July 2009 - Haskell-Cafe

Re: [Haskell-cafe] Laziness enhances composability: an example
by Marcin Kosiba 10 Jul '09

10 Jul '09

On Thursday 09 July 2009, you wrote: > 2009/7/9 Marcin Kosiba <marcin.kosiba(a)gmail.com>: > > I thought I'll go the "smart" way and rely on the Python yield construct > > to do a CPS transformation of my code. > > While this worked to a certain extent, composability was a > > problem, because any sub-procedure which used yield (and thus was an > > iterator) required being called in a rather inelegant way. > > i confess to not understanding the inelegance. > can you provide an example where an iterator inside CPS requires > special support? Hi, To illustrate what I meant I'm attaching two examples. In example_1.py I've written code the way I think would be elegant (but it doesn't work). In example_2.py I've written code so that it works, but it isn't elegant. I know I'm abusing Python iterators here. Also, I'm not sure the way to compose iterators shown in example_2.py is the only option. Actually I'd love to see a better solution, because it would remove a lot of bloat from my code ;) Thanks! Marcin Kosiba

2 2

Re: [Haskell-cafe] Leaner Haskell.org frontpage
by haskell＠kudling.de 10 Jul '09

10 Jul '09

> I've seen book providing a "chapters at a glance" part, just before the real table of content. Such an inverted pyramid is exactly the consequence Nielson draw from the "F shape pattern" (http://www.useit.com/alertbox/reading_pattern.html) And that's my critque: i don't see the most important things there are to be sayed about Haskell in the top left corner. > For the hompage we're talking about, glancing is even simpler since everything is on the same page and you can scroll it quite easily. I don't agree that "everything on one page" makes comprehension easier. > I'm not sure hiding a level of the hierarchy of information behind a few clicks make things easier. That depends on which task we are talking about: - getting an overview of all available information, or - finding exactly what you are looking for I think we should optimize for the latter, where "What is Haskell?" being the most improtant question.

3 2

iteratee enumHandle
by Paolino 10 Jul '09

10 Jul '09

I'm testing iteratee. This is the possible bug I've found import Data.Iteratee.IO import Data.Iteratee.Base import Data.Iteratee.Char import System.IO import Control.Exception main = do h <- openFile "mamma23" ReadWriteMode hPutStrLn h "ciao" hSeek h AbsoluteSeek 0 l <- (enumHandle h stream2list :: IO (Iteratee IO String)) >>= run print $ assert (l == "ciao") () assertion failed This aside, I'd like to know if there is a way to use parsec library to make an Iteratee, so if it's possible to make parsec spit out a continuation on EOF, or maybe as in delcont language to capture it. thanks paolino

2 1

c2hs confusion with a simple function
by Jeff Heard 09 Jul '09

09 Jul '09

The c2hs documentation at http://www.cse.unsw.edu.au/~chak/haskell/c2hs/docu/implementing.html#id3149… gives me an example to follow for this case: {#fun notebook_query_tab_label_packing as ^ `(NotebookClass nb, WidgetClass cld)' => {notebook `nb' , widget `cld' , alloca- `Bool' peekBool*, alloca- `Bool' peekBool*, alloca- `PackType' peekEnum*} -> `()'#} so I took the original C binding: void SHPGetInfo(SHPHandle, int*, int*, double*, double*) And wrote a c2hs binding: {#fun SHPGetInfo as getInfo { fromSHPHandle `SHPHandle' , alloca- `Int' , alloca- `Int' , alloca- `Double' , alloca- `Double' } -> `()' #} The error I get: Internal.chs:85:2: Couldn't match expected type `Int' against inferred type `(Ptr b -> IO c) -> IO c' Expected type: IO (Int, Int, Double, Double) Inferred type: IO ((Ptr b -> IO c) -> IO c, (Ptr b1 -> IO c1) -> IO c1, (Ptr b2 -> IO c2) -> IO c2, (Ptr b3 -> IO c3) -> IO c3) In the expression: alloca $ \ a2' -> alloca $ \ a3' -> alloca $ \ a4' -> ... In the expression: let a1' = fromSHPHandle a1 in alloca $ \ a2' -> alloca $ \ a3' -> alloca $ ... I've tried ending it with withCIntConv and withCFloatConv to no avail. The original c2hs file is attached. I'm sure there are many more problems, but this is the one I'm working on at the moemnt... -- Jeff

1 0

[Haskell-cafe] ANN: hsparql, a SPARQL query generator/DSL and client
by Jeff Wheeler 09 Jul '09

09 Jul '09

I'm excited to announce the first version of hsparql. HSparql makes it easy to query SPARQL-compliant servers using a relatively intuitive DSL and very simple client. SPARQL is a SQL-ish query language for stores of RDF data. So, SPARQL lets you search the structured data in several big databases. I wrote it so that I could search DBpedia [1], a neat site that pulls structured data from Wikipedia's infoboxes, and also ties it to remote data services like MusicBrainz and US Census information. HSparql isn't tied to DBpedia in any way, though, and can search any server. The DSL looks something like this: > simple :: Query [Variable] > simple = do > resource <- prefix (iriRef "http://dbpedia.org/resource/") > dbpprop <- prefix (iriRef "http://dbpedia.org/property/") > foaf <- prefix (iriRef "http://xmlns.com/foaf/0.1/") > > x <- var > name <- var > page <- var > > triple x (dbpprop .:. "genre") (resource .:. "Web_browser") > > triple x (foaf .:. "name") name > triple x (foaf .:. "page") page > > return [name, page] It's currently available on Hackage [2], and at the moment the source is only available from the tarball hosted there. I'd greatly appreciate any feedback on this package. Thanks! Jeff Wheeler [1] http://dbpedia.org/ [2] http://hackage.haskell.org/package/hsparql

2 2

Re: [Haskell-cafe] Leaner Haskell.org frontpage
by haskell＠kudling.de 09 Jul '09

09 Jul '09

> I find it very to the point and not overwhelming at all : it's easy to glance over it and find quickly what I want. Thanks for your feedback. Most people feel overwhelmed when confronted with more than 7+-2 items: http://www.smashingmagazine.com/2007/10/09/30-usability-issues-to-be-aware-… On the current Haskell frontpage there are over 60 links competing for attention. I am not sure whether we should design interfaces solely with few people having exceptional abilities in mind. This could be understood as a statement about who Haskell is made for in itself. Bye, Lenny

3 2

About return type
by ZhangXu 09 Jul '09

09 Jul '09

_________________________________________________________________ Messenger安全保护中心，免费修复系统漏洞，保护Messenger安全！ http://im.live.cn/safe/

1 0

Leaner Haskell.org frontpage
by haskell＠kudling.de 09 Jul '09

09 Jul '09

Hi, i find the current www.haskell.org frontpage quite overwhelming. Compare it for example with the home pages of other programming languages : http://caml.inria.fr/ http://factorcode.org/ http://sbcl.sourceforge.net/ http://www.ruby-lang.org/en/ http://www.falconpl.org/ Here is my sketch of a leaner, more structured Haskell front page: http://www.haskell.org/haskellwiki/User:Lenny222/Haskell Comments? Bye, Lenny

3 2

ANN: darcs 2.3 beta 2
by Petr Rockai 09 Jul '09

09 Jul '09

Hello! Two weeks passed and it is about time to release darcs 2.3 beta 2 (it's already a day late, sorry about that). As with beta 1, there is only a single installation package for this release of darcs: cabalised source. (Please note that the final version with also come with the legacy autoconf-based buildsystem, for the last time.) You can either download a tarball from <http://repos.mornfall.net/darcs/darcs-2.2.98.2.tar.gz> and build manually (see the build instructions in README inside the tarball), or, alternatively, you can use cabal-install to obtain a copy (the beta release is now available on Hackage): $ cabal update $ cabal install darcs-beta This, if successful, will give you a darcs binary in `~/.cabal/bin` -- you should probably add that to your PATH. What's new since beta 1 ----------------------- There are basically two relatively intrusive changes since beta 1. First, I have added index upgrade functionality to hashed-storage and this is now used by darcs, so that any bad or incompatible indexes are recreated transparently by darcs. Moreover, the index format is now architecture-independent, meaning that repositories shared across multiple architectures should not suffer from excessive index rebuilding. Second, Ganesh has done further work on his gzip CRC correction code. There is a very slight risk of regressions, so if you have around any repositories with broken CRCs in them, please test this functionality. Moreover, there is now an option to limit the repair to current repository, avoiding changes in related branches or caches. Trent, Thorkil and Eric have done cleanup work and a number of minor fixes for this release, improving overall quality and polish. One bug in the new indexing code has been found and reported by Eric and Guillaume independently, and is fixed in hashed-storage 0.3.4 (which required by this beta). You can track the release plan and progress at our wiki: <http://wiki.darcs.net/development>. The question of GHC 6.8 ----------------------- Using GHC 6.10.3 or newer is *strongly recommended*. You may compile darcs with GHC 6.8, but there are several caveats. If you are using 6.8.2 or older, please disable mmap support (pass -f-mmap to cabal install or runghc Setup configure below). Note that the GHC 6.8.2 that ships with Debian Lenny is not affected and it should be safe to keep mmap enabled. It is also recommended to disable use of Hackage zlib when compiling with GHC 6.8.2 (including the Debian Lenny version): pass -f-zlib to cabal. When using zlib, we have seen occasional crashes with error messages like "openBinaryFile: file locked" -- this is a known GHC 6.8.2 bug (and is fixed in GHC 6.8.3). Last, if you are using a 64-bit system, darcs may hang when you exit a pager when compiled with GHC older than 6.10.3. Although this is harmless, it is quite inconvenient. Overall, the status of GHC 6.8 is "semi-supported": for many cases, things will work just fine, especially if you take a little extra caution with compilation flags. Yours, Petr. -- Petr Ročkai | http://web.mornfall.net A physicist is an atom's way of knowing about atoms. (George Wald)

1 0

ANN: Hayoo! beta 0.4
by Timo B. Hübel 09 Jul '09

09 Jul '09

Hello! We are pleased to announce the next beta version 0.4 of Hayoo!, the Haskell API search engine with find-as-you-type and suggestions. Visit Hayoo! here: http://holumbus.fh-wedel.de/hayoo This release includes some major changes to the web interface: - The full description of a function can now be displayed upon request, using the small "+" symbol in front of the description text - The most common root modules for all functions found are shown as list, providing a shortcut for further restricting the current query - Some example queries on the start page for easier use of Hayoo! Additionally, we have again updated the search index. It contains all packages from Hackage as well as gtk2hs as of 06.07.2009, a total of 111.946 function and type definitions. Please let us know about any problems, hints and suggestions. Happy hacking, Timo

2 1