February 2009 - Haskell-Cafe

Video phone application
by Jamie 09 Feb '09

09 Feb '09

Hello! I like to create H.323 (primarly) & SIP based video phone application that will run on Linux, Windows & Mac platforms and of course, written in Haskell. (Writing similar application in Haskell is not "reinventing the wheel" :) I did look/try various current H.323/SIP based video phone software that runs on Windows, Mac and Linux such as Ekiga, XMeeting, Linphone, Mirial and some others. Ekiga (despite version 3.x) is unstable, keeps crashing a lot on me. XMeeting (Mac) does not work good (choppy videos, dropped packets et al). So do Linphone. The best of the lot so far is Mirial Softphone, it has nice decent features to begin with, but there are lot more features I want to add. Please let me know if there is good H.323 based videophone software that I might miss (I think I have looked at all of them). Since my video phone application mainly focus more on deaf people thus strong emphasize on video part while all other (if not all) video softphone put stronger emphasize on audio part & traditional phone setup. Also it has to be primarly H.323 because I would say about over 95% or so of all deaf people (with high speed Internet access) use Sorenson VP-200 as their "main" videophone (Sorenson gives away VP-200 video phone device for free to all deaf people in USA in exchange for using Sorenson Video Relay Service which they make profit from). Sorenson VP-200 only supports H.323, H.263 video codec and G.711/G.723.1 audio codec. Being able to work with VP-200 well is this application's first step. Skype, ooVoo, SightSpeed, QuteCom, VZOchat, iChat AV and some others are excellent video phone/chat applications but they are not H.323 complaint so they are "out of question". It would take me a lot of time (more like forever) if I have to write my own H.323 protocol stack along with video/audio codecs as well the GUI in Haskell. So in order to write this application in timely manner (say one year from now or so) I will need external library/functions for H.323/SIP protocol stack, the audio/video codecs and the GUI. I have been contacting serveral companies that offer licensing of their H.323/SIP/codecs SDK/API/Libraries for my program to be able to access to those functions. Naturally all of them ues C/C++. Yes there are some free/open source ones such as h323plus.org, x264, ffmpeg and some others, but still all of them are written in C/C++. I think I am more likely to go for commercial libraries due to stability (least bugs), higher performance, better documentation and whole set of supported features. What do I ask to make sure those libraries/SDK would work well with Haskell? As for GUI, since I would like for this application to run on three popular platforms: Windows, Mac and Linux, so I think I would have to use Gtk+, wxWidgets or Qt which mean I would use Gtk2Hs, wxHaskell or qtHaskell. Even nicer is to use FRP GUI library such as Fruit, Grapefruit, wxFruit, et al. This application would need to support localization i.e. support different languages. What do you suggest for me to use one of middle level GUI library or even the higher level FRP GUI library (with localization support)? Fast and smooth video (i.e. 30 FPS) is the key for this application. As it stands now, looks like I'll write the "main program" entirely in Haskell, then use some kind of bindings to call to those C/C++ libraries for H.323, SIP, codecs, webcam/video camera I/O, video display, GUI et al. Any better ideas for doing this? I know it would be dream come true to see acclerated vector/raster based drawing primitives with FRP GUI on top written entirely in Haskell along with H.323, codecs et al being written in Haskell too (we may have to wait quite while for that), but we live in real world so I will have to learn to deal with what we have now and the workarounds. Thanks in advance! Jamie

1 1

GHC 6.11 missing time package?
by Lyle Kopnicky 08 Feb '09

08 Feb '09

Hi folks, I'm getting ready to release a piece of software. Unfortunately due to a bug in GHC 6.10 on Windows it does not handle Ctrl+C properly. Since the bug has been fixed (thank you Simon Marlow), I figured I'd download a 6.11 build (I grabbed the 2009-01-29 version). Unfortunately, my project won't build with it because it's missing the time-1.1 package. This is sad because I had gone through the trouble to rewrite my use of oldtime to use time, thinking it was more future-proof. Is this an oversight in the nightly build, or is this package going out of GHC? Thanks, Lyle

3 4

haddock-2.3.0 literate comments discarded from .lhs input
by Alistair Bayley 08 Feb '09

08 Feb '09

I have this test case for Haddock (2.3.0): -------------------------------------------------- | Module : Test.Haddock Copyright : (c) 2009 Alistair Bayley License : BSD-style Maintainer : alistair(a)abayley.org Stability : stable Portability : portable Test case for Haddock. > module Test.Haddock > ( > -- $named_block > Fail(..) > ) > where > data Fail = Fail | Succeed $named_block This is some hadock documentation. -------------------------------------------------- This fails with: [1 of 1] Compiling Test.Fail ( Test\Fail.hs, Test\Fail.o ) Test\Fail.hs:11:26: Can't make a derived instance of `Typeable Fail' (You need -XDeriveDataTypeable to derive an instance for this class) In the data type declaration for `Fail' If I manually unlit, then Haddock is happy: -------------------------------------------------- -- | -- Module : Test.Haddock2 -- Copyright : (c) 2009 Alistair Bayley -- License : BSD-style -- Maintainer : alistair(a)abayley.org -- Stability : stable -- Portability : portable -- -- Test case for Haddock. module Test.Haddock2 ( -- $named_block Fail(..) ) where data Fail = Fail | Succeed -- $named_block -- -- This is some hadock documentation. -------------------------------------------------- so it looks as though it's discarding the literate comments. Is this intended? I was under the impression that because it used the ghc parser, it could now properly handle .lhs input. Ona related note, we have a nice unlitter in cabal that would preserve these comments before invoking haddock. Is this still used with haddock2, or does it now assume haddock2 will do the right thing? Alistair

4 7

ANNOUNCE: X Haskell Bindings 0.1
by Antoine Latter 08 Feb '09

08 Feb '09

I'd like to announce a version bump for the X Haskell Bindings (XHB) library, to 0.1.* from 0.0.*. The goal of XHB is to provide a Haskell implementation of the X11 wire protocol, similar in spirit to the X protocol C-language Binding (XCB). On Hackage: http://hackage.haskell.org/cgi-bin/hackage-scripts/package/xhb This release focuses on making the API a bit friendlier: + 'type BOOL = Word8' has been replaced in the API by Prelude.Bool + type synonyms BYTE, CARD8, CARD16 and CARD32 for the Data.Word types have been eliminated + type synonyms INT8, INT16 and INT32 for the Data.Int types have been eliminated + Previously, all protocol replies were represented by their own distinct data type. Now, if the reply to a request only includes a single field, the request returns that field directly. In more concrete terms: > internAtom :: Connection -> InternAtom -> IO (Receipt InternAtomReply) becomes: > internAtom :: Connection -> InternAtom -> IO (Receipt ATOM) Further work to make the API more "Haskelly" is ongoing. Related projects: X C Bindings: http://xcb.freedesktop.org/ -Antoine

2 2

how to link statically with a c lib ?
by minh thu 08 Feb '09

08 Feb '09

Hi, I'm writing bindings for the Tiny C Compiler. It seems that tcc provide a libtcc.a but no libtcc.so. In my cabal file, I have extra-libraries: dl, tcc but when using the generated haskell module, I have the following message : ⟨thu@jones samples⟩ ghc -e "main" Test.hs <interactive>: <command line>: can't load .so/.DLL for: tcc (libtcc.so: cannot open shared object file: No such file or directory) How can I generate a module linked statically against libtcc ? Thanks, Thu

2 4

any implementation of ACIO?
by Jianzhou Zhao 08 Feb '09

08 Feb '09

Hi, http://www.haskell.org/pipermail/haskell-cafe/2004-November/007715.html mentioned an interesting (A)ffine and (C)entral IO. Are there any packages or extensions to support ACIO in Haskell? Thanks JZ

2 1

[ANN] Working with HLint from Emacs
by Alex Ott 08 Feb '09

08 Feb '09

Hello For Emacs users it could be interesting - I wrote small module for more comfortable work with HLint from Emacs. It has same functionality as compilation-mode - navigation between errors, etc. To use it, just add following code to emacs init file: (require 'hs-lint) (defun my-haskell-mode-hook () (local-set-key "\C-cl" 'hs-lint)) (add-hook 'haskell-mode-hook 'my-haskell-mode-hook) Module is available from http://xtalk.msk.su/~ott/common/emacs/hs-lint.el -- With best wishes, Alex Ott, MBA http://alexott.blogspot.com/ http://xtalk.msk.su/~ott/ http://alexott-ru.blogspot.com/

6 7

Re: [Haskell-cafe] Just how unsafe is unsafe
by Yitzchak Gale 08 Feb '09

08 Feb '09

Peter Verswyvelen wrote: >>> I do have asked myself the question whether >>> a "really random generating" >>> function could be regarded as "pure" somehow I wrote: >> Not really... Alberto G. Corona wrote: > What is pure randomness? .When the algorithmic complexity of the list of > random number is equal to the lenght of the list, that is, is incompressible > by any means. But this can not be probed. the distribution of values gives a > hint of whether if the list obey any algorithm, but even so, many > compressible series have a random distribution. in practical terms, it is > random has a random distribution, the receiver does not know the algorithm > AND the initial value. In this case it can be pure. Perhaps philosophically you can consider it pure, but it is unsafe to treat it as pure in your program. If you write x = genRandomR (1, 100) then the compiler may or may not replace the variable x by the function application genRandomR (1, 100) anywhere in your program. That is probably not what you want. -Yitz

1 0

Moggi :: CT -> Hask
by Gregg Reynolds 07 Feb '09

07 Feb '09

Hi, I had a monadic revelation at about 3 am. The answer to the question "what is an IO value, really?" is "who cares?" I just posted a blog entry discussing how CT found it's way from Moggi into Haskell at http://syntax.wikidot.com/blog (hence the title; Moggi as functor). It addresses the question of what such things are and why Moggi's insight is so brilliant. Feedback welcome, but please remember this is coming from a non-mathematician who likes to write. If you find anything there that outrages your inner Russell, please correct me, but be gentle. Thanks, gregg

2 3

Just how unsafe is unsafe
by Andrew Wagner 07 Feb '09

07 Feb '09

So we all know the age-old rule of thumb, that unsafeXXX is simply evil and anybody that uses it should be shot (except when it's ok). I understand that unsafeXXX allows impurity, which defiles our ability to reason logically about haskell programs like we would like to. My question is, to what extent is this true? Suppose we had a module, UnsafeRandoms, which had a function that would allow you to generate a different random number every time you call it. The semantics are relatively well-defined, impurity is safely sectioned off in its own impure module, which is clearly labeled as such. How much damage does this do? Can we push the lines elsewhere? Is sectioning unsafeXXX into Unsafe modules a useful idiom that we can use for other things as well?

12 14