March 2008 - Haskell-Cafe

Thanks for the help: image of draft visualization
by Jefferson Heard 07 Mar '08

07 Mar '08

Thanks for everyone's help on the list re my Haskell woes with the latest visualization effort. I've been making my code more generic for the last week, and I plan on releasing a visualization framework back to the community at some point. Gotta get approval from my boss before releasing code back to the wild, but here's an image of the draft of the visualization I've been working on: http://vizdata.renci.org/projects/jeff/ProteinViz/ProteinViz.png Basically, what you're seeing is 18,500 genes arranged in a full heirarchical clustering (the clustering technique uses a metric I'm unfamiliar with, and I got the dataset pre-packaged from the guy who's using it). The final visualization is fully interactive and runs fine under Linux, dies a miserable death under windows (runs out of RAM, and I can't figure out why, nor do I particularly care), and primarily runs on our 14 foot by 9 foot (4.5x3 metres for those of us in metric) linux display wall. The full program is about 800 linux of Haskell code using nothing but the standard library under GHC 6.8. Again, thanks for all your help! -- Jeff

2 1

Network.Socket Woes
by Scott Bell 07 Mar '08

07 Mar '08

Hi all, I'm trying to do some standard socket stuff using the Network.Socket library. This example is a UDP server: import Network.Socket proto_udp :: ProtocolNumber proto_udp = 17 port = 10162 main = do s <- socket AF_INET Datagram 17 bindSocket s (SockAddrInet (PortNum port) iNADDR_ANY) putStrLn $ "Listening on port " ++ (show port) ++ "... " (msg, n, addr) <- recvFrom s 1024 putStrLn $ "Received " ++ (show n) ++ " bytes from " ++ (show addr) ++ ": " putStrLn (show msg) Using the equivalent C program (socket, bind, recvfrom) I'm able to successfully receive a 64 byte message from my test program. The Haskell version, however, does not return from recvFrom. I've also tried wrapping this in a withSocketsDo, with no effect. How can I proceed to debug/troubleshoot this program? Thanks, - Scott Bell

7 9

Small displeasure with associated type synonyms
by Stefan Holdermans 07 Mar '08

07 Mar '08

Dear all, I was doing a little experimentation with associated type synonyms and ran into a small but unforeseen annoyance. The following, contrived, example snippets illustrate my pain: First, let us declare some of the simplest classes exhibiting associated type synonyms that I can think of, class C a where type T a val :: T a together with a trivial instance for nullary products: instance C () where type T () = () val = () But then, let us try an instance for binary products: instance (C a, C b) => C (a, b) where type T (a, b) = (T a, T b) val = (val, val) I really thought this would work out nicely, but GHC (version 6.8.2) gracefully gives me Couldn't match expected type `T a2' against inferred type `T a' Expected type: T (a2, b) Inferred type: (T a, T a1) In the expression: (val, val) In the definition of `val': val = (val, val) Couldn't match expected type `T b' against inferred type `T a1' Expected type: T (a2, b) Inferred type: (T a, T a1) In the expression: (val, val) In the definition of `val': val = (val, val) while I think I deserve better than that. Can someone (Tom?) please explain (a) why the required unifications fail, (b) whether or not it is reasonable to expect the unifications to succeed, and (c) how I can overcome problems like these? Surely, I can have val take a dummy argument, but I am hoping for something a bit more elegant here. I tried lexically scoped type variables, but to no avail: instance forall a b. (C a, C b) => C (a, b) where type T (a, b) = (T a, T b) val = (val :: T a, val :: T b) gives me Couldn't match expected type `T a2' against inferred type `T a' In the expression: val :: T a In the expression: (val :: T a, val :: T b) In the definition of `val': val = (val :: T a, val :: T b) etc. Cheers, Stefan

7 10

Analysing Haskell with Graph Theory
by Ivan Miljenovic 07 Mar '08

07 Mar '08

Back in December, I posted that I was thinking about doing a project for my honours year on using graph theory for analysis on Haskell source code, amongst others: http://thread.gmane.org/gmane.comp.lang.haskell.cafe/32912 Well, I've officially started this project, and my supervisor and I have come up with the following functions that would probably be relevant to analysing _any_ piece of source code (the actual code will be tentatively stored as a directed graph with the nodes being functions and the edges being function calls, i.e. if f calls g, there is a directed edge from f to g): * root finding -> find nodes with no incoming edges (in a program, you'd ideally want only main to be such a node) * cycle detection -> find a possibly recursive cycle in your code: if the cycle is too big, then you may wish to consider refactoring * depth analysis -> find leaves that have a depth from the root node that is extremely large compared to the others (if a function is 50 calls down from main compared to an average of 15, you may wish to refactor) * chain detection -> find connected functions that have only one incoming and one outgoing edge, e.g. : f -> g -> h : if g isn't used anywhere else, you may wish to combine it inside either f or g * node popularity -> get a count of how many functions use a particular function and how many other functions it calls (related to chain detection above) * clique detection -> probably not as relevant to source code, but if you have a large number of functions that are all pairwise co-recursive than you may wish to refactor Can anyone think of any other kind of functions that would be useful in this kind of source code analysis? Also, I'm going to be using haskell-src for for the parsing of the Haskell code so that I don't waste my time writing a parser, when my project is technically on the graph-theory side of things. I know that it focuses mainly on H98 with only a few extensions... is this likely to be a problem if I want to try running my eventual program on say ghc or xmonad as test examples? -- Ivan Lazar Miljenovic

4 7

Build dependency problem with bytestring.
by Jon Strait 07 Mar '08

07 Mar '08

Hi all, I'm upgrading HAppS from my 0.9.2 to the current 0.9.2.1 in Hackage. Some HAppS components have built and installed (IxSet and Util). Halfway through the HAppS-State build, the GHC runtime linker gives a fatal error on finding a duplicate definition for symbol fps_minimum while loading bytestring 0.9.0.4 after having already loaded bytestring 0.9.0.1. From trial and error, I'm guessing that HAppS-State needs 0.9.0.1 but one of its build dependencies was built with 0.9.0.4. Using GHC 6.8.2 Linux here. Any suggestions on how best to currently approach this situation? Thanks, Jon

3 3

ANN: Emping 0.4
by Hans van Thiel 07 Mar '08

07 Mar '08

Hello All, Emping 0.4, a utility which can derive heuristic rules from facts in Open Office Calc spreadsheet .csv format, has been released on HackageDB. Version 0.4 supports blank fields, treated as 'none of the others', which allows for split attribute values. For example, "Owns Car" could have values "Yes" or 'No", and then "Yes" could be split into "Make", "Price Range" and so on, while the fields for "No" could be blank for these attributes. Data still have to be in the spreadsheet format of rows and columns (and in OO Calc .csv). Version 0.4 has a Gtk2Hs based GUI and the code has been changed in several places. Compiled with GHC 6.8 it appears to be noticeably faster than version 0.3. The package comes with a user guide, a white paper, and two data files from the UCI Machine Learning Repository. To get an impression see: http://j-van-thiel.speedlinq.nl/emp/empug.html The compiled binary for Linux, with documentation, can also be downloaded directly from my web site. Thank you, Hans van Thiel

1 0

Re: [Haskell-cafe] Analysing Haskell with Graph Theory
by Ivan Miljenovic 07 Mar '08

07 Mar '08

(CC'd to the list) On 07/03/2008, Dusan Kolar <kolar(a)fit.vutbr.cz> wrote: > I'm just wondering what algorithms you use for analysis. For instance, > clique is an NP complete problem. Do you use some approximation? If yes, > which? I'm particularly interested in the algorithm. I haven't gotten that far yet ;-) As it stands, my supervisor have discussed this slightly and I'll probably start off with a more-or-less brute-force approach. This should be OK as the graph will probably be reasonably sparse and (hopefully) not too big! -- Ivan Lazar Miljenovic

1 0

CABAL: conditional executable?
by Magnus Therning 06 Mar '08

06 Mar '08

I'm putting together a package consisting of 2 executables. Only one of them is pure Haskell and thus buildable on all platforms, the other relies on Windows API calls and can only be built on that platform. I found the “if os(...)” conditional in the CABAL docs but I'm having problems getting it to do what I want. if os(mingw32) executable foo ... Results in the error “Section expected”. Swapping the two lines like this executable foo if os(mingw32) ... results in “Setup.hs: Error: No 'Main-Is' field found for executable foo”. Is there a way to get CABAL to do what I want or should I raise a feature request on the CABAL trac? /M -- Magnus Therning (OpenPGP: 0xAB4DFBA4) magnus＠therning．org Jabber: magnus．therning＠gmail．com http://therning.org/magnus What if I don't want to obey the laws? Do they throw me in jail with the other bad monads? -- Daveman

4 10

"Hello World" in Haskell for web browser
by Dimitry Golubovsky 06 Mar '08

06 Mar '08

Hi, I have noticed that some people tried to compile a traditional Haskell program using IO monad to show "Hello World" using Yhc Web Service. Thanks for your interest, and here are some clues. The programming paradigm chosen for Haskell Web programming is not based on monads. It is based on CPS-style calls to DOM functions by Haskell code compiled to Javascript. Further on, additional layers may stack up (such as Haskell Web Toolkit) to provide more convenient APIs, but DOM is the basis. So here is an example of "Hello World" program written this way: -- begin pasteable code -- module HelloWorldDOM where import CPS import UnsafeJS import CDOM.Level2.DomUtils import DOM.Level2.Dom import DOM.Level2.Html2 import DOM.Level2.HTMLElement import DOM.Level2.HTMLDivElement main = getHTMLDocument $ \doc -> documentBody doc $ \body -> mkText doc "Hello World" $ \txt -> mkDiv doc $ \dv -> addChild txt dv $ \_ -> addChild dv body $ id -- end pasteable code -- The meaning of this: * get reference to the HTML document node first (it is the parent of everything), * extract the <BODY> tag node, create at text element with "Hello World", * create a <DIV> tag node, * insert the text node into div, * insert div into body Or, same in HTML: <html> <body> <div> Hello World </div> </body> </html> but filled in dynamically. Using Haskell Web Toolkit API, the same may be expressed in more compact fashion: -- begin pasteable code -- module HelloWorldHsWTK where import DOM.Level2.HTMLDivElement import Graphics.UI.HsWTK main = docBodyC (mkDiv |<< textP "Hello World") -- end pasteable code -- Here, docBodyC is roughly equivalent of the first two DOM calls, mkDiv is same as above, |<< means "insert into container", and textP is a wrapper around mkText. Earlier, I posted the link to Haddock-generated documentation which also includes the Haskell DOM API. Here it is again: http://www.golubovsky.org:5984/_utils/yhcws/index.html Hopefully this example along with the documentation provided helps shed some light on Haskell Web programming techniques. Feel free to ask questions. Thanks. -- Dimitry Golubovsky Anywhere on the Web

1 0

Re: [Haskell-cafe] Connection helpers: for people interested in network code
by brad clawsie 06 Mar '08

06 Mar '08

"Gwern Branwen" <gwern0(a)gmail.com> writes: >> i offered some time ago to look at building a cabal package and >> documentation for this. i would offer up excuses as to why this hasn't >> appeared yet, but between kids, work and ski season i just haven't >> allocated the time yet. sorry to all. i still hope to look into this, >> but fair to say i dropped the ball. > > Dunno. I just downloaded the git repo, and I'm not sure what's > stopping anyone from uploading it to Hackage. It builds cleanly and > with essentially no warning on 6.8.2, the Haddock docs build, and so > on. The Cabal file itself is quite good; out of boredom, I tweaked it > a little: awesome! that means someone has indeed recognized my lameness and subbed in to do what i didn't. thanks to whoever you are! sorry again for promising and not delivering [gwern, hope you don't mind if i redirect your personally reply to the list for context preservation]

1 0