April 2008 - Haskell-Cafe

Mutually recursive types?
by Ron Alford 16 Apr '08

16 Apr '08

Here's the setup: I have a series of problems that use various logical connectives. The problem is that they're not all the same. So instead of creating one giant datatype (or duplicating much code), I'd like to assemble them like toy blocks. I've boiled down an example here: data LogicalConnective a = Not a | And [a] | Or [a] data BasicGoal a = Atomic String [Term] | Empty | Logical (LogicalConnective a) deriving (Show, Eq) data PreferenceGoal1 = Basic1 PreferenceGoal1 | Prefer1 PreferenceGoal1 This works OK, but PreferenceGoal1 is a dead end. I can't combine it with other connectives. So I try: data PreferenceGoal2 a = Basic2 (PreferenceGoal2 a) | Prefer2 (PreferenceGoal2 a) And this works fine, but seems impossible to explicitly type (ie, there is nothing to substitute for 'a' in a type declaration). Or am I wrong? Also, it could be that this is just an ugly way to represent things (it does require a huge number of constructors). Any suggestions? -Ron

2 2

ANN: datapacker 1.0.0
by John Goerzen 16 Apr '08

16 Apr '08

Hi, I'm pleased to announce the first release of datapacker. datapacker is a tool to pack files into a minimum number of CDs, DVDs, or any other arbitrary bin. It groups file by size. It is designed to group files such that they fill fixed-size containers (called "bins") using the minimum number of containers. This is useful, for instance, if you want to archive a number of files to CD or DVD, and want to organize them such that you use the minimum possible number of CDs or DVDs. In many cases, datapacker executes almost instantaneously. Of particular note, the hardlink action can be used to effectively copy data into bins without having to actually copy the data at all. datapacker is a tool in the traditional Unix style; it can be used in pipes and call other tools. More information, including downloads and a full manual, can be found at: http://software.complete.org/datapacker datapacker is also available on Hackage.

4 4

Module for parsing C (pre-processed header files only)?
by Magnus Therning 16 Apr '08

16 Apr '08

Is there such a beast out there? If this were a list on some OO language I'd ask for something that created an AST and then offered an API based around a visitor pattern. The pre-processor (platform specific most likely, but could be cpphs of course) would be run prior to the tool I'm considering writing. After that I only need access to declarations of C functions and definitions of types (structs, enums, other typedefs). I had a quick look at the code in c2hs yesterday, but found it a little hard to make heads or tails of it without some guidance, so I couldn't tell whether it would be suitable. Would it? Any help or pointers appreciated. /M

3 4

Searching for possible PhD supervisors/projects
by Ivan Miljenovic 16 Apr '08

16 Apr '08

Next year, I'm thinking of going overseas (I'm currently in Australia) to do a PhD. Preferably, I'd like to do something in the area of computational combinatorics using Haskell. Does anyone know of any particular unis/supervisors I should be looking at/talking to about this? -- Ivan Lazar Miljenovic

1 0

Funny State monad dependency
by Hans Aberg 16 Apr '08

16 Apr '08

When I load the State module in Hugs, then I can define the function f below, but I do not immediately see exactly what function "return" returns. Explanation welcome. For example: > f [2..4] [6..9] [6,7,8,9,6,7,8,9,6,7,8,9] That is, it just repeats the second argument as many times as the length of the second argument. Hans Aberg -------- import Control.Monad.State f :: Monad a => a b -> a c -> a c f x y = x >>= (return y) --------

4 6

monadic debugging
by Galchin, Vasili 16 Apr '08

16 Apr '08

Hello, I have an Linux executable of my Haskell library and test case. I see there are several debuggers, e.g. Buddha, Hat, etc. Which debugger is currently preferred for monadic (imperative) code? Thanks. Vasili

5 4

Installing HaXml
by rodrigo.bonifacio 15 Apr '08

15 Apr '08

Hi all, I've tried to install HaXml as explained in the documentation: > runhaskell Setup.hs configure However, I get as response: dyld: Library not loaded: GNUreadline.framework/Versions/A/GNUreadline Referenced from: /usr/local/bin/runhaskell Reason: image not found Trace/BPT trap Any idea about this problem? Some comments about my configuration: Mac/OS Darwin 8.11.1 GHC-6.6 Hugs - Version September 2006 Thanks in advance, Rodrigo.

3 2

Re: [Haskell-cafe] Module for parsing C (pre-processed header files only)?
by Magnus Therning 15 Apr '08

15 Apr '08

On Tue, Apr 15, 2008 at 4:12 PM, Brian Sniffen <bts(a)evenmere.org> wrote: > 2008/4/15 Magnus Therning <magnus(a)therning.org>: > > Is there such a beast out there? > > Well, there's CIL (http://manju.cs.berkeley.edu/cil/) an OCaml > library attacking the same problem. It has a very positive > reputation. Interesting answer on a Haskell list ;) I've been toying with the idea of looking at OCaml though, so this might be just the opportunity I was waiting for... /M

1 0

Help understanding sharing
by Patrick Surry 15 Apr '08

15 Apr '08

I'm new to Haskell and trying to get a better understanding of sharing (and ultimately memoization). I've read SOE and various of the tutorials, as well as browsing around the wiki and old mailing lists. Most of the examples of memoization seem to revolve around Fibonacci, and are based either on the fact that a list defined within the function will get shared between calls, or on doing some 'unsafeIO' (which I haven't dug too deeply into.) I've read various discussions that explain why function calls are generally not automatically memoized (e.g. f x gets recalculated rather than looked up based on the prior result). The rationale for that (big space leak and no guarantee of improved performance) makes sense. (Though I did like one poster's suggestion of a compiler pragma that hints that a particular function should be memoized.) I've seen other discussions that suggest that lists are always shared while in scope (so the fibs trick works). But is that just a feature of the standard compilers, or is it somewhere mandated in the Hakell spec (I don't see anything obvious in the Haskell Report tho haven't read it cover to cover)? The wiki page http://www.haskell.org/haskellwiki/Performance/Strictness says laziness == non-strictness + sharing but again nowhere gives a set of rules that guarantees what will be shared and what won't. I was hoping I might find it here: http://www.haskell.org/haskellwiki/Sharing but no such luck. Or are there no guarantees and you just have to know how your particular compiler works?? Cheers, Patrick ________________________________________________________________________ DISCLAIMER: This e-mail is intended only for the addressee named above. As this e-mail may contain confidential or privileged information, if you are not the named addressee, you are not authorised to retain, read, copy or disseminate this message or any part of it. If you received this email in error, please notify the sender and delete the message from your computer.

3 2

retrospective on 'seq' -> 'unsafeSeq' ?
by Henning Thielemann 15 Apr '08

15 Apr '08

When reading the section "10.3 Controlling Evaluation Order" in "History of Haskell" I thought that the example that justified the 'seq' to be unrestricted polymorphic was mainly a debugging problem. I wondered if the better solution would have been to provide an 'unsafeSeq' which has no type restriction but must be absent from production code just like 'trace'. With 'unsafeSeq' one could nicely play around until time and space behaviour is as wanted and then one can replace all 'unsafeSeq's by 'seq' while adding type constraints accordingly. (Analogously there could be an unsafeShow that allows showing offending values in an 'error' without adding a Show constraint to the type signature.)

6 5