- Haskell - Haskell.org

New monadic implementation of Control.Exception
by Brian Hulley 25 Apr '06

25 Apr '06

Hi - Following indications in http://www.haskell.org/pipermail/haskell/2006-February/017547.html I've wrapped all the functions in Control.Exception so that they can be used with any monad constructed using the common state transformers and based (ultimately) on IO. I'd suggest something like this could be incorporated into Haskell' perhaps as a replacement for Control.Exception. This also relates to http://hackage.haskell.org/trac/haskell-prime/ticket/110 but I don't know how to attach a file to a ticket therefore I've just enclosed it with this email. (In the discussion following Oleg's original post my experience implementing the generalised exception functions showed that getUnliftIO :: m(m a -> IO a) mentioned in a later post proved to be a red herring since StateT can't support this operation yet it *can* support all the exception ops in a sensible way.) So far, I've only tested catch, catchJust, throwM using a (StateT s IO) monad, so I can't be absolutely sure that I've got the semantics for the other monads correct, but at least they all compile.... :-) Regards, Brian.

1 1

Parallel list comprahensions
by Iavor Diatchki 25 Apr '06

25 Apr '06

Hello, Does the parallel list comprahension notation support guards that involve both variables that are being generated? I was trying to write something that is essentially: [ f x y | (x,y) <- zip xs ys, p x y ] and I thought that it would be nice to rewrite it as a parallel list comprahension, but I am stuck on the `p x y` part. I can define my own function to do that: pzip f p xs ys = concat (zipWith zipper xs ys) where zipper x y = if p x y then [f x y] else [] or use the normal list comprahension version so it is not a problem, but I thought I'd check if there already was a notation that I didn't know about. -Iavor

1 0

Haskell Summer School (Camp) for Teens
by John Peterson 24 Apr '06

24 Apr '06

The Computer Science department at Western State College is offering a very unique summer camp opportunity for high school Juniors / Seniors. Although our brochure and web page never mention Haskell, we're using Haskell on top of the Panda 3D game engine to allow students with no computer experience to do interesting things with computing, math, and physics. In addition, our program includes some recreational opportunities that those boring summer schools can't compete with! :-) If you know anyone that would be interested in this, please pass the message on. Our brochure is here: http://www.western.edu/computerscience/computer_camp.html And you don't have to tell them that this is a secret Haskell Indoctrination Camp. John Peterson Now at jpeterson(a)western.edu

1 0

Haskell.org and Google Summer of Code 2006
by Paolo Martini 24 Apr '06

24 Apr '06

Hello everybody, nice to meet you all. Last year I did work on an Haskell project during the first year of the Google Summer of Code programme[1]. I wrote the bindings to the Cairo Vector Graphics library and integrated them in Gtk2Hs. You can see last year's announcement on the website[2]. I had to take Google itself as my mentoring organization, in absence of a better-suited one, no wonder I was the only one with an Haskell project, eventually I got one of its 13 slots available -- Chris DiBona[3], the project leader, seem to like Haskell very much! They recently announced they would be running the programme again this year, and I thought we could surely do better! My idea is that the best way we are going to get more Haskell projects done is having Haskell.org as a mentoring organization this year. Let me briefly summarize how the project works; it starts with a number of mentoring organizations that support active open source projects. They are required to publish a list of projects for students to apply for, and some mentors which will take care of the pupils during their work. Part of the fun is of course that the students can propose their own projects too. Each successful applicant gets an initial $500, three months of coding time, a mid- term $2,000 of are doing well and a final $2,000 on successful completion of the project (The site[4] does explain it all.) I contacted many of the people who usually reside in the #haskell IRC channel asking for collaboration, and they promptly accepted, forming a quite big number of well organized people -- as now, those three people are in charge for the administrative work: - Isaac Jones <http://www.syntaxpolice.org/>, - Donald Bruce Steward <http://www.cse.unsw.edu.au/~dons/>, - David Himmelstrup <http://darcs.haskell.org/~lemmih/aboutMe.html>. Not surprisingly, since the people of the channel are known to be very kind and helpful, answering questions at every time of the day, I've got a good number of volunteers for mentoring too (which I'm very happy about): - David Himmelstrup, - John Goerzen <http://www.complete.org/>, - Cale Gibbard <http://www.haskell.org/hawiki/CaleGibbard>, - Einar Karttunen <http://www.cs.helsinki.fi/u/ekarttun/>, - Shae Matijs Erisson <http://www.ScannedInAvian.com/>, - Duncan Coutts <http://haskell.org/gtk2hs/development/>. - Andres L\¨oh <http://www.iai.uni-bonn.de/~loeh>, - Ian Lynagh <http://web.comlab.ox.ac.uk/oucl/work/ian.lynagh/>, And the list is getting bigger and bigger! Unsure about where to get an "authoritative answer" about things like this I have mailed Simon Peyton-Jones, he likes the idea (Thanks very much!) He explained to me that since the Haskell community is very decentralized I would better send a message to the Haskell mailing list to propose this project. I hope you will be as enthusisatic as I am. Hacking Haskell code during the summer, funded enough to keep up with the other things (and getting some pretty cool hardware lately), was a *great* experience. I sincerely hope that many students will apply for Haskell works. We need projects done! I'll be mailing you the url of the wikipage collecting projects and details. Thank you very much, Paolo. -- [1] <http://code.google.com/summerofcode.html> [2] <http://haskell.org/gtk2hs/?p=30> [3] <http://dibona.com>

4 6

Example Code Wiki area
by Paul Johnson 23 Apr '06

23 Apr '06

I've created a new area for Example Code on the Wiki. I'd like to call on all those who contributed code to the old "CodeOnTheWiki" area to bring it over to the new area. However you should be aware that doing so implicitly licenses the code under the Simple Permissive License. Paul.

1 0

How to avoid both duplicate instances and extraneous instance declarations?
by Robin Bate Boerop 22 Apr '06

22 Apr '06

I am having trouble making a type system for my EDSL in Haskell. Can someone help? A simplified description of my problem follows. I have three classes, A, B1, and B2. I want all instances of B1 to be instances of A. I want all instances of B2 to be instances of A. None of the classes have methods. The following does NOT work, because of a duplicate instance declaration for A: class A a class B1 b1 class B2 b2 instance B1 x => A x instance B2 x => A x -- duplicate instance, won't compile data T = T instance B1 T The following DOES work, but it requires that I explicitly give instance declarations of A for all instances of B1 and B2: class A a class A a => B1 a class A a => B2 a data T = T instance A T -- I don't want to have to specify this! instance B1 T Of course, this is a simplified example. In a complicated type class hierarchy, in a program with many instances of, say B1, having to provide instance declarations for A is laborious. Is there a way to avoid this? How can I tell Haskell "all instances of B1 are automatically instances of A". -- Robin Bate Boerop

4 10

Haskell' status & summary
by Isaac Jones 22 Apr '06

22 Apr '06

Greetings! I'll try to update the Haskell community periodically on the status of the Haskell' language standard. As mentioned previously, we are currently focusing on two topics, concurrency and the class system. If you feel that you have anything important to contribute to those topics, now is the time to review the proposals, join in the Haskell' mailing list and let us know what you know! Stephanie Weirich has posted a summary of the class system discussion here: http://hackage.haskell.org/trac/haskell-prime/wiki/ClassSystem Stephanie says: This page is important because it lists all of the proposals not related to MPTCs as well as trying to capture the big picture about where we stand with respect to MPTCs. I've been trying to not duplicate text that appears elsewhere in the wiki, but just provide a consistent picture of the state of the discussions on the mailing list. Please take a look at this page and help me fill it out. In particular, 've been trying to take a pulse of where we stand on some of these issues, and some of you may not agree! Tell me if I'm off the wall. Also, I've (mostly) concentrated on issues that have tickets, so I may have missed some issues that were only discussed in the mailing list. Please let me know if there is anything I've forgotten. Simon Marlow has posted a summary of the concurrency discussion here: http://hackage.haskell.org/trac/haskell-prime/wiki/Concurrency Simon says: I have tried to summarise the various points that have arisen during the discussion. If anyone feels they have been mis-paraphrased, or I have forgotten something, please feel free to edit, or send me some text for inclusion. I don't want to include long gobs of text in here, though: just summarise the main points, and if necessary link to relevant mailing list posts. Thanks, Simon & Stephanie for keeping things moving forward! peace, isaac

1 0

(HS)XML queries via SYB3-like generic transformations
by oleg＠pobox.com 21 Apr '06

21 Apr '06

This message demonstrates SYB3-like generic term processing for transformations and selections from (HS)XML-like documents. Our running example is an HSXML document that includes Haskell sample code, which is to be formatted specially. Moreover, we wish to guarantee that the sample code is syntactically and type-correct. Furthermore, we wish to evaluate expressions in the context of the code -- and include the result in the rendered HTML document. The present message is inspired by the example posted by shelarcy(a)gmail.com on the Haskell list, Mar 25, 2006, and benefited from discussions with shelarcy(a)gmail.com. We show a quite non-trivial generic term processing: extracting textual data from a particular subterm. The subterm in question is identified by its type; as there may be several occurrences of subterms of a given type, we further identify the desired subterm by its content (i.e., a string label). Some of the extracted strings must be concatenated, while the others must be kept distinct -- depending on their occurrences within terms of a particular type. Thus our extraction has both static aspects (i.e., type-directed) as well as dynamic ones (i.e., guards). We should point out that in HSXML, the notions of the element type (in the sense of the XML Recommendation) and of the Haskell data type coincide. Our generic term processing is like SYB3 in spirit, but different in implementation. In particular, we do not require a GHC extension of recursive instances, although this point is moot here as HSXML so far has had no recursive datatypes. Like SYB3, we do need overlapping instances to specify generic term processing, which the user can later override for particular types of terms. Our HSXML transformations are like SXSLT, SXML pre-post-order transformations described in http://pobox.com/~oleg/ftp/Scheme/xml.html#SXSLT only more modular. The complete code discussed in the present message is in the file sample-code.hs included in the (updated) archive http://pobox.com/~oleg/ftp/Haskell/HSXML.tar.gz Here's our running example: > test_haskell = > (document > (head > [meta_tag [description "Haskell"]]) > (body > [h1 "Haskell examples"] > [h2 "The first sample code"] > [hcode "sample1" > "module Main where" > "fact n = product [1..n]" > ] > [hcode "sample2" > [[thide "module Main where"]] > [[thcode "fix f = f $" [[em "fix f"]]]] > "fact self n = if n <= 0 then 1 else n * (self (n-1))" > ] > [p "the result:" [[thrunh test_haskell "sample2" > "show $ fix fact 5"]]] > )) It includes two pieces of Haskell code, gabelled "sample1" and "sample2". The sample code is comprised of lines, which may include formatting annotations such as 'em' to emphasize certain parts of code. Some other lines of code, like module declarations, can be hidden in rendering for the sake of clarity. We may wish to extract a particular piece of code from the HSXML document, with formatting annotations removed. For example, we may wish to type-check that code, and evaluate some expressions in its context. We may want to do that while we render the HSXML document into HTML. For instance, [[thrunh test_haskell "sample2" "show $ fix fact 5"]], when rendered on a particular formatter, will extract "sample2", evaluate the expression "show $ fix fact 5" in its context, and insert the result in the rendered HTML document: <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd"> <html> <head> <meta name="description" content="Haskell"></head> <body bgcolor="#FFFFFF"> <h1>Haskell examples</h1> <h2>The first sample code</h2> <pre> > module Main where > fact n = product [1..n]</pre> <pre> > fix f = f $ <em>fix f</em> > fact self n = if n <= 0 then 1 else n * (self (n-1))</pre> <p>the result: <code>show $ fix fact 5</code> is 120</p> </body> </html> To accomplish evaluation during rendering, we do invoke a Haskell system (runghc). To render Haskell code in a particular way, we define a new HSXML context, CT_hcode, and provide for interpretation of the text data and elements in that context: > -- Rendering of Strings in the HCODE context. > instance Render (HW CT_hcode String) where > render (HW str) = emit "\n> " >> emit str > > -- `Transparent' grouping of inline elements into one > -- hcode element > -- This is an administrative element > newtype THCODE a = THCODE a deriving Show > thcode x = build (as_hcode . HW . THCODE . rev'apppend HNil) nil_inline x > > -- An THCODE element is rendered as one string > instance RenderInline b => Render (HW CT_hcode (THCODE b)) where > render (HW (THCODE b)) = emit "\n> " >> (render_inline False b > >> return ()) Of a particular interest is SYB3-like generic processing for extracting sections of HSXML document. Just like in SYB3, the generic term processing code naturally splits into three distinct modules. In sample-code.hs file however, all three `modules' are placed within one file, for convenience. Again, our processing is SYB3 is spirit only -- but not in the implementation. The first part of the library defines the class Dat and its instances for term types of interest. This part of the library depends exclusively on the structure of terms -- irrespective of operations one may wish to perform on terms. > class Dat gctx a where > gmapq :: ([gctx] ->w) -> a -> w We also define a class of generic functions, identified by labels. We don't provide any instances of this class here. > class FN label a where > fn :: a -> label Unlike SYB3, Dat is not a subclass of any class (cf. the SAT class in SYB3). Thus the dictionary for Dat does not depend on other dictionaries. Furthermore, the function gmapq has _no_ higher-rank type. The type gctx is the label that identifies a particular generic function to apply. Within this part of the library, we treat gctx as an (abstract) type variable. Here are a few instances of the Dat class > instance Dat gctx (HW ctx String) where > gmapq u _ = u [] > instance Dat gctx HNil where > gmapq u _ = u [] > instance (FN gctx a, FN gctx b) => Dat gctx (HCons a b) where > gmapq u (HCons a b) = u [fn a, fn b] > -- Instances of Dat for more specific containers > instance (FN gctx h, FN gctx b) => Dat gctx (HW ctx (Document h b)) where > gmapq u (HW (Document h b)) = u [fn h, fn b] > > instance (FN gctx h) => Dat gctx (HW ctx (Head h)) where > gmapq u (HW (Head h)) = u [fn h] As we can see, all the instances are quite regular and could well be automatically derived (e.g., via DrIFT). The second part of the generic term traversal library defines generic functions, irrespective of the structure of the types to which they may apply. Here we define a traversal function to extract a part of the term according to some criterion (to be specified later by the user of the library) > class ExtractD a r where > extractd :: a -> r > newtype EXD r = EXD{unEXD :: r} We use EXD in two ways. First, we use its type as a label to select the proper instance of FN. Second, we use EXD to encapsulate the result type. So, we can select the proper instance of generic function based on the argument type and the result type. The following `registers' ExtractD to be a generic `function'. > instance ExtractD a r => FN (EXD r) a where fn = EXD . extractd That's all there is to the second part of the library. Incidentally, none of the previous parts needed overlapping instances. Finally, we wish to instantiate the ExtractD function to extract all strings from a part of the term that has the type (HW CT_block (HCode b)) and which has the desired label (given as a string): First we define a generic extraction, which applies to any term type: > instance Dat (EXD (String->[String])) a > => ExtractD a (String->[String]) where > extractd a label = gmapq (concatMap (\x -> unEXD x $ label)) a This traversal rule obviously does nothing but propagating accumulated strings. To get any result we need to define specific extraction rules. In our case, we have to find the HCode sub-term of the desired label and extract all strings from it: > instance ExtractC b => > ExtractD (HW CT_block (HCode b)) (String->[String]) where > extractd (HW (HCode label b)) label' = > if label == label' > then extractc b > else [] -- otherwise, ignore the whole block Here extractc is yet another generic function, which extracts the string content from terms and returns the array of lines. The following is the entire definition. A three-line boiler-plate: > class ExtractC a where extractc :: a -> [String] > newtype EXC = EXC{unEXC :: [String]} > instance ExtractC a => FN EXC a where fn = EXC . extractc General extraction instance, which only propagates what other instances might have produced: > instance Dat EXC a => ExtractC a where > extractc a = gmapq (concatMap unEXC) a Specific instances: > instance ExtractC (HW ctx String) where > extractc (HW str) = [str] > instance ExtractC b => ExtractC (HW CT_hcode (THIDE b)) where > extractc (HW (THIDE b)) = [concat $ extractc b] > instance ExtractC b => ExtractC (HW CT_hcode (THCODE b)) where > extractc (HW (THCODE b)) = [concat $ intersperse " " (extractc b) ]

1 0

ANNOUNCE: FPS - FastPackedStrings 0.2
by dons＠cse.unsw.edu.au 20 Apr '06

20 Apr '06

I'm pleased to announce version 0.2 of FPS, the fast, packed string library for Haskell. FPS allows you to have time and space efficient arrays of bytes accessed via a List interface, along with fast IO on those strings. FPS is, in particular, suited for heavy duty string and IO projects. It is also useful for applications that must pass strings back and forward from C. Version 0.2 features a number of improvements over v0.1. * Its faster! * There is a richer interface. * More support for converting between C, Addr# and Haskell strings. (in particular, there are 0-copy functions to create FPS strings from Addr# and to create CStrings from FPS') Get it here: Homepage: http://www.cse.unsw.edu.au/~dons/fps.html Interface: http://www.cse.unsw.edu.au/~dons/fps/Data.FastPackedString.html FTP: ftp://ftp.cse.unsw.edu.au/pub/users/dons/fps/fps-0.2.tar.gz darcs: darcs get --partial http://www.cse.unsw.edu.au/~dons/code/fps Cheers, Don ------------------------------------------------------------------------ Here are benchmarking results for 20M strings, for versions 0.1 and 0.2 of FPS, compared against Simon Marlow's prototype packedstring code, the current Data.PackedString library and traditional [Char] functions. Functions that are only provided by FPS (such as the various CString routines) are not tested. Key: FPS2 = Fast Packed String v2 FPS1 = Fast Packed String v1 SPS = Simon Marlow's packedstring prototype PS = Data.PackedString [a] = [Char] ~ = unchanged from FPS2 - = no function exists ! = stack or memory exhaustion Size of test data: 21256k Time in seconds. FPS2 FPS1 SPS PS* [a] ++ 0.078 ~ ! ! 1.288 length 0.000 ~ 0.000 0.000 0.131 pack 0.345 2.043 0.502 0.337 - unpack 1.596 ~ 1.630 7.445 - compare 0.000 ~ 0.000 0.000 0.000 index 0.000 ~ 0.000 0.000 0.000 map 2.664 4.283 2.917 4.813 7.286 filter 0.282 0.482 2.805 0.954 0.305 take 0.000 ~ 0.000 0.024 0.005 drop 0.000 ~ 0.000 11.768 0.130 takeWhile 0.000 ~ 1.498 0.000 0.000 dropWhile 0.000 ~ 1.985 8.447 0.130 span 0.000 ~ 9.289 11.144 0.131 break 0.000 ~ 9.383 11.268 0.133 lines 0.421 ~ 1.114 1.367 2.790 unlines 0.121 ~ ! ! 10.950 words 2.115 3.202 2.128 5.644 4.184 unwords 0.058 ~ ! ! 1.305 reverse 0.024 4.606 12.997 13.018 1.622 concat 0.029 ~ 12.701 11.459 1.163 cons 0.016 3.094 2.064 8.358 0.131 snoc 0.017 1.536 - - - empty 0.000 ~ 0.000 0.000 0.000 head 0.000 ~ 0.000 0.000 0.000 tail 0.000 ~ 0.000 14.490 0.130 last 0.000 ~ - - 0.143 init 0.000 ~ - - 1.147 inits 5.350 - - - ! tails 6.634 - - - 1.136 intersperse 0.034 4.590 - - 10.517 concatMap ! - - - 1.131 any 0.000 ~ - - 0.000 all 0.000 ~ - - 0.000 sort 14.380 15.773 - - ! maximum 0.024 ~ - - 0.183 minimum 0.025 ~ - - 0.185 replicate 0.008 ~ - - 0.053 elem 0.000 ~ 1.490 0.001 0.000 find 0.278 0.366 - - 0.000 elemIndex 0.000 ~ - - 0.000 elemIndicies 4.192 ~ - - 0.314 * Note that it is not possible to directly read a string larger than 1M into a Data.PackedString (due to a space leak). Instead you need to use packString

4 6

ANN: Debian From Scratch 0.99.0
by John Goerzen 20 Apr '06

20 Apr '06

Here's why I'm sending a copy of this announcement to the Haskell list: The tool that generates these ISO images (dfsbuild) is written in Haskell (having been ported from OCaml). The generated ISO images also contain full, working GHC and Hugs environments. -------------------------------------------------------------- Debian From Scratch (DFS) is a single, full rescue CD capable of working with all major filesystems, LVM, software RAID, and even compiling a new kernel. The DFS ISO images also contain a small Debian mirror subset that lets you use cdebootstrap, along with the other utilities on the CD, to perform a manual, "Gentoo-like" installation. It also serves as an excellent rescue CD, with a full compliment of filesystem tools, backup/restore software, and a development environment complete enough to build your own kernels. DFS also refers to dfsbuild, the tool that generates DFS images. dfsbuild is available as a Debian package. dfsbuild is designed to make it trivial to build your own custom DFS images. You can have your own set of Debian packages on your images, your own kernels, etc. Unlike many other systems, you can go from the example dfs.cfg to a customized DFS build in just a few minutes, even if you've never used dfsbuild before. Today I am announcing the availability of DFS 0.99.0 images for i386/amd64 and dfsbuild 0.99.0. dfsbuild is available in sid and (hopefully shortly) testing. My DFS ISOs are available from http://people.debian.org/~jgoerzen/dfs/ DFS is not a separate distribution. DFS images *are* Debian. What you get when you boot a DFS CD is etch (or whatever you built your own images with), with a few minor tweaks to make it run from CD-ROM. These are the major changes since the previous incarnation of DFS: * Entire program rewritten from scratch. Ported from OCaml to Haskell, with some Busybox scripts. * DFS now supports Debian initramfs kernels out of the box. It continues to support kernels with enough drivers statically compiled to load the CD. * Switched from apt-move to reprepro for the generation of the on-CD mini Debian mirror, saving space on the generated image. * DFS images now use (and require) udev. 2.4.x kernels are no longer supported. * CD-ROM autodetection via /sys. * My DFS images are now built with testing (etch). * My DFS images now come with multiple versions of GCC, plus development environments for C, Java, Perl, Haskell, Pyton, OCaml, shell, and tcl. * My DFS images now come with various version-control programs: CVS, darcs, bzr, subversion, git, cogito, arch2darcs, tailor, etc. * Numerous additions of other packages to the DFS images; full list at [1]. My DFS images now contain 575 packages. * My DFS images now built with etch, and come with enough .debs to cdebootstrap sarge, etch, or sid (assuming sid is stable enough at this time) * My DFS images continue to be able to cdebootstrap either an i386 or amd64 system. A more comprehensive list of DFS features is available at [2]. [1] http://people.debian.org/~jgoerzen/dfs/pkglist-0.99.0_i386.txt [2] http://people.debian.org/~jgoerzen/dfs/html/intro.html#FEATURES

1 0