- Haskell-Cafe - Haskell.org

How to instance class with functional dependency?
by Magicloud Magiclouds 26 Nov '10

26 Nov '10

Hi, In System.Posix.IOCtl, there is such a class: class Storable d => IOControl req d | req -> d where ioctlReqSource :: req -> CInt How to instance it? I do it as: data TIOCGWINSZ = TIOCGWINSZ #starttype struct winsize #field ws_row , CUShort #field ws_col , CUShort #field ws_xpixel , CUShort #field ws_ypixel , CUShort #stoptype instance IOControl TIOCGWINSZ C'winsize where ioctlReq _ = #const TIOCGWINSZ Then I got: topIO.hs:19:0: Couldn't match expected type `C'winsize' against inferred type `Int' When using functional dependencies to combine IOControl (Foreign.C.Types.CUShort -> Foreign.C.Types.CUShort -> Foreign.C.Types.CUShort -> Foreign.C.Types.CUShort -> C'winsize) Int, arising from a use of `ioctl'' at topIO.hs:21:35-60 IOControl (Foreign.C.Types.CUShort -> Foreign.C.Types.CUShort -> Foreign.C.Types.CUShort -> Foreign.C.Types.CUShort -> C'winsize) C'winsize, arising from a use of `ioctl'' at topIO.hs:20:12-37 When generalising the type(s) for `main' -- 竹密岂妨流水过山高哪阻野云飞

2 1

Eq instance for Chan
by Mitar 25 Nov '10

25 Nov '10

Hi! Why is there no Eq instance for Chan? There is Eq for MVar so it is quite possible to define also Eq for Chan? What I would like to do is keep track how many consumers I have for each Chan so duplicating them with dupChan as necessary. So I was thinking of storing a list of Chans which already have a consumer and then duplicating it if another (and every additional) consumer would be added. But without Eq it is not possible to check if there is Chan already in a list (and I hope Eq would compare just Chan and not its content). Is there some other way to achieve this? Mitar

4 5

Precedence of "if then else"
by Sean Leather 25 Nov '10

25 Nov '10

This may be a silly, but I occasionally run into a situation where I'm wondering what the effective precedence of "if then else" is. I was finally motivated to do a few experiments, and it seems like the precedence level is reset (below 0) after "if", "then", and "else". λ> if not $ False then "here" else "x" "here" λ> if False then (++ "y") $ "x" else "here" "here" λ> if True then "here" else (++ "y") $ "x" "here" I suppose I knew this was the case with "if", but I had not given it much thought for "then" or "else". Of course, it makes sense. It's the same for "case" and many other syntactic constructs. I was wondering where this was defined in the Language Report. I had looked for prose, but found nothing. Now, I look at it again (with more motivation) and see it (clearly) in the following excerpt of the grammar (Ch. 3 Expressions): exp → infixexp :: [context =>] type (expression type signature) | infixexp infixexp → lexp qop infixexp (infix operator application) | - infixexp (prefix negation) | lexp lexp → \ apat1 … apatn -> exp (lambda abstraction, n ≥ 1) | let decls in exp (let expression) | if exp [;] then exp [;] else exp (conditional) | case exp of { alts } (case expression) ... Plainly and simply, this says that all infix expressions (infixexp) are expressions (exp), and these can be found in "if then else" and "case" among others. So, while I was looking for something regarding precedence, it is actually defined straightforwardly in the grammar. In case others happen to search for the same thing I did, maybe this will provide them with an answer of sorts. Regards, Sean

1 0

Monadic function purity
by Gregory Propf 25 Nov '10

25 Nov '10

I have a pretty basic question. I've been wondering about whether monadic functions that do NOT us IO can be pure or not. There seems to be some confusion on this topic on the web. I'm especially interested in whether they can be memoized. It seems to me that something like a function in the State monad should be pure provided the same initial state and same function arguments are present. Likewise with the list monad and most other monads in fact.

4 3

Equality constraint "synonyms"
by Hugo Pacheco 25 Nov '10

25 Nov '10

Dear Haskellers, When programming with type functions, I often find myself with a set of invariants that are frequent in my programs and can be expressed as equality constraints on the context of the functions. I wonder if there is any way in current GHC to express some kind of synonyms for equality constraints, to avoid a fixed set of constraints for a smaller expression. The natural way for type classes would be to create a superclass containing all the desired instances class (A a, B a) => C a where {} but as from http://hackage.haskell.org/trac/ghc/ticket/2715, equality constraints are currently not supported in class contexts, what might be reasonable. However, in fact I am just looking for some kind syntactic sugar. Would this be a desired feature for other people? Regards, hugo -- www.di.uminho.pt/~hpacheco

3 3

ANN: tskiplist-0.0.0
by Peter Robinson 25 Nov '10

25 Nov '10

This package [1] provides an implementation of a skip list using STM. A skip list is a probabilistic data structure with Data.Map-like operations. In contrast to a balanced tree, a skip list does not need any (expensive) rebalancing, which makes it particularly suitable for concurrent programming. You can find some documentation on how to use it here: http://darcs.monoid.at/tskiplist/dist/doc/html/tskiplist/Control-Concurrent… Feedback appreciated! Peter [1] http://hackage.haskell.org/package/tskiplist-0.0.0

1 0

Женская логика
by Vovan 25 Nov '10

25 Nov '10

Хотите понимать женщин? Знать реальные их мысли, когда они пытаются спрятать их за совершенно отвлеченными речами? На практических тренингах Академии Знакомств вы в реальных условиях научитесь быть предупредительным и понимающим мужчиной для женщин. http://samec.org.ua/?p=202 Не верите? Попробуйте, вам понравится. Желаю вам больших научных успехов.

1 0

Пикап
by Vovan 24 Nov '10

24 Nov '10

Пикап блог один из лучший информативных сайтов о в Украине, который помогает мужчинам в разы улучшить их личную и сексуальную жизнь дает ответы на вопрос как познакомиться с девушкой? , и решить множество проблем, найти ответы которые помогут Вам стать настоящими мужчинами!!!

1 0

Re: [Haskell-cafe] Confused about missing data constructor
by Antoine Latter 24 Nov '10

24 Nov '10

I meant that 'pop' and 'push' should have been written with 'modify', 'get', and 'set' instead of the raw constructor, not as a drop-in replacement. I can show you examples later if this isn't clear, unless I'm not understanding your code above. Antoine On Nov 24, 2010 10:46 AM, "Adam Miezianko" <adam(a)theorylounge.org> wrote: > On Wed, 24 Nov 2010 06:17:47 -0800, Antoine Latter <aslatter(a)gmail.com> > wrote: > >> On Tue, Nov 23, 2010 at 6:58 PM, Adam Miezianko <adam(a)theorylounge.org> >> wrote: >>> I'm working through Learn You a Haskell for Great Good [1] and getting >>> a compiler error while playing around with some of the code. I have >>> this: >>> >>> pop :: State Stack Int >>> >>> But when I try to load it into ghci I get the following errors: >>> >>> /home/admi/.pe/state.hs:6:6: Not in scope: data constructor `State' >>> >>> Now, I'm not exactly sure how to read the documentation for >>> Control.Monad.State [2] but it seems that newtype State s a = State >>> {...} defines a constructor, or am I wrong on that point too? So, >>> what am I missing here? In case it matters, I am using mtl-2.0.1.0 >>> and ghci 6.12.3. >>> >>> [1] http://learnyouahaskell.com/for-a-few-monads-more >>> [2] >>> http://cvs.haskell.org/Hugs/pages/libraries/mtl/Control-Monad-State.html >>> >> >> It looks like your documentation doesn't match the library you're using. >> >> The documentation for mtl-2.x is here: >> http://hackage.haskell.org/package/mtl > > Thanks! > > My documentation didn't have any version identification, so discovering > that there was a breaking interface change is no surprise now. Next time > hoogle doesn't return me what I want I'll first suspect it's been removed > before questioning the accuracy of the search results. > > On a somewhat related note, I also found the link you provide, but I found > hackage.haskell.org to be intermittently timing out for me this last week. > >> However you might be better served using the 'modify' function. > > Looking at the type signature of modify, I'm unsure how to use it. It > does not look like I can just drop it in as a replacement for State > (StateT, state). Could someone provide an example? The tutorials seem to > mostly use the State constructor that's now gone. > > -- > Adam Miezianko

2 1

Another Quasi-Quotation question
by jean-christophe mincke 24 Nov '10

24 Nov '10

Hello, I am still playing with template-haskell... I wonder, is there any reason why a quasiquoter cannot create haskell statements and declarations in addition to expressions and patterns? Or more generally create any legal Haskell syntax tree. I.e Suppose I would like to create a quasiquoter for the C language (please, imagine that it could be useful). I could write sth such as : [$c| int a = 6; int b = 7; int c = a +b; struct S { int x; int y} |] It could be nice to generate the appropriate haskell code: a = 6 b= 7 c= a + b data S = S { x::Int; y:Int} etc That would allow to really embed any kind of language in a more or less easy way into haskell, provided that code can be translated into legal haskell. Is there anything that prevent these features. Thank you Regards J-C

2 1