August 2003 - Haskell

Re: Need some help please.
by b.i.mills＠massey.ac.nz 27 Aug '03

27 Aug '03

What is Nick Name doing? Trolling? I don't approve of people getting homework done on this emailing list, you learn more from doing, but baiting newbies is worse. You want to give them the idea that Haskell is complex and abtruse? > I need to define a function called safetail; it's like tail > except that this one maps the empty list to the empty list. > It has to be defined using the following: Ok, so logicaly, we are saying that we want (safeTail x) to be (tail x) if x is not [], but for x==[] (normally an error) we want (safeTail []) == []. Just write this down in Haskell ... safeCond x = if x==[] then [] else tail x Guards are just another way of producing conditional expressions safeGuard x | x==[] = [] | True = tail x And pattern matching likewise, for this simple example safePat [] = [] safePat x = tail x If you want to test the result ... tester 0 safe = [(safe [])] tester n safe = (safe [1 .. n]) : (tester (n-1) safe) call it as (tester 10 safePat) for example, to get a list of some of the outputs from the safePat routine. Regards, Bruce.

1 0

RE: Type class problem
by oleg＠pobox.com 27 Aug '03

27 Aug '03

Simon Peyton-Jones wrote: > > instance (Show (f (Mu f))) => Show (Mu f) where > > show (In x) = show x > > > > instance Show (N (Mu N)) where > > show Z = "Z" > > show (S k) = "S "++show k > But again, it's fragile isn't it? You are dicing with non-termination > if you have instance declarations where the stuff before the '=>' is more > complicated than the stuff after the '=>'. A less carefully written > instance for Show (N (Mu N)) would make the type checker loop. That's true that care is required. However, the same care is required when we want to write 'fix' or a set of mutually recursive type aliases. We have to introduce a real constructor somewhere to break the loop. I'm curious if you could post an example of > You are dicing with non-termination > if you have instance declarations where the stuff before the '=>' is more > complicated than the stuff after the '=>'. I'm wondering if the general method of avoiding non-termination can be made to work in these more complex cases. Incidentally, the constraint solver stack overflow problem can be turned to our advantage. The typechecker's exhausting the stack should be considered a failure to match the instance -- and so the typechecker should mark the current instance inappropriate and try another one, if any. I have just implemented this strategy in a logic programming system I'm working on, and it works beautifully. I can find counter-examples of Haskell instances violating functional dependencies no matter in which order the instances are given. Thank you!

1 0

AW: Need some help please
by Markus.Schnell＠infineon.com 27 Aug '03

27 Aug '03

If you post some of your tries, we could give you specific tips. -- Markus Schnell -----Ursprüngliche Nachricht----- Von: Rajiv Patel [mailto:raj_patel@cwcom.net] Gesendet: Mittwoch, 27. August 2003 16:09 An: haskell(a)haskell.org Betreff: Need some help please Hi, I need to define a function called safetail; it's like tail except that this one maps the empty list to the empty list. It has to be defined using the following: 1. a conditional expression 2. guarded equation 3. pattern matching It's probably quite simple, but I am struggling on this, so help would be appreciated. Thanks Raj

1 0

PVM-free GpH
by mgross＠dorsai.org 27 Aug '03

27 Aug '03

I am pleased to be able to announce that the Metis Project at Brooklyn College has been able to modify the gum RTS for ghc-5.02.3 so that it runs without using PVM. We have verified operation under the Mosix clustering system and are currently working on implementing a version that we will be able to use for parallel execution on our Solaris machines. It is to be expected that our code will run without modification on any Linux SMP. While our code is still alpha (among other things, it generates numerous annoying but harmless informational messages we hope will be soon be silenced) and we lack the resources to provide technical support, I will be pleased to provide copies of our CVS repository to those who ask (sorry, no anonymous CVS access at present). We would greatly appreciate assistance in improving the code, and particularly in locating and fixing the bug whose symptoms are either a failed assertion in the garbage collector, or a message about improper packet packing. Anyone who would like a copy of the CVS archive should send e-mail to magross(a)its.brooklyn.cuny.edu. I will reply as soon as possible to arrange for a transfer. Murray Gross Brookly College, City University of New York

1 0

POSIX enviornment standard?
by John Meacham 26 Aug '03

26 Aug '03

I was wondering if perhaps there was a need for a standard (or at least documentation for all the major compilers) on how one can expect the haskell runtime to interact with the OS enviornment? in particular, questions which I can think of which I would like answered (for POSIX systems, I imagine Win32 would have a seperate set of concerns) are: SIGPIPE, does it kill the process or is it ignored and an IO exception is raised? I would prefer to be able to count on the second behavior. is setlocale(LC_ALL,"") called before main to initialize locale? is SIGALRM safe to use or is it reserved for the haskell implementation? is the full path of the program stored in $_ ? is argv[0] modified in place? (ghc appears to do so brokenly on linux at the moment) There are probably more, Perhaps some sort of document listing such concerns for a given enviornment stating that compilers should document their behavior (or specify a specific behavior when appropriate) and perhaps set some #defines based on them? John -- --------------------------------------------------------------------------- John Meacham - California Institute of Technology, Alum. - john(a)foo.net ---------------------------------------------------------------------------

1 0

ANN: Konka doc
by Jan Skibinski 23 Aug '03

23 Aug '03

Hi All, You might be interested in Konka documentation, which is available either in HTML format: http://www.lun.pl/konka/doc/konka-doc.html or as a raw source -- from which either HTML or a View format can be generated in two seconds: http://www.lun.pl/konka/files/konka-doc.zip http://www.lun.pl/konka/files/konka-doc.tar.gz The preface to Konka doc explains how to generate the View format on the fly, so if you get tired of scrolling the big 360K HTML version (with images and tabularized 25 Remote API functions) download one of the above plus the Rebol/View interpreter from http://www.rebol.com This small 530 executable installs in seconds and deinstalls even easier - by erasing. It is worthy to have - even just for the sake of making and reading the Konka documentation. Konka is a portable software for local and remote generation, transformation, viewing and "snapshoting" of 3D scenes. Any language, which speaks TCP can interface to Konka via its set of well documented Remote API functions. The documentation has one chapter on Haskell and one on O'Haskell and Timber. Two simple simbot examples demonstrate how to use Konka API in GHCI. The overview of Haskell-Konka API is provided. Low level clients are presented for Haskell and O'Haskell. Regards, Jan _____________________________________________________________ Finally, a free email address your friends will remember! Become you(a)EmailAccount.com at http://www.emailaccount.com/

1 0

AW: container for different types, avoiding boiler plate
by Markus.Schnell＠infineon.com 22 Aug '03

22 Aug '03

> As you can see, functions ggender and gnumber are so trivial that > hardly need to be declared at all. You can just use fetch > directly. There is no need for Maybe. In fact, if you attempt to fetch > a type that you didn't put into the attrs, you get a _compile-time_ > error. Before I use your code, I have to understand it - this may take a while... :) But I'm not sure if it's suitable for what I'm doing, as the attributes get inserted at run time, and not all of them appear everywhere. So Maybe is quite natural here. Markus

2 1

ANNOUNCE: Alex version 2.0
by Simon Marlow 22 Aug '03

22 Aug '03

I am pleased to announce version 2.0 of Alex, the lexical analyser generator for Haskell. Based on Chris Dornan's original version of Alex, this new version contains some significant changes: - A revised syntax, more closely matching that of lex and Happy. - A rewritten back-end, producing smaller, faster, lexers. - A more flexible programmer interface, allowing simple tokenisers to be produced with very little code, and allowing monads to be threaded through the lexer. Read the release notes for the full story: http://www.haskell.org/alex/doc/html/about.html#RELNOTES This version has been tested on GHC: I have a development version of GHC using an Alex lexer, which beats the orignal hand-optimised lexer. Distributions can be obtained from Alex's new home: http://www.haskell.org/alex/ Alex is now distributed under a BSD-style license. Share and enjoy, Simon Marlow: simonmar(a)microsoft.com Chris Dornan: cdornan(a)arm.com Isaac Jones: ijones(a)syntaxpolice.org

3 3

Re: container for different types, avoiding boiler plate
by oleg＠pobox.com 22 Aug '03

22 Aug '03

You might also wish to look at the typed heaps, which have been discussed here on many occasions. Given the constant piece of code in the appendix (which does *not* depend on user types and can be put into a separate, constant module) you can write > data Gender = Masc | Fem | Neutr deriving (Show) > data Number = First | Second | Third deriving (Show) > > attrs = Cons Masc $ Cons Second $ Nil > > ggender attrs = (fetch undefined attrs) :: Gender > gnumber attrs = (fetch undefined attrs) :: Number As you can see, functions ggender and gnumber are so trivial that hardly need to be declared at all. You can just use fetch directly. There is no need for Maybe. In fact, if you attempt to fetch a type that you didn't put into the attrs, you get a _compile-time_ error. *Main> ggender attrs Masc *Main> gnumber attrs Second Updates are allowed as well: *Main> let attrs' = alter Neutr attrs in (ggender attrs', gnumber attrs') (Neutr,Second) Appendix: the constant part of the code. Flags used: -fglasgow-exts -fallow-undecidable-instances -fallow-overlapping-instances > data Nil t r = Nil > data Cons t r = Cons t r > > class PList ntype vtype cdrtype where > cdr:: ntype vtype cdrtype -> cdrtype > empty:: ntype vtype cdrtype -> Bool > value:: ntype vtype cdrtype -> vtype > > instance PList Nil vtype cdrtype where > empty = const True > > instance (PList n v r) => PList Cons v' (n v r) where > empty = const False > value (Cons v r) = v > cdr (Cons v r) = r > > class TypeSeq t s where > type_index:: t -> s -> Int > fetch:: t -> s -> t > alter:: t -> s -> s > > instance (PList Cons t r) => TypeSeq t (Cons t r) where > type_index _ _ = 0 > fetch _ (Cons v _) = v > alter newv (Cons v r) = Cons newv r > > instance (PList Cons t' r', TypeSeq t r') => TypeSeq t (Cons t' r') where > type_index v s = 1 + (type_index v $ cdr s) > fetch v s = fetch v $ cdr s > alter newv (Cons v' r') = Cons v' $ alter newv r'

1 0

Re: overlapping instances and functional dependencies
by oleg＠pobox.com 21 Aug '03

21 Aug '03

Wolfgang Jeltsch has observed: > I have this code: > class C a b c | a b -> c where > f :: a -> b -> c > > instance C a b c => C a (x,y,b) c where > f a (_,_,b) = f a b > > instance C a (a,c,b) c where > f _ (_,c,_) = c > ghci -fglasgow-exts -fallow-overlapping-instances compiles it without > complaint but hugs -98 +o says: > ERROR "ClassProblem.hs":7 - Instances are not consistent with > dependencies > *** This instance : C a (a,b,c) b > *** Conflicts with : C a (b,c,d) e > *** For class : C a b c > *** Under dependency : a b -> c > Can anyone tell me what the reason for this is and, maybe, how to avoid > these problems with Hugs? I believe it's helpful to think about this problem in terms of logical programming, and proceed in phases. Phase 1. The compiler sees the instances of a class C and compiles them into a database of instances. In Prolog terms, given the instances instance C a b c => C a (x,y,b) c instance C a (a,c,b) c the compiler asserts c(A,tuple(A,C,B),C,dict2). c(A,tuple(X,Y,B),C,dict1) :- c(A,B,C,_). In Haskell, lower-case type identifiers are variables and upper-case one are type constants. In Prolog, it's the other way around: lower-cased identifiers are constants and upper-cased are variables. I hope that won't cause too much confusion. I happen to have SWI-Prolog handy, in which I assert the above two facts. Phase 2. Suppose the compiler sees a type with a class constraint ``C Int (Int,Char,Bool) t'' The compiler can determine the appropriate instance by issuing a query: ?- bagof(Dict,c(int,tuple(int,char,bool),T,Dict),Dicts). Dict = _G158 T = char Dicts = [dict2] Yes In our case, only one instance matched, and we found the corresponding dictionary. If Dicts were a list of two or more dictionaries, we would have needed to choose the most specific, if any. Phase 1a. Before we get to Phase 2 however, we have to take care of one more problem. The class declaration specified a dependency: class C a b c | a b -> c which all members of that class must satisfy. The instances (which define the membership) must be consistent with the dependency "a b -> c". The consistency would be violated if there were two types "C a b c1" and "C a b c2" such that c1 /== c2. Thus we can check the consistency of the instances by searching for a counter-example: ?- c(A,B,C1,_), c(A,B,C2,_), \+ C1 == C2. A = _G151 B = tuple(_G151, _G153, tuple(_G151, _G158, _G302)) C1 = _G153 C2 = _G158 Yes Alas, in this case we found a counter-example: types C a (a,c,(a,c',d)) c C a (a,c,(a,c',d)) c' with c' /= c match some instance but contradict the functional dependency. Therefore, the compiler complaints.

3 2