- Haskell-Cafe - Haskell.org

User group in Stockholm?
by Sergey Bushnyak 01 Apr '13

01 Apr '13

I'll be in Stockholm next month, so I'm curious about haskellers in Stockholm. I saw attempts to organize regular meetings in old mailing threads, but can't find *up to date* information to correct my schedule and see fellow haskellers. Sorry, if this offtopic -- Best regards, Sergey Bushnyak

1 0

Haskel let
by A Smith 01 Apr '13

01 Apr '13

Hi I think I've mastered much of functional programming with Haskell, or at least I can write programs which process as desired. However the existence of the "let" statement evades me apart from a quick way to define a function. Then there is the "in" and "where" parts. Its been suggested its to do with scope and is obvious. I think I need to understand it in a C/C++/Perl style of code to have a Eureka moment when it will be obvious. Can someone refer me to something which will assist me ? Andrew

3 2

Assembly EDSL in Haskell
by C K Kashyap 01 Apr '13

01 Apr '13

Hi Cafe, I am trying to embed x86 assembly in Haskell. I'd like the EDSL to not allow invalid movements into registers - for example, should not allow moving into RIP. I was not able to get it to work. I ended up using DataTypeContexts - which is considered misfeature anyway. I was wondering if I could get some suggestions. {-# LANGUAGE DatatypeContexts #-} data SREG = RIP data DREG = RBX data SNDREG = RAX data (Source s, Destination d) => Instruction s d = MOV s d class Source a class Destination a instance Source SREG instance Source SNDREG instance Destination DREG instance Destination SNDREG move :: (Source s, Destination d) => s -> d -> Instruction s d move s d = MOV s d hello = [move RAX RAX, move RAX RAX] hello = [move RAX RAX, move RAX RBX] -- this is still not allowed. Regards, Kashyap

2 2

graphical editor links ?
by Luc TAESCH 01 Apr '13

01 Apr '13

I am designing some graphical editor ( visio - dia like) and would be interested to hear any pointer to help me think on the architecture of this, in term of data structure/ EDSL/ persistence/ paradigm (FP/ FRP/...) for the different "layers of abstraction " I may see , like rendering libraries ( gtk/wx..)<-> model of the diagram<-> semantics ( this is a state diag/ archimate / mind mapping , i.e. what the glyph and lines means ) <-> persistence Any help & pointers welcomed ! ( and which place is best to discuss this ?) Luc

2 1

Haskell SBV Package with Z3
by J. J. W. 01 Apr '13

01 Apr '13

Dear all, I have a small question regarding the installation of the SBV package. I first installed the SBV 2.10 package with cabal with the following instructions: cabal install sbv Next I installed the Z3 theorem prover and adding the path to my system variables (Windows 7 x64). Next I tested whether I could find it by opening cmd.exe and then typing z3, I get an error message of Z3, so I can safely assume the system can find it. (I added the path to the bin of Z3, I have not included the include directory, I see no reason why I should add a path to this directory, but maybe I am wrong). I ran the program: SBVUnitTests. However it had some errors in the beginning and afterwards a few failures. Having no idea how to fix this, I continued to check whether I can get the SBV to work. So I started to execute the SBV package: ghci -XScopedTypeVariables ghci> :m Data.SBV ghci> prove $ \(x::SWord8) -> x `shiftL` 2 .== 4*x Now this should return Q.E.D., however it returned the following: An error occured. Failed to complete the call to z3 Executable: "C:\\Program Files\\z3-4.3.0-x64\\bin\\z3.exe Options: /in /smt2 Exit code: 0 Solver output: =========================================================== ; :smt.mbqi ; :pp.decimal_precision =========================================================== Giving up.. It does seems like that the Z3 has a normal output, however not a result. Can someone help me to figure out what I actually did wrong? I am using Z3 version 4.3.0, SBV 2.10 and GHCi 7.4.2 Thank you for your help! Yours sincerely, Jun Jie

2 6

Organising code in FRP
by luc taesch 31 Mar '13

31 Mar '13

I am just learning FRP ( via reactive-banana) . so possibly opening a topic already mentionned , apologises in advance, and pointers welcomed. I am wondering about "code structuration" ? ie spagetti code for the network 1/ in a non trivial FRP applicaiton, how to manage "modulatity" or structuration of the Network ? ( most of the example I scanned are toy-size). I ca imagine it may not be trvial to find its way around in a large setup ... 2/ what about the separation of "layers". I mean fine, we extract the reactive ness from the GUI, and separate the layout of widgets from the behaviors of these great, but is it to then mingle the UI with the "business part" ( say in banking), and the infrastructure side ( say persistence, comms, transactionnality, etc) in other words, how tdoes the nice abstraction developped ( module/ interface / categoty patterns, monad, etc) coexsit with a netork of FRP event and behaviors ? Any prior art on this ? ( and which place si best to discuss this ?)

1 0

Specialized Computer Architecture - A Question
by OWP 31 Mar '13

31 Mar '13

Hi, I was reading the paper "A History of Haskell: being lazy with class" ( http://www.haskell.org/haskellwiki/History_of_Haskel) and came on a section (I 2.1) which briefly explained the work people did on making non-von Neumann architectures. It concluded by saying "Much (but not all) of this architecturally oriented work turned out to be a dead end, when it was later dis- covered that good compilers for stock architecture could outper- form specialised architecture. But at the time it was all radical and exciting." These "stock architectures", were they really so good that they out performed the specialized ones on it's own merits or was this mainly due to Moore's Law on transistors? In other words, suppose we separate Moore's Law from the stock architecture, would it still outperform the specialized ones? Thank you for any response

6 8

Threadscope 0.2.2 goes in segmentation fault on Mac Os X 10.8.3
by Alfredo Di Napoli 31 Mar '13

31 Mar '13

Hi Cafè, I've tried installing threadscope, but when I run it from console with "threadscope" I get a laconic segmentation fault. Some info to help the debugging: * Installed gtk via brew * gtk-demo runs correctly * I'm using Mac Os X 10.8.3 * Running gtk and threadscope through a virtual environment (provided by hsenv) * Using XQuartz 2.7.4 I know it's a bit difficult to debug this way, I can try debugging with gdb if it can help. Cheers, A.

3 4

[ANN] hoodle-0.2
by Ian-Woo Kim 30 Mar '13

30 Mar '13

Hi, all, Pen note-taking program hoodle, which is being developed entirely in haskell, is updated to version 0.2. The previous version was 0.1.1. About what is hoodle, please refer to http://ianwookim.org/hoodle The changes in this version are - hoodle data format updated. now hoodle document has a unique id. - linking between documents implemented - vertical space inserting tool implemented - a widget introduced for panning and zooming - drag and drop linking implemented - drag and drop image embedding implemented - poppler pdf support is mandatory - pdf can be embedded in a file - toolbox ui reflects current hoodle state more correctly - slimmer select box design - several rendering glitches fixed Note that poppler is mandatory for hoodle (in the previous versions, poppler support was optional). One needs to install poppler-glib before installing hoodle. After gtk2hs and poppler installed, installing hoodle should be simply > cabal install hoodle As for the detail information about wacom tablet support, please read installation page on the webpage. If you were using the previous versions of hoodle, you will find this version gives you much smoother experience and many interesting improvements. (especially adding vertical space tool and pan-zoom widget ) Enjoy~ best, Ian-Woo Kim

2 2

Idea: Source maps for GHC's -pgmF
by Christopher Done 30 Mar '13

30 Mar '13

Here's an idea. For the -pgmF flag, because compile errors become really obscure and line numbers rendered unhelpful, why don't we use source maps as recently done for JavaScript in the browser? Info: https://wiki.mozilla.org/DevTools/Features/SourceMap https://docs.google.com/document/d/1U1RGAehQwRypUTovF1KRlpiOFze0b-_2gc6fAH0… Here is an example I experimented with the Fay (Haskell→JS compiler): https://dl.dropbox.com/u/62227452/Screenshots/Screenshot%20from%202012-09-2… This shows the original source code and a backtrace of an exception thrown in JavaScript. We can use Mozilla/Google's format: http://hackage.haskell.org/package/sourcemap Though it's kind of an obscure format because it's aimed at being small, but if it's already been done, why not? Otherwise a simple JSON format is probably sufficient. It could be used both in GHC's compilation, pattern match exceptions, and even in debugger stepping, if desired. Provided with, e.g. -pgmFsm foo.map. This would make things like the XML mode or other syntactical extensions feel quite natural and not awkward in GHC's error messages. What say you, -pgmF users and GHC hackers? Ciao!

1 0