January 2011 - Haskell-Cafe

GHC ParseTree Module
by Jane Ren 03 Jan '11

03 Jan '11

Hi, Does anyone know what GHC module gets the AST and type info of some source code? This is the GHC module that converts all of Haskell into an AST with a small number of pattern cases, where each AST node is annotated with the Haskell type. Thanks

2 1

Data.Typeable TypeRep Ord instance.
by Serguey Zefirov 02 Jan '11

02 Jan '11

Why TypeRep does have equality and doesn't have ordering? It would be good to have that. Right now when I have to order two type representations I convert them to string and then compare. This is somewhat inefficient and not quite straightforward.

7 10

How to write Manatee application?
by Andy Stewart 02 Jan '11

02 Jan '11

Hi all, How to write Manatee extension? Now have answer: Please look http://www.flickr.com/photos/48809572@N02/5031811365/lightbox/ to understand framework before continue. Template code at http://hackage.haskell.org/package/manatee-template First, i explain Manatee package hierarchy: manatee-core contain DBus protocol detail and toolkit use in other manatee packages. manatee-anything is multi-task search interface for user input. manatee contain daemon process, Daemon process is root process to manage other manatee processes: such as render application process' graphics for sandbox protected, window manage ... etc. Daemon process just do manage job and don't running any special application code in it. Other manatee-* package, such as manatee-editor, manatee-browser...etc running in it's own process, you can write any code to implement your application, just need follow some interface: class Typeable a => PageBuffer a where -- Get buffer name. pageBufferGetName :: a -> IO String -- Set buffer name. pageBufferSetName :: a -> String -> IO () -- DBus client to handle DBus signal pageBufferClient :: a -> Client -- How to create view with buffer for MVC design. pageBufferCreateView :: a -> PagePlugId -> IO PageViewWrap -- Page buffer mode. pageBufferMode :: a -> PageMode -- Get package name to update user's configure. pageBufferPackageName :: a -> IO String class Typeable a => PageView a where -- Buffer for view. pageViewBuffer :: a -> PageBufferWrap -- GtkPlug id to send DBus signal to other process. pageViewPlugId :: a -> TVar PagePlugId -- How to handle key action from daemon process. pageViewHandleKeyAction :: a -> Text -> SerializedEvent -> IO () -- How to do when manatee framework focus on current application. pageViewFocus :: a -> IO () -- Scrolled window to contain application widget. pageViewScrolledWindow :: a -> ScrolledWindow -- How to handle cut action, can ignore. pageViewCut :: a -> IO Bool -- How to handle copy action, can ignore. pageViewCopy :: a -> IO Bool -- How to handle paste action, can ignore. pageViewPaste :: a -> IO Bool -- How to scroll top postion, can ignore. pageViewScrollToTop :: a -> IO () -- How to scroll bottom postion, can ignore. pageViewScrollToBottom :: a -> IO () -- How to scroll vertical page, can ignore. pageViewScrollVerticalPage :: Bool -> a -> IO () -- How to scroll vertical step, can ignore. pageViewScrollVerticalStep :: Bool -> a -> IO () -- How to scroll to left side, can ignore. pageViewScrollToLeft :: a -> IO () -- How to scroll to right side, can ignore. pageViewScrollToRight :: a -> IO () -- How to scroll horizontal page, can ignore. pageViewScrollHorizontalPage:: Bool -> a -> IO () -- How to scroll horizontal step, can ignore. pageViewScrollHorizontalStep:: Bool -> a -> IO () If you have any question, feel free to ask me (lazycat.manatee(a)gmail.com) -- Andy

1 0

OpenCL Raw package
by Pavel Perikov 02 Jan '11

02 Jan '11

Hi list (and hopefully Jefferson Heard!) I'm trying to play with OpenCL Raw package. While it has a great amount of work put into it it's not currently usable. Frankly, it does not even link (due to foreign import typo). It does not compile with GHC 7.0.1, has badly specified dependencies on mtl, wrongly defined data types etc. Anyway, many thanks to Jefferson for doing such body of work. I think having OpenCL package is very handy and opens many possibilities for using haskell in high performance computing. Sadly, OpenCL Raw does not seem to have any maintainer currently. I personally not ready to become a maintainer but I started to accumulate patches and I'm willing to send them somebody. Anyone? I set up mercurial repo at http://bitbucket.org/ppavel/opencl-raw and plan to publish patches there. Anyone interested please don't hesitate to contact me. Pavel

1 0

forcing a type constraint on data
by Mathijs Kwik 02 Jan '11

02 Jan '11

Hi all, I would like to make a more restrictive version of a certain datatype. I tried wrapping it in a newtype, fundeps, type families, really got lost :) What I'm trying to achieve is the following: The original package http://hackage.haskell.org/packages/archive/reaction-logic/2010.11.17/doc/h… has a constructor function (mkReactor) with exactly the constraints I'm after. However, the constraint is just on this constructor function, not on the data itself. All functions I write are inferred as "Reactor (State s) c", while the use of that constructor ensures that s == c. Also, I might switch "State s" for some "StateT s m" later on, so I would really like to be able to use "(MonadState c m) => Reactor m c" in my functions as well, since this is the most generic. Trying that blows up however: Could not deduce (m ~ StateT s Identity) from the context (ValidEvent event, MonadState s m) `m' is a rigid type variable bound by the type signature for `stepper2' at Eventer.hs:77:45 In the first argument of `runState', namely `(insertExternals reactor [Serial event])' In the expression: runState (insertExternals reactor [Serial event]) state In an equation for `stepper2': stepper2 (Just reactor, state) event = runState (insertExternals reactor [Serial event]) state Can someone please explain the cause of this problem? Are there workarounds? Language extensions? Thanks in advance, Mathijs

2 1

Defining a type depending on the word size of platform
by Robert Clausecker 02 Jan '11

02 Jan '11

Hi guys! I'm working on a random generator package and want to ask, whether it's possible to generate a type like Int, that can have different sizes at runtime. Motivation: The package ships to generators, one for 32 (lets name it MyGen32) and one for 64 bits (MyGen64). As it is not good for performance to use the wrong one, I also want to provide a type MyGen, that is either a type synonym for MyGen32 or MyGen64, depending on the platform. Any ideas how to do this at compile time? I want this to compile with Cabal so that anybody can download it as a package. I guess the easiest way would be to use CPP, but is there any flag for the wordsize? Yours, Robert Clausecker

3 2

OpenSSL question
by Charles-Pierre Astolfi 02 Jan '11

02 Jan '11

Hi -cafe, I'm using the OpenSSL package, and I don't see any way to forge my own keys. (http://hackage.haskell.org/packages/archive/HsOpenSSL/0.9/doc/html/OpenSSL-…) For example, in the case of RSA, I'm given the exponent and modulus and I would like to create a public key from that, of type RSAPubKey. Is it possible ? How should I do? I'm not particularly tied to OpenSSL, anything that let me use RSA and DSA would be fine. Regards, -- Cp

3 2

getting last char of String
by Aaron Gray 01 Jan '11

01 Jan '11

Is there an easy Haskell function that gets the last Char of a [Char] or String ? Many thanks in advance, Aaron

8 12

V.I.P.s and the associativity of merge'
by Leon Smith 01 Jan '11

01 Jan '11

Ok, after mulling over the issues that Will Ness has brought up in the last few days [1], I think I have a partial explanation for the apparent tension between Will's observations and Heinrich Apfelmus's Implicit Heaps article [2], which both concern the implementation of mergeAll [3]. The merge' function takes two ordered lists, with the head of the first list less than the head of the second, and merges their contents: merge' [] ys = ys merge' (x:xs) ys = x : merge xs ys The nice thing about this function is we can merge an infinite number of lists by folding right, if assume that the heads of those lists are appropriately ordered. This appears in Richard Bird's code at the end of Melissa O'Neill's "Genuine Sieve of Eratosthenes", though undoubtedly this observation has been independently made by many people. Now, in an ordinary sense, merge' *is* an associative operator, in that given three fully defined ordered lists with ordered heads, merge' xs (merge' ys zs) == merge' (merge' xs ys) zs. This allows us to merge an infinite number of lists using an arbitrary tree of merge' operations, without ever worrying that we will return an incorrect result. (However, we might get stuck in a non-productive loop, for example if we fold left over an infinite list of ordered lists) However, Heinrich's article uses a stronger sense of associativity which includes laziness properties and thus captures some operational characteristics. In this sense, merge' *is not* associative. To remedy this, Heinrich uses VIPs to strengthen the associativity property of merge'. This raises the question, is there some combination of the shape of the merge' tree and some input for which using VIPs dramatically changes the efficiency of a mergeAll operation? I suspect the answer is yes, though I don't know for sure at this point in time. However, I do tacitly believe that the current tree that mergeAll uses doesn't exhibit this property for any input, and so I have simplified the implementations of mergeAll and unionAll in the latest version of data-ordlist-0.4.4 by avoiding the use of VIPs. This has the nice side benefit of modestly improving performance when the elements of the result are highly biased towards the first list. Best, Leon [1] http://permalink.gmane.org/gmane.comp.lang.haskell.cafe/84666 [2] http://apfelmus.nfshost.com/articles/implicit-heaps.html [3] http://hackage.haskell.org/packages/archive/data-ordlist/0.4.4/doc/html/Dat…

3 6

Could someone give me a small code review
by Aaron Gray 31 Dec '10

31 Dec '10

I have attached a tiny parser and lexer. This is really key to getting the lexing correct. The only thing I could not get neatly was to disallow hyphens at the end of identifiers, so have left it off until a good solution arrives. I plan to move over to monads next. Any comments welcome, Aaron

2 2