June 2010 - Haskell-Cafe

design question/font metrics
by Gery Debongnie 03 Jun '10

03 Jun '10

Dear Haskell-Cafe list, Since I am learning Haskell, I decided to try to do a real program, to practice and give me some experience. I choose to do a basic typesetting program (like TeX/LaTeX). Now, obviously, such a program needs to manipulate font objects, and in particular, font metrics. Basically, I would like some function like this : stringWidth :: Font -> String -> Double charWidth :: Font -> Char -> Double which take some appropriate font type, string or char, and gives me a width (also height) in some given unit. I'd like it to take into account appropriate kerning. Question 1 : Now, I may be missing something, but a quick search on Google/Hackage didn't yield anything. Is there an existing package, or any nice and simple way to deal with that problem? Question 2 : In case the answer to the previous question is negative, I guess I will need to do it myself, the old-fashioned way. Now, this is where I have a design question. The thing is that font metrics information are encoded into some file (either .tfm or .pl files, for TeX fonts), so my function stringWidth will need to read a file, which is, gasp, an IO operation. Therefore, its type will be something like stringWidth :: Font -> String -> IO Double and it will "infect" pretty much everything else. It doesn't sound really appealing, so I'd like your opinions on this subject. To start the conversation, here are the solutions that I have in mind, so far : 1. well, do nothing, and let IO infect everything. 2. use unsafePerformIO to get rid of IO. (I am weary to do this, since I am a newbie, and I don't really want to start using that function everywhere... However, in this case, it looks kinda legitimate...) 3. Perform a reading of the font metrics file in the main program, put the results into some FontMetrics object, and give that to stringWidth :: FontMetrics -> Font -> String -> Double. Pros : allow me to avoid problems of solution 1 and 2, cons : it doesn't sound right for the main program to have to do this, it should be confined into a Font module. Any other ideas, suggestions? Gery D.

5 5

Parselib sample
by C K Kashyap 03 Jun '10

03 Jun '10

Hi, Is there a not-so-trivial parser implementation with Parselib? Parser for a "C" like language would be good. I searched and found Haskell++ -> http://www.cs.chalmers.se/~rjmh/Software/h++.html However, I'd prefer to look at a parser for a "C" like language. -- Regards, Kashyap

5 9

cabal update doesn't work on Windows XP
by John O'Donnell 03 Jun '10

03 Jun '10

Hi, Has anyone managed to get cabal to run on Windows? I'm running Windows XP. There are lots of other things, including old version of Haskell tools, installed, but it's too costly to try installing cabal on a fresh machine image. Whenever I do cabal update, the response is "Unsuccessful HTTP code 403". That looks like the hackage server isn't responding, but other software tools (browsers, etc) on the computer can download the hackage database; the only program that can't get it is cabal. This problem occurs with all cabal 1.6, 1.8, with downloaded executables, with versions I've compiled, with the version in the Haskell Platform. Needless to say, these problems don't occur on Linux. Best wishes, John

2 1

Wire GUI
by Andrew Coppin 03 Jun '10

03 Jun '10

I'm looking at a project which involves a GUI where you can insert components and wire up connections between them. Obviously the details of what the components are and what code gets executed for them is domain-specific, however the general idea of graphically wiring things together is quite generic. Has anybody already put together a Haskell library for doing this kind of thing?

6 7

How to build an "Indicator Type" for a type class?
by Steffen Schuldenzucker 03 Jun '10

03 Jun '10

Dear Cafe, let: > data True > data False > > class C a (arbitrary instances for C may follow) Now, how to obtain an "Indicator Type" for C, i.e. a type IndC that is defined via a type family / fundep / ... , so that IndC a = True forall a which are instances of C IndC a = False for all other a. I've collected some failed approaches here[1]. My key problem is that if I define (in the 3rd try): > instance (C a) => IndC3 a True , it does *not* mean "Define this instance for all a which are an instance of C", but "Define the instance IndC3 a True for all types a, but it's not gonna work if a is not an instance of C". Does anyone have another idea? Background: After having implemented type-level lists[2] and a quicksort on them[3], I'd like to have type-level sets. In their most simple implementation, sets are just (unsorted) lists like this: > data Nil > data Cons a b > class Elem x l (instances for Elem so that Elem x l iff x is an element of the list l) Now I want: > type family Insert x s :: * Insert x s = s forall (x, s) with (Elem x s) Insert x s = Cons x s for all other (x, s). Thanks a lot! Steffen [1] http://hpaste.org/fastcgi/hpaste.fcgi/view?id=25832#a25832 [2] Kiselyov, Peyton-Jones, Shan: "Fun with type functions" http://research.microsoft.com/en-us/um/people/simonpj/papers/assoc-types/fu… [3] I rewrote this algorithm using type families instead of fundeps: http://www.haskell.org/haskellwiki/Type_arithmetic#An_Advanced_Example_:_Ty…

2 2

Re: [web-devel] Re: [Haskell-cafe] Google Summer of Code: BlazeHTML RFC
by Jasper Van der Jeugt 03 Jun '10

03 Jun '10

On Thu, May 27, 2010 at 8:31 PM, Matt Parker <moonmaster9000(a)gmail.com> wrote: > will it be possible to easily interleave IO values into the HTML? like > instead of the [1,2,3] > > ul $ forM_ [1, 2, 3] (li . string . show) > > what if it was a function that returned IO [1,2,3] (maybe 1,2,3 came out of > a database). will the forM_ handle that OK? Depends on how you write the code. More precisely, it depends on in which monad the forM_ takes place. Also, I would not recommend interleaving IO values into the HTML, because the HTML is pure -- you usually don't want to mix that with IO. I'd rather write something like: numbers <- getFromDataBase return $ ul $ forM_ numbers (li . string . show) Or it could be possible to store your templates in a separate modules, and they would all have a pure type signature: templateX :: Arg1 -> Arg2 -> Html a I think this would be the preferred way, because it allows a separation between the pure View code and the IO-interleaved Controller code in a web application. Kind regards, Jasper Van der Jeugt On Thu, May 27, 2010 at 8:31 PM, Matt Parker <moonmaster9000(a)gmail.com> wrote: > will it be possible to easily interleave IO values into the HTML? like > instead of the [1,2,3] > > ul $ forM_ [1, 2, 3] (li . string . show) > > what if it was a function that returned IO [1,2,3] (maybe 1,2,3 came out of > a database). will the forM_ handle that OK? > > > On Thu, May 27, 2010 at 12:19 PM, Matt Parker <moonmaster9000(a)gmail.com> > wrote: >> >> Blaze HTML looks wonderful to me (especially with the do notation), and >> better then Hamlet for one reason: writing idiomatic haskell for stuff like >> this: >> >> ul $ forM_ [1, 2, 3] (li . string . show) >> >> instead of the Hamlet style: >> >> %ul >> $forall children.person child >> %li $child$ >> >> which, i don't know, for some reason, made me feel like i was back in PHP >> land. >> >> -matt >> >> >> On Thu, May 27, 2010 at 7:08 AM, Alberto G. Corona <agocorona(a)gmail.com> >> wrote: >>> >>> As a user, I have too many HTML generators, a few of them with Ajax and >>> none with server-side event handling (like ASPX or JSPX). Ajax is >>> complicated but server side event handling is what I really miss because it >>> is simple from the user point of view, my ervents could be handled in >>> haskell code rather than in javaScript and I implicitly could use the >>> advantages of dinamic HTML and Ajax without the need to know them at all. >>> Imagine a dynamic Web application with 100% haskell code made with >>> dynamic widgets created by third party developers. >>> So, anyone want to create a HTML templating system with server side event >>> handling? It is not terribly hard to do. (I refer to ASP.NET documentation >>> or the JavaServer Faces framework). >>> By the way, I vote for XML templating or else, combinator templating that >>> produce XHML templating because it can be handled by a future graphical IDE. >>> >>> 2010/5/27 Jasper Van der Jeugt <jaspervdj(a)gmail.com> >>>> >>>> Hey Bas, >>>> >>>> > How about also providing an enumerator back-end? >>>> > >>>> > http://hackage.haskell.org/packages/archive/iteratee/0.3.5/doc/html/Data-It… >>>> > >>>> > Then your library can integrate more easily with the snap framework: >>>> > http://snapframework.com >>>> >>>> Sure, I can do that. But I already tested integration with the snap >>>> framework, the best path here seems to call the `writeLBS` function >>>> from the snap framework on the `L.ByteString` that BlazeHtml produces >>>> (`writeLBS` internally uses an enumerator). >>>> >>>> Kind regards, >>>> Jasper Van der Jeugt >>>> >>>> On Thu, May 27, 2010 at 10:38 AM, Bas van Dijk <v.dijk.bas(a)gmail.com> >>>> wrote: >>>> > Q14: Do you see any problems with respect to integrating BlazeHtml in >>>> > your favourite web-framework/server? >>>> > >>>> > How about also providing an enumerator back-end? >>>> > >>>> > http://hackage.haskell.org/packages/archive/iteratee/0.3.5/doc/html/Data-It… >>>> > >>>> > Then your library can integrate more easily with the snap framework: >>>> > http://snapframework.com >>>> > >>>> > Regards, >>>> > >>>> > Bas >>>> > >>>> _______________________________________________ >>>> Haskell-Cafe mailing list >>>> Haskell-Cafe(a)haskell.org >>>> http://www.haskell.org/mailman/listinfo/haskell-cafe >>> >>> >>> _______________________________________________ >>> web-devel mailing list >>> web-devel(a)haskell.org >>> http://www.haskell.org/mailman/listinfo/web-devel >>> >> > >

1 0

Happy GLR parser creation problem
by Paul Jensen 02 Jun '10

02 Jun '10

[Apologies if this is the wrong place to ask -- there appears to be no Happy-specific list.] I'm interested in experimenting with using a GLR parser to parse some data that could be interpreted in different ways. I'm trying to use Happy's GLR mode but I am encountering an error building one of the examples (specifically http://darcs.haskell.org/happy/examples/glr/expr-eval/Expr.y) [Mac OS X 10.6.3] rhoda:expr-eval psj$ ghc --version The Glorious Glasgow Haskell Compilation System, version 6.12.1 rhoda:expr-eval psj$ happy --version Happy Version 1.18.4 Copyright (c) 1993-1996 Andy Gill, Simon Marlow (c) 1997-2005 Simon Marlow rhoda:expr-eval psj$ happy --glr --ghc Expr.y shift/reduce conflicts: 9 happy: src/ProduceGLRCode.lhs:190:11-76: Irrefutable pattern failed for pattern (pre, drop_me : post) It's the same error someone else encountered last year, described at http://www.haskell.org/pipermail/haskell-cafe/2009-July/063770.html . Does anyone have any insights into this? If Happy's GLR support isn't working just now, are there any other parser generators that produce GLR parsers that anyone would recommend having a look at? Many thanks, psj

2 1

Re: [web-devel] Google Summer of Code: BlazeHTML RFC
by Matt Parker 02 Jun '10

02 Jun '10

> Q3: Which HTML version would you preferably use? > HTML 5. google summer of code should be about pushing the new and exciting.

7 8

Resource module
by Arie Peterson 02 Jun '10

02 Jun '10

Dear list, For some time, I have maintained a small private module centred around the following type of "resource": > newtype Resource cap m = Resource { with :: forall a. (cap -> m a) -> m a } Interpretation: @Resource cap m@ is a resource, providing a "capability" of type @cap@, which does administration (like opening and closing handles) in the monad @m@. I would like to use a standard module instead of this howe-grown one, so I can publish things that depend on it. Does anyone know a package that provides this? Things I tried: - The type is suspiciously similar to the continuation monad transformer; in fact @Resource cap m@ is isomorphic to @forall t. ContT t m cap@. However, I can't use this latter type directly, because I would like functions on Resources like the instance > instance (Monoid cap) => Monoid (Resource cap m) , which is not possible without a newtype wrapper. - I tried using the "regions" package, since it has a very similar purpose, but this seems impossible: some resources I could not express in the form required for its class 'Resource' (methods 'open' and 'close'). If something like this is not yet around, I'll upload my version, but I'd like to reuse, if possible. Regards, Arie

2 6

ANN: threads-0.2
by Bas van Dijk 02 Jun '10

02 Jun '10

Hello, I just released threads-0.2. The library which lets you fork threads and wait for their result. INSTALL: $ cabal update $ cabal install threads API DOCS: http://hackage.haskell.org/package/threads-0.2 DEVELOPMENT: darcs get http://code.haskell.org/~basvandijk/code/threads CHANGES: * I removed the regular 'Control.Concurrent.ThreadId' out of our own 'ThreadId α' type and renamed the type to 'Result α': data ThreadId α = ThreadId { result ∷ TMVar (Either SomeException α) , threadId ∷ Control.Concurrent.ThreadId } to: newtype Result α = Result { unResult ∷ TMVar (Either SomeException α) } * The various fork functions now return the regular 'ThreadId' paired with a 'Result α' value: forkIO ∷ IO α → IO (ThreadId α) to: forkIO ∷ IO α → IO (ThreadId, Result α) * The various wait functions now just receive a 'Result α' value: wait ∷ ThreadId α → IO (Either SomeException α) to: wait ∷ Result α → IO (Either SomeException α) * I removed: throwTo ∷ Exception e ⇒ ThreadId α → e → IO () and: killThread ∷ ThreadId α → IO () because they aren't needed anymore since we have direct access to the regular 'ThreadId'. * When you build the package with GHC the GHC specific 'forkOnIO' function is additionally build in: forkOnIO ∷ Int → IO α → IO (ThreadId, Result α) * For the rest I added documentation and made some internal changes like a simplified implementation of ThreadGroups and made small stylistic changes all over the place. FUTURE PLANS: Inspired by the 'parIO' combinator from the 'async' package, I'm thinking about adding a module which offers similar functionality: data ParGroup α new ∷ IO (ParGroup α) forkIO ∷ ParGroup α → IO α → IO ThreadId wait ∷ ParGroup α → IO (Maybe α) The idea is that when you have multiple IO computations, that each compute the same answer but differ in how fast they compute that answer, you can compute them in parallel using 'forkIO'. Finally 'wait' blocks until one of the threads has computed an answer and will return Just that answer. When all of the threads fail to compute an answer (throw an exception) wait returns Nothing. Regards, Bas

1 0