Parsec backend is about 4 times slower, and the Posix library scales poorly, being far slower on large data sets. The code is in darcs http://evenmere.org/~chrisk/trl/head/ under regex-devel/bench Benchmark: find all matches and substring capture for "b(..?c)?d" in a one million character string on disk. Print the count, the first, and the last match (with captures). The benchmark program was, for String: > module Main(main) where > > import Text.Regex.XXX > > filename = "datafile" > regex = "b(..?c)?d" > main = do > input <- readFile filename > let a :: [[String]] > a = input =~ regex > b :: Int > b = length a > print (b,head a,last a) and for ByteString: > module Main(main) where > > import Text.Regex.XXX > import qualified Data.ByteString as B > > default (Int) > > filename = "datafile" > regex = "b(..?c)?d" > > main = do > input <- B.readFile filename > let a :: [[B.ByteString]] > a = input =~ regex > b :: Int > b = length a > print (b,head a,last a) where XXX was replaces by PCRE, Parsec, DFA, PosixRE, TRE and compile with "ghc -O2" Data file is 10^6 characters from permutations of the set "abcdbcdcdd\n" Using the 10^6 character datafile and String or ByteString. (The DFS uses a different semantics so a dot matches a newline, so the matching is different.) The output and user+sys reported by the time command: BenchPCRE (102363,["bcdcd","cdc"],["bbccd","bcc"]) total is 1.294s BenchTRE (102363,["bcdcd","cdc"],["bbccd","bcc"]) total is 2.128s BenchDFA (107811,["bcdcd"],["bbccd"]) total is 2.313s BenchParsec (102363,["bcdcd","cdc"],["bbccd","bcc"]) total is 8.094s BenchPosixRE (102363,["bcdcd","cdc"],["bbccd","bcc"]) total is 91.435s BenchBSPCRE (102363,["bcdcd","cdc"],["bbccd","bcc"]) total is 0.932s BenchBSTRE (102363,["bcdcd","cdc"],["bbccd","bcc"]) total is 1.297s BenchBSDFA (107811,["bcdcd"],["bbccd"]) total is 1.437s BenchBSParsec (102363,["bcdcd","cdc"],["bbccd","bcc"]) total is 8.496s BenchBSPosixRE (102363,["bcdcd","cdc"],["bbccd","bcc"]) total is 89.780 For 10^5 characters on String: PCRE 0.077s DFA 0.131s TRE 0.206s PosixRE 0.445s Parsec 0.825s Old Posix 43.760s (Text.Regex using splitRegex) Old Text.Regex took 43.76 seconds on 10^5 characters to do a task comparable to the one the others did (it ran splitRegex). The new PosixRE wrapping took 0.445 seconds instead. Yes it is two orders of magnitude faster, and this is because my wrapping only marshals the String to CString once. Laziness cannot be worth 2 orders of magnitude of runtime. This is why we needed a new wrapping, which has grown into the new library. On a 10^7 length data set: PCRE (979261,["bcdcd","cdc"],["bd",""]) time 17.388s TRE (979261,["bcdcd","cdc"],["bd",""]) time 17.880s DFA (1063961,["bcdcd"],["bd"]) time 21.617s Parsec (979261,["bcdcd","cdc"],["bd",""]) time 87.330s BenchBSPCRE (979261,["bcdcd","cdc"],["bd",""]) time 8.322s BenchBSTRE (979261,["bcdcd","cdc"],["bd",""]) time 12.644s BenchBSDFA (1063961,["bcdcd"],["bd"]) time 14.115s BenchBSParsec (979261,["bcdcd","cdc"],["bd",""]) time 83.395s

5 6

Cabal vs base
by Neil Mitchell 10 Aug '06

10 Aug '06

Hi, I'm interested in using Cabal to build the base library, just like any other library. Does this have any chance of working? How much work would it take? At the moment the libraries build script has broken on Windows loads of times, and if Cabal was used it would be nicer, less code, and more cross platform. I'd also like to compile libraries for Hat, Yhc and Catch, and if I can reuse cabal to do this that would be great. If this work was done, would it be possible to reduce the installation script for the full haskell.org libraries to get library, cabal config, cabal build, cabal install? So just a general question about what needs doing, who is capable of doing it, and what I can do to get it working. Thanks Neil

4 3

My documentation of System.Posix.Files
by Johan Tibell 10 Aug '06

10 Aug '06

I've started to write some docs for System.Posix.Files but before I finish that and move on to other packages in System.Posix.* it would be great with some feedback on style (and grammar/spelling of course, English is my second language). http://www.itstud.chalmers.se/~larssont/docs/System-Posix-Files.html * Currently most of the POSIX docs are rather incomplete and are in the spirit of "read the man pages". Often a function's documentation states what C function is called and very little beyond that. The man pages contains a wealth of information but: - Everyone doesn't know C (hopefully Haskell is their first language ;) ) - Everyone might not have them - It takes longer time if you have to go and look in the man pages - There might be subtle differences + They are likely to contain a more in depth information than my documentation of the library ever will! Perhaps I could still annotate each function with something like: NOTE: Uses the POSIX system call stat. Or something like that. Perhaps put a mapping of functions to C calls in the module description. * Perhaps I should put a general description of terms (symbolic links, everything is a file, etc.) in the module description? * Lots of error conditions are undocumented. I added some references to a function called throwErrnoIfMinus1_ which lives in System.Posix.Error. It is currently undocumented and frankly it looks like something that "should" be an internal function. Perhaps I could just mention which computations might throw an IOError? Cheers, Johan

2 1

cabal (./Setup.hs) options
by Conrad Parker 09 Aug '06

09 Aug '06

Hi, I like Cabal. There are a few things that I'm used to from autotools which I can't seem to find: dry-runs, uninstall and a "make distcheck" equivalent. Here's my wishlist (ie. stuff I wish someone, possibly me, would hack on ;-) Are the following planned, possible, far-fetched? ./Setup.hs -n, --dry-run For any command, does not actually do anything (build, install, modify the filesystem), but prints the steps that would be executed (ghc commandlines, install locations etc.) uninstall Removes all files and directories created by install sdist-test 1. runs sdist 2. unpacks the resulting tarball 3. runs configure --prefix=_inst build install (into given _inst) test uninstall 4. Checks that no files remain in _inst 5. Cleans up the source tmpdir 6. If all the above succeed, prints a message saying the tarball is ready for distribution cheers, Conrad.

5 6

darcs patch for fgl - Array import-fix
by Esa Ilari Vuokko 09 Aug '06

09 Aug '06

Hi, attached is a patch that allows fgl to compile after the recent Array-changes in base. I don't know if semantics are correct, but atleast it seems to build now. Best regards, --Esa New patches: [Remove now unneccesary hiding Data.Array.IO.indices Esa Ilari Vuokko <ei(a)vuokko.info>**20060809144302] { hunk ./Data/Graph/Inductive/Monad/IOArray.hs 16 -import Data.Array.IO hiding (indices) +import Data.Array.IO } Context: [Silence warnings. Ian Lynagh <igloo(a)earth.li>**20060730164944] [Remove a variable shadowing warning Ian Lynagh <igloo(a)earth.li>**20060730163933] [Version June 2006 Martin Erwig <erwig(a)eecs.oregonstate.edu>**20060718181211] [[project @ 2006-01-18 10:54:19 by simonmar] simonmar**20060118105419 Stop shadowing a variable (darcs patch from Ian Lynagh) ] [TAG Initial conversion from CVS complete John Goerzen <jgoerzen(a)complete.org>**20060112154129] Patch bundle hash: 8642b9c82dd2177887f7dfc84036d669ff94a27b

2 1

System.Posix.Process.executeFile question
by Frederik Eaton 09 Aug '06

09 Aug '06

Hi, I'm wondering what the second argument to 'executeFile' is. It's not clear from the documentation. http://www.haskell.org/ghc/docs/latest/html/libraries/unix/System-Posix-Pro… Also, it seems the documentation source is being parsed incorrectly. This is what it looks like to me: executeFile :: FilePath -> Bool -> [String] -> Maybe [(String, String)] -> IO () literal>executeFile cmd args env</literal calls one of the function>execv*</function family, depending on whether or not the current PATH is to be searched for the command, and whether or not an environment is provided to supersede the process's current environment. The basename (leading directory names suppressed) of the command is passed to function>execv*</function> as <varname>arg[0]</varname; the argument list passed to function>executeFile</function therefore begins with varname>arg[1]</varname Thanks! Frederik -- http://ofb.net/~frederik/

3 3