September 2006 - Libraries

ANNOUNCE: Codec.Compression.GZip & .BZip
by Duncan Coutts 21 Sep '06

21 Sep '06

I'm pleased to announce two new packages: zlib and bzlib which provide functions for compression and decompression in the gzip and bzip2 formats: Both provide pure functions on streams of data represented by lazy ByteStrings: compress, decompress :: ByteString -> ByteString This makes it easy to use either in memory or with disk or network IO. For example a simple gzip compression program is just: > import qualified Data.ByteString.Lazy as ByteString > import qualified Codec.Compression.GZip as GZip > > main = ByteString.interact GZip.compress Or you could lazily read in and decompress @.gz@ file using: > content <- fmap GZip.decompress (readFile file) The code is available via darcs: darcs get http://haskell.org/~duncan/zlib darcs get http://haskell.org/~duncan/bzlib (Note that if you are using GHC-6.5 then you'll need to edit the .cabal file to remove the dependency on fps, since ByteStrings are provided in the base package in GHC-6.5.) There is API documentation too: http://haskell.org/~duncan/zlib/docs http://haskell.org/~duncan/bzlib/docs Both packages are bindings to the corresponding C libs, so they depend on those C libraries. Fortunately both zlib and bzlib2 are available on every OS. It also means that the compression speed is as you would expect since it's the C lib that is doing all the work. I'm very happy to get feedback on the API, the documentation or of course any bug reports. Duncan

2 2

Patch: Allow testsuite to run under MSYS/MinGW using native Python (not Cygwin Python)
by Brian Smith 16 Sep '06

16 Sep '06

This is based on a patch by *Esa Ilari Vuokko*. I had a lot of difficulty getting Cygwin set up so I didn't test the changes under Cygwin, but I tested them under Ubuntu and MSYS/MinGW. - Brian

1 0

Patch to fix haddock syntax error in GHC.Handle (Windows only)
by Brian Smith 16 Sep '06

16 Sep '06

Without this very patch, the Haddock docs for the libraries cannot be built on Windows. Tested in MSYS/MinGW current on Windows XP.

1 0

please try this tiny script to test cabal-install
by Isaac Jones 15 Sep '06

15 Sep '06

Paolo has been hacking on cabal-install, as many of you know. Cabal-install is a tool for installing cabal packages and their dependencies. We put together this little script that _should_ just install cabal-install, along with testing it out to install hnop. I know this works on my Debian box, your results may vary, especially if you don't use sudo, but these commands should be pretty clear for you to try them yourself. Please copy and paste this into a terminal, or make minor tweaks and do the same. Email cabal-devel to let us know if it worked or if it didn't! wget http://hackage.haskell.org/~paolo/darcs/cabal-install/test-cabal-install && chmod 755 test-cabal-install && ./test-cabal-install (Now that you've got that working, you can "cabal-install list" and see the packages on the server.) peace, isaac p.s. of course, if you don't trust hackage or paolo, you should look at the code before you run it as root ;) p.p.s. the script itself, in case you want to copy, mod, and paste is: #!/bin/sh darcs get http://hackage.haskell.org/~paolo/darcs/cabal-install/ && \ cd cabal-install && \ make && \ sudo make install && \ sudo cabal-install update && \ cabal-install install --prefix=/tmp --user hnop && \ ls -l /tmp/bin/hnop

1 0

Build failures in packages built with GHC 6.6
by Duncan Coutts 14 Sep '06

14 Sep '06

Hi folks, Over the weekend Lennart Kolmodin tested all of Gentoo's Haskell packages with the latest GHC 6.6 RC snapshot. Here is his report of what failed, and how: http://www.haskell.org/~gentoo/gentoo-haskell/projects/GHC-6.5-failures.html You'll note that there are some common ones, like Data.FiniteMap going away and the old API in Data.Set disappearing. There seemed to be several programs which used the old lang or text packages. The change in the array bounds api broke some things. Text.Regex moved into a different package. Most of these look fairly trivial. Class instance changes broke a couple packages, which might be a bit more tricky to fix. One Cabal api change broke one package, again easily fixable. We made some Cabal modules that had been private by convention properly private (ei Distribution.Compat.*). This broke lhs2tex. I'm not sure exactly what Cabal scripts should be using for portable file path handling since Setup.hs modules are not supposed to have external dependencies on other packages (otherwise there's a chicken and egg problem with compiling Setup.hs). Advice here would be apreciated, I could let Distribution.Compat.FilePath remain public in this release and seek a better solution for the next release. BTW, you can see the complete set of Gentoo's Haskell packages here: http://www.haskell.org/~gentoo/gentoo-haskell/ (mostly under dev-haskell/) Gentoo users & developers can use this package collection as a portage overlay. Duncan (wearing his Cabal release manager hat and his Gentoo hat)

4 5

ANN: hsgnutls 0.2.3
by Einar Karttunen 13 Sep '06

13 Sep '06

Hello Hsgnutls version 0.2.3 is released. This is mainly a bug fix release supporting new versions of both GHC and GnuTLS. * Support for GHC 6.6 * Support GnuTLS 1.0-1.4 * Many bug fixes * tlsCheckPending, getPeersX509DN http://www.cs.helsinki.fi/u/ekarttun/hsgnutls/ - Einar Karttunen

1 0

3rd Cabal-1.1.6 release candiate
by Duncan Coutts 12 Sep '06

12 Sep '06

Hi all, I've put up a tarball for the third 1.16 Cabal release candidate: http://www.haskell.org/cabal/release/cabal-1.1.5.9.3/cabal-1.1.5.9.3.tar.gz This is also included with very recent GHC 6.5 snapshots. Please test this with your packages and report any issues to cabal-devel(a)haskell.org We're particularly interested of course in anything that worked before and is now broken. Duncan

1 0

darcs patch: Added Data.Map.insertWith'
by John Goerzen 10 Sep '06

10 Sep '06

Tue Sep 5 13:11:49 CDT 2006 John Goerzen <jgoerzen(a)complete.org> * Added Data.Map.insertWith' Use of insertWith can lead to memory leaks. This function was suggested by Udo Stenzel at http://article.gmane.org/gmane.comp.lang.haskell.cafe/14864 This type of algorithm solved a real problem discussed in this thread: http://thread.gmane.org/gmane.comp.lang.haskell.cafe/14852

3 3

darcs patch: Fix building problem for Mac OS X
by shelarcy 06 Sep '06

06 Sep '06

Hi all, I made OpenAL patch almost one year ago. http://www.haskell.org/pipermail/libraries/2005-September/004399.html Sven merged that ... but He accidentally removed Mac OS X part. I posted again and again when someone can't build under Mac OS X or OpenAL was changed newer version. But His patch check priority is lowest .... and it seems that He fade out development of OpenAL and OpenGL packages current a few months. Now, GHC 6.6 comes and this version bumps OpenAL version, it's good time to apply my patch ... but I don't know how to apply that. (Because I don't have darcs.haskell.org account.) Does anyone apply this patch? I already tested this patch under Windws and other Unix platform. http://www.haskell.org/pipermail/libraries/2006-May/005405.html Best Regards, New patches: [Fix building problem for Mac OS X shelarcy <shelarcy(a)capella.freemail.ne.jp>**20060907023009] { hunk ./aclocal.m4 171 + +# FP_CHECK_OPENAL +# ------------- +AC_DEFUN([FP_CHECK_OPENAL], +[AC_REQUIRE([AC_CANONICAL_TARGET]) +AL_CFLAGS= +case $target_os in +darwin*) + AL_LIBS= + AL_FRAMEWORKS=OpenAL + ;; +*) + AC_SEARCH_LIBS([alGenSources], [openal openal32], [AL_LIBS="$ac_cv_search_alGenSources"]) + test x"$AL_LIBS" = x"none required" && AL_LIBS= + AC_SUBST([AL_LIBS]) + AL_FRAMEWORKS= + ;; +esac +AC_SUBST([AL_CFLAGS]) +AC_SUBST([AL_FRAMEWORKS]) +])# FP_CHECK_OPENAL + hunk ./configure.ac 11 -# Shall we build this package at all? -FP_ARG_OPENAL hunk ./configure.ac 13 -if test x"$enable_openal" = xyes; then - hunk ./configure.ac 15 -# The following test for the OpenAL library is a bit of a hack, but gives nice -# feedback even for the case when no additional library is needed. (Mac OS X?) -AL_LIBS=no -AC_SEARCH_LIBS([alGenSources], [openal openal32], [AL_LIBS="$ac_cv_search_alGenSources"]) -test x"$AL_LIBS" = x"none required" && AL_LIBS= -AC_SUBST([AL_LIBS]) +# Shall we build this package at all? +FP_ARG_OPENAL + +if test x"$enable_openal" = xyes; then + +# Check for ALUT include paths and libraries +FP_CHECK_OPENAL hunk ./configure.ac 27 -# Ugly... -AL_FRAMEWORKS= -case $host_os in -darwin*) - AL_FRAMEWORKS=OpenAL - ;; -esac -AC_SUBST([AL_FRAMEWORKS]) - +# check for ALUT include files hunk ./configure.ac 30 -if test -z "$AL_FRAMEWORKS" && test x"$fp_found_al_header$fp_found_alc_header" != xyesyes; then + +if test x"$fp_found_al_header" = xno || test x"$fp_found_alc_header" = xno; then hunk ./configure.ac 185 -*-mingw32) CALLCONV=stdcall ;; +# WinDoze DLL hell +# NOTE: For some obscure/unknown reason, OpenAL uses ccall, even on WinDoze... +# *-mingw32) CALLCONV=stdcall ;; } Context: [includes -> install-includes Ross Paterson <ross(a)soi.city.ac.uk>**20060829123744] [exclude Setup.hs even if not building package Ross Paterson <ross(a)soi.city.ac.uk>**20060825222702] [exclude Setup.hs from build Ross Paterson <ross(a)soi.city.ac.uk>**20060824183534] [add boilerplate Setup.hs Ross Paterson <ross(a)soi.city.ac.uk>**20060824115102] [add files used by configure Ross Paterson <ross(a)soi.city.ac.uk>**20060518174356] [Fixed bug #715 (OpenAL package fails to compile with older headers) Sven Panne <sven.panne(a)aedion.de>**20060507164755] [TAG Initial conversion from CVS complete John Goerzen <jgoerzen(a)complete.org>**20060112154136] Patch bundle hash: f96adc3351c29eb8bdee0deeebd721ee8c0812a1 -- shelarcy <shelarcy capella.freemail.ne.jp> http://page.freett.com/shelarcy/

1 0

Disclipline for re-use: Slim instances
by kahl＠cas.mcmaster.ca 06 Sep '06

06 Sep '06

Although the trigger for this comes from a thread in glasgow-haskell-users, I think that for the general problem, the libraries list is the right forum. Of course, the prelude and standard libraries aspect of haskell-prime should be influenced as well, but I am hesitant to cross-post. Serge D. Mechveliani <mechvel(a)botik.ru> wrote: > > I need to print in a special way the data of > [Equation], (Term, Term), [(Term, Term)], (Equation, Equation). > > The first can be by defining showList in instance Show Equation. > But Show has not a method of showPair. So, I need to write the > function showsTermPair and to use it together with another home-made > function showsListGeneric. I have been bitten by this, too, so let me make the point more explicit: For lists, the Haskell98 report contains: | class Show a where | showsPrec :: Int -> a -> ShowS | show :: a -> String | showList :: [a] -> ShowS | | -- Mimimal complete definition: | -- show or showsPrec | showsPrec _ x s = show x ++ s | | show x = showsPrec 0 x "" | | showList [] = showString "[]" | showList (x:xs) = showChar '[' . shows x . showl xs | where showl [] = showChar ']' | showl (x:xs) = showChar ',' . shows x . | showl xs The proposed function showsListGeneric would of course be: > showsListGeneric shows [] = showString "[]" > showsListGeneric shows (x:xs) = showChar '[' . shows x . showl xs > where showl [] = showChar ']' > showl (x:xs) = showChar ',' . shows x . > showl xs If this was defined in the prelude, then the default definition in the class declaration would become: > class Show a where > -- [...] > > showList = showsListGeneric shows Similarly, for pairs the prelude has: | instance (Show a, Show b) => Show (a,b) where | showsPrec p (x,y) = showChar '(' . shows x . showChar ',' . | shows y . showChar ')' The suggested function showPair is of course: > showPair shows1 shows2 (x,y) = showChar '(' . shows1 x . showChar ',' . > shows2 y . showChar ')' Again, if this was defined in the prelude, the instance declaration would become: > instance (Show a, Show b) => Show (a,b) where > showsPrec p (x,y) = showPair shows shows The problem we encounter here is that instance declarations and default definitions contain unnamed functions, which however frequently are also useful in other contexts. Since it does not appear to be attractive or easy to introduce (into the language definition) a naming scheme for these functions, I propose to introduce the following discipline for all library developers: ********************************************* * All instance declarations should be slim. * ********************************************* I.e.: * Instance declarations (and default definitions) should never contain non-trivial function definitions. * Instance declarations (and default definitions) should only provide ``plumbing'' to make existing functions accessible via the type class resolution mechanism. * The ``plumbed'' functions should always be exported (since instances are always exported). Since the naming will not always be straight-forward, the last point is particularly important and would enable more re-use and less re-invention of mostly trivial wheels. I append the ``TextFunctors'' module which I mostly pulled out of GHC's internals; the commented-out part of the export list shows where these belong in the code base of GHC; in my opinion, they should of course also be in the standard libraries. Wolfram ======= TextFunctors.lhs ================================================ \section{Read and Show Functors} \begin{code} module TextFunctors ( module TextFunctors -- -- the exported functions should really be imported -- -- from the following locations: -- , GHC.Show.showsPrecMaybe -- , GHC.Show.showsPrecEither -- , GHC.Show.showsPrecTup2 -- , GHC.Show.showsPrecTup3 -- , GHC.Show.showsPrecTup4 -- , GHC.Show.showsPrecTup5 -- , GHC.Read.readPrecMaybe -- , GHC.Read.readPrecEither -- , GHC.Read.readPrecArray -- , GHC.Read.readPrecTup2 -- , GHC.Read.readPrecTup3 -- , GHC.Read.readPrecTup4 -- , GHC.Read.readPrecTup5 , GHC.Read.Read(..) ) where import qualified GHC.Show import qualified GHC.Read -- the following imports are necessary for the read functions: import Text.ParserCombinators.ReadPrec import Array import qualified Text.Read.Lex as L \end{code} %{{{ \subsection{Show} \subsection{Show} \begin{code} type ShowSPrec a = Int -> a -> ShowS showsPrecList :: ShowSPrec a -> ShowSPrec [a] showsPrecList showsPrec p = GHC.Show.showList__ (showsPrec 0) \end{code} \begin{code} showsPrecMaybe :: ShowSPrec a -> ShowSPrec (Maybe a) showsPrecMaybe _showsPrec _p Nothing s = showString "Nothing" s showsPrecMaybe showsPrec p (Just x) s = (showParen (p > appPrec) $ showString "Just " . showsPrec appPrec1 x) s \end{code} \begin{code} showsPrecEither :: ShowSPrec a -> ShowSPrec b -> ShowSPrec (Either a b) showsPrecEither showsPrecA showsPrecB p e s = (showParen (p > appPrec) $ case e of Left a -> showString "Left " . showsPrecA appPrec1 a Right b -> showString "Right " . showsPrecB appPrec1 b ) s \end{code} \begin{code} showsPrecTup2 showsPrecA showsPrecB _ (x,y) s = (showChar '(' . showsPrecA noPrec x . showChar ',' . showChar ' ' . showsPrecB noPrec y . showChar ')' ) s \end{code} \begin{code} showsPrecTup3 showsPrecA showsPrecB showsPrecC _ (x,y,z) s = (showChar '(' . showsPrecA noPrec x . showChar ',' . showChar ' ' . showsPrecB noPrec y . showChar ',' . showChar ' ' . showsPrecC noPrec z . showChar ')' ) s \end{code} \begin{code} showsPrecTup4 showsPrecA showsPrecB showsPrecC showsPrecD _ (x,y,z,u) s = (showChar '(' . showsPrecA noPrec x . showChar ',' . showChar ' ' . showsPrecB noPrec y . showChar ',' . showChar ' ' . showsPrecC noPrec z . showChar ',' . showChar ' ' . showsPrecD noPrec u . showChar ')' ) s \end{code} \begin{code} showsPrecTup5 showsPrecA showsPrecB showsPrecC showsPrecD showsPrecE _ (x,y,z,u,v) s = (showChar '(' . showsPrecA noPrec x . showChar ',' . showChar ' ' . showsPrecB noPrec y . showChar ',' . showChar ' ' . showsPrecC noPrec z . showChar ',' . showChar ' ' . showsPrecD noPrec u . showChar ',' . showChar ' ' . showsPrecE noPrec v . showChar ')' ) s \end{code} \begin{code} noPrec :: Int noPrec = 0 appPrec = GHC.Show.appPrec appPrec1 = GHC.Show.appPrec1 \end{code} %}}} %{{{ \subsection{Read} \subsection{Read} \begin{code} readPrecList = GHC.Read.list \end{code} \begin{code} readPrecMaybe :: ReadPrec a -> ReadPrec (Maybe a) readPrecMaybe readPrec = parens (do L.Ident "Nothing" <- lexP return Nothing +++ prec appPrec ( do L.Ident "Just" <- lexP x <- step readPrec return (Just x)) ) \end{code} \begin{code} readPrecEither :: ReadPrec a -> ReadPrec b -> ReadPrec (Either a b) readPrecEither readPrecA readPrecB = parens ( prec appPrec ( do L.Ident "Left" <- lexP x <- step readPrecA return (Left x) +++ do L.Ident "Right" <- lexP y <- step readPrecB return (Right y) ) ) \end{code} \begin{code} readPrecArray :: Ix i => ReadPrec i -> ReadPrec a -> ReadPrec (Array i a) readPrecArray readPrecI readPrecA = parens $ prec appPrec $ do L.Ident "array" <- lexP bounds <- step (readPrecTup2 readPrecI readPrecI) vals <- step (GHC.Read.list (readPrecTup2 readPrecI readPrecA)) return (array bounds vals) \end{code} \begin{code} readPrecTup2 readPrecA readPrecB = parens ( paren ( do x <- readPrecA L.Punc "," <- lexP y <- readPrecB return (x,y) ) ) \end{code} \begin{code} readPrecTup3 readPrecA readPrecB readPrecC = parens ( paren ( do x <- readPrecA L.Punc "," <- lexP y <- readPrecB L.Punc "," <- lexP z <- readPrecC return (x,y,z) ) ) \end{code} \begin{code} readPrecTup4 readPrecA readPrecB readPrecC readPrecD = parens ( paren ( do w <- readPrecA L.Punc "," <- lexP x <- readPrecB L.Punc "," <- lexP y <- readPrecC L.Punc "," <- lexP z <- readPrecD return (w,x,y,z) ) ) \end{code} \begin{code} readPrecTup5 readPrecA readPrecB readPrecC readPrecD readPrecE = parens ( paren ( do v <- readPrecA L.Punc "," <- lexP w <- readPrecB L.Punc "," <- lexP x <- readPrecC L.Punc "," <- lexP y <- readPrecD L.Punc "," <- lexP z <- readPrecE return (v,w,x,y,z) ) ) \end{code} \begin{code} paren = GHC.Read.paren parens = GHC.Read.parens lexP = GHC.Read.lexP \end{code} %}}} %{{{ EMACS lv % Local Variables: % folded-file: t % fold-internal-margins: 0 % eval: (fold-set-marks "%{{{ " "%}}}") % eval: (fold-whole-buffer) % end: %}}}

2 1