- Haskell-Cafe - Haskell.org

Re: [Haskell-cafe] [Haskell] ANN: Portackage - a hackage portal
by Andrew Seniuk 30 Apr '12

30 Apr '12

> On Apr 27, 2012 6:18 PM, "Andrew Seniuk" <rasfar(a)gmail.com> wrote: > Oh, the idea wasn't that you would make the app, but that by providing a web > service interface you would make it easier for me or anyone else to make an > app or other alternative interface. It would also reduce bandwidth usage > over screen scraping. Yes, well as I mentioned in the part of my response you didn't quote, I agree it would be desireable, and am working on it, but my motivation was to provide people with something immediately useful. Once I get something reasonably efficient worked out, I'll look at uploading the Haskell code to Hackage, providing JSON data, etc.. Meanwhile if someone wanted to work on their own app or interface, it's not too hard to get CSV from the table. I made a mistake and the link was down for about 12 hours. It's back now, with a number of improvements and up-to-date hackage info. It is an expensive page to run, you may get "Script not responding" pop-ups depending on your hardware, but I think it's worth it. http://fremissant.net/portackage Kind Reg'ds, Andrew

2 4

ANN: hflags-0.1, a command line flags library similar to Google's gflags
by Gergely Risko 30 Apr '12

30 Apr '12

Hello, I would like to announce the release of HFlags, a command line flags library that makes it very easy to define and use flags. The API is similar to Google's gflags, here is a very simple example program: -=- #!/usr/bin/env runhaskell {-# LANGUAGE TemplateHaskell #-} import HFlags defineFlag "name" "Indiana Jones" "Who to greet." defineFlag "r:repeat" (3 + 4 :: Int) "Number of times to repeat the message." main = do remainingArgs <- $(initHFlags "Simple program v0.1") sequence_ $ replicate flags_repeat greet where greet = putStrLn $ "Hello " ++ flags_name ++ ", very nice to meet you!" -=- As you can see, we have TemplateHaskell functions to help with the definition (and initialization) of the flags, and the values themselves are pure, no need for being in the IO monad to use the value of the flags. Also, the initHFlags function automagically gathers all the flags defined anywhere in the whole program, so if a library defines flags, you don't have to mention it in the main, but the user will still be able to set those flags too. More details behind this design and more example can be found in the following post: http://blog.risko.hu/2012/04/ann-hflags-0.html Comments and criticism is welcome. Code: http://github.com/errge/hflags Examples: https://github.com/errge/hflags/tree/master/examples Hackage: http://hackage.haskell.org/package/hflags Best regards, Gergely Risko

1 0

Hong Kong Haskell hackathon - call for participation
by Matthias Fischmann 30 Apr '12

30 Apr '12

Hi everybody, if you are in or around Hong Kong in May and want to play, please check out this: http://www.haskell.org/haskellwiki/HkHac2012 Otherwise, any wider circulation or suggestion of channels / university people / companies that I should invite directly is welcome. cheers, Matthias

1 0

Printing call site for partial functions
by Michael Snoyman 30 Apr '12

30 Apr '12

I had a bug in a site of mine[1] for a few weeks, where it would just print: Prelude.head: empty list It took a long time to track down the problem, as it came from some other library I was depending on. Eventually I tracked it down, reported it, and the problem was fixed the next day. The weak link in this chain was identifying what package was to blame. I noticed the other day that the docs for undefined[2] state: > It is expected that compilers will recognize this and insert error messages which are more appropriate to the context in which undefined appears. I'm not sure if the current output of `Prelude.undefined` is meant to fit the criterion of "appropriate to the context", but in my opinion it does not. I'd like to propose a bit of a strawman to improve the situation for not just `undefined`, but any arbitrary partial function (including `head`). What we really want is to define some function like: headContext :: String -> [a] -> a headContext context [] = error $ "Prelude.head: empty list (" ++ context ++ ")" headContext _ (x:_) = x And then have the compiler automatically include (optional) package name, module name, and line number where `headContext` was called. How about we borrow a bit from rewrite rules, and have a pragma such as: {-# WITH_CONTEXT head headContext #-} If the compiler supports the feature, and is able to provide context in the given location, it will call `headContext` with an appropriate `String`. Otherwise, normal `head` would be called. I'm sure there are many better ways to approach the problem, and I can't speak to the complexity of implementation within GHC. I *can* say, however, that this would have saved me a lot of time in the example I gave above, and I'd bet many Haskellers have similar stories. This could be a huge debugging win across the board. Michael [1] It was actually yesodweb.com, and resulted in empty Javascript files being generated, thus disabling some features of the site. [2] http://hackage.haskell.org/packages/archive/base/4.5.0.0/doc/html/Prelude.h…

8 10

ANN: unfoldable-0.4.0
by Sjoerd Visscher 29 Apr '12

29 Apr '12

I am pleased to announce the 5th version of the unfoldable package. (This is the first announcement, you didn't miss anything.) http://hackage.haskell.org/package/unfoldable-0.4.0 Just as there's a Foldable class, there should also be an Unfoldable class. This package provides one: class Unfoldable t where unfold :: Unfolder f => f a -> f (t a) Writing instances of Unfoldable is similar to writing Traversable instances. For example, given a data type data Tree a = Empty | Leaf a | Node (Tree a) a (Tree a) a suitable instance would be instance Unfoldable Tree where unfold fa = choose [ pure Empty , Leaf <$> fa , Node <$> unfold fa <*> fa <*> unfold fa ] The choose function comes from the Unfolder class: class Applicative f => Unfolder f where choose :: [f x] -> f x (If f is an Alternative instance, choose is simply Data.Foldable.asum.) Different unfolders provide different ways of generating values, for example: - Random values - Enumeration of all values (depth-first or breadth-first) - Convert from a list - An implementation of QuickCheck's arbitrary should also be possible (still working on that) Some examples can be found in the examples directory in the github repo: https://github.com/sjoerdvisscher/unfoldable Ideas and comments are welcome! greetings, Sjoerd

4 7

Can we write a parser with both name resolution and error handling by using Happy?
by Tsuyoshi Ito 29 Apr '12

29 Apr '12

Hello, I have a question about Happy, the parser generator. Section 2.1 of Happy User Guide explains a neat trick to perform the name resolution inside a parser by turning the semantic values of nonterminal symbols into Reader monads. http://www.haskell.org/happy/doc/html/sec-using.html#sec-other-datatypes (The example uses bare functions instead of Reader monads, but this difference is insignificant.) A compilable Calc.y file is available from https://gist.github.com/2554028. The important part of the example is the following: Exp : let var '=' Exp in Exp { \p -> $6 (($2, $4 p) : p) } | Exp1 { $1 } Exp1 : ... Term : ... Factor : int { \p -> $1 } | var { \p -> case lookup $1 p of Nothing -> error "no var" Just i -> i } | '(' Exp ')' { $2 } The example code calls `error` on the failure in the name resolution. Although this is fine as an example, it is not good in an actual parser because that means that I have to run the parser in IO monad to catch a parse error. I would like to handle this error just like a parse error by using a monadic parser as explained in Section 2.5.1. http://www.haskell.org/happy/doc/html/sec-monads.html#sec-exception However, I cannot make this work. I tried the following approaches: 1. Just combine these two techniques. That is, write `%monad { E } { thenE } { returnE }` in .y file, where `E t` is equivalent to `Either String t` as defined in Section 2.5.1, and make the semantic value a Reader monad. This does not achieve what I want: now Happy expects that the semantic expression of the form {% ... } will have type `E (Env -> t)` (or `E (Reader Env t)`), but I want to write an expression of type `Env -> E t` (or `ReaderT Env E t`) because whether name resolusion succeeds or not depends on the environment. See https://gist.github.com/2554041 for a code (which does not compile). 2. Write `%monad { P } { thenP } { returnP }`, where `P t` is equivalent to `Env -> E t` or `ReaderT Env E t` and `thenP` and `returnP` are defined accordingly. Now I cannot write the semantic expression for the production rule `Exp : let var '=' Exp in Exp`, because the semantic values of children ($2, $4, and $6) accessible in the `{ ... }` or `{% ... }` part is no longer Reader monads but plain values, meaning that I cannot evaluate a subexpression ($6) using a different environment from the current one. Moreover, this feels wrong because not every symbol requires the environment to evaluate. See https://gist.github.com/2554044 for a code (which does not run because it contains `undefined`). 3. Maybe I cannot use the `%monad` support for this purpose, so shall I give up using it and just use `Env -> E t` as a type of semantic values? This is cumbersome at best because now in each production rule, I have to handle the case where the semantic values of some of the children are `Nothing` (after giving an environment). But more importantly, this does not work because Happy expects the `parseError` function to have type `forall a. [Token] -> a` (which is understandable), which means that it cannot return anything like `Nothing`, defeating the purpose of the use of a monadic parser for error handling. See https://gist.github.com/2554122 for a code (which does not run). >From this, it does not seem that name resolution can be performed inside a parser if we need any error handling better than calling `error`. Is this really the case? Because the combination of parsing and name resolution is very natural (which is why the example in Section 2.1 is so appealing), it is hard to believe that it cannot be done nicely. Am I missing something? In case it matters, I am using Haskell Platform 2011.4.0.0 on Windows. (Just in case, I am aware that we can split parsing and name resolution, in which case a parser generates an abstract syntax tree (AST) with unresolved names and a name resolver converts it to an AST with resolved names. But is that a recommended way to use Happy? If so, the example in Section 2.1 is misleading at best.) Best regards, Tsuyoshi

1 0

Announce: HaTeX 3.3
by Daniel Díaz Casanueva 29 Apr '12

29 Apr '12

Greetings Haskellers! A new version of HaTeX is ready to use in Hackage: http://hackage.haskell.org/package/HaTeX-3.3 == Description == HaTeX is a heavily maintained LaTeX code parser/generator. Write LaTeX documents with all the advantages you already have in Haskell (recursion, type system, high order functions, ...), create a LaTeX backend for your own program, make analysis of LaTeX code through its Abstract Syntax Tree (AST), find another way to pretty-printing values, ... == Release notes == Please, to read the release notes go to the following blog post: http://deltadiaz.blogspot.com.es/2012/04/hatex-33-release-notes.html Thanks and good luck, Daniel Díaz.

1 0

Cabal try to install a package I already have
by Daniel Díaz Casanueva 29 Apr '12

29 Apr '12

Hi Cafe, today I opened the command prompt and typed "cabal install ftphs". Then I read: > Resolving dependencies... > Configuring network-2.3.0.5... And I started to feel confused. Because I already have installed that package with that version! I continued reading: > cabal: The package has a './configure' script. This requires a Unix > compatibility toolchain such as MinGW+MSYS or Cygwin. Yes, I knew this would happen! (thought I don't like it with a so common-used library) But I can't understand why cabal is trying to install network-2.3.0.5. If I type "ghc-pkg list" I can see precisely network-2.3.0.5 is installed. In fact, 2.3.0.5 is the version that comes with the Haskell Platform I recently installed (2011.4.0.0) as you can see here: http://code.haskell.org/haskell-platform/haskell-platform.cabal. Could someone, please, explain me this? I will be very glad. I was working with ghc-7.4 but I needed the network package so I switched to the Haskell Platform, because install network in Windows is painful. Yes, I read the Mitchell guide (found in http://neilmitchell.blogspot.com.es/2010/12/installing-haskell-network-libr…) but once I installed the library I was having compile time error with "unknown symbol" errors. I felt very odd so I switched this machine to the Haskell Platform. I hope someone could help me. Thanks in advance, Daniel Díaz.

2 3

Embed Haskell
by Rosario Borda 28 Apr '12

28 Apr '12

Hi All, Is there a simple method for embed haskell in other languages, like latex? An example in quasiquotation notation: \documentclass{article} \title{Hello World} \begin{document} \maketitle Hello world! and reversed: [haskell: putStrLn reverse "Hello, World!"] \end{document} Many thanks, :) Rosario

3 2

Automatic discovery of tests
by Graham Berks 28 Apr '12

28 Apr '12

Hi, trying to learn haskell and the associated build env. I come from a java/scala world so trying to understand the how to invoke tests. Have a cabal file that invokes a main function to execute a test. But wondering if there is any automatic discovery of tests as per junit. I have seen test-framework-th but to me that implies putting a testGroupGenerator into each module. Am i correct about that and if so is there an example of calling multiple modules tests ? Or is there a better way just to invoke specific functions prefixed with case_ prop_ etc in the entire src/test directory of the cabal build ? Thanks

2 1