- Haskell-Cafe - Haskell.org

Hide the data constructor for type family instance
by David Sorokin 29 Jan '16

29 Jan '16

Hi Cafe, When instantiating a type class with type family, I see that haddock shows the data constructor for the type family as visible. This is what I would like to avoid. I would like to hide the details of the data constructor in the documentation. An example is stated below: -- | An implementation of the 'FCFS' queue strategy. instance QueueStrategy BrIO FCFS where -- | A queue used by the 'FCFS' strategy. newtype StrategyQueue BrIO FCFS a = FCFSQueue (LL.DoubleLinkedList BrIO a) newStrategyQueue s = fmap FCFSQueue LL.newList strategyQueueNull (FCFSQueue q) = LL.listNull q Here the FCFSQueue data constructor is visible together with all its contents in the documentation! I would like it would be hidden completely. At least, I would like to hide the contents. Is it possible to do? Thanks, David

2 1

How does -N allot cores?
by Hong Yang 29 Jan '16

29 Jan '16

Hi Cafe, On an 1P system, there are 8 cores. Does -N8 send one thread to each core no matter how loaded each core is? Assuming 6 cores are busy, will -N8 send threads only to the two idle cores? If -N8 sends threads to eight cores no matter what, is there any function that detects number of idle (or usage less than certain percentage) cores? Suppose a 4P system has 32 cores (4 processors with each processor having 8 cores). Assuming the whole system is idle, how will -N32 work? Thanks, Hong

2 1

FICS client in Haskell
by Tilmann 29 Jan '16

29 Jan '16

Hi, I've been working on a FICS (Free Internet Chess Server) Client for a bit over a year now and it is finally in a presentable state. It would be great if you would like to have a look at the source code and let me know what you think! When I started with this project I was just starting to learn Haskell and I would like to know if the code is accessible/readable/idiomatic, if the overall organization seems reasonable as well or if you might have any other suggestion about what I could improve. (ie: I haven't decided on what logging framework to use.) The code is on github: https://github.com/tgass/macbeth I thank you all very much in advance! Tilmann https://github.com/tgass/macbeth/blob/master/src/Macbeth/Fics/FicsConnectio… Opens a telnet connection to freechess.org. Using conduit and attoparsec messages from the server are parsed to FicsMessages and put into a Chan. https://github.com/tgass/macbeth/blob/master/src/Macbeth/Fics/FicsMessage.hs The domain model The UI is using wx widgets (wxHaskell). Each wx-frame gets a copy of Chan FicsMessage and updates the UI when new Messages are available, ie here: https://github.com/tgass/macbeth/blob/master/src/Macbeth/Wx/ToolBox.hs https://github.com/tgass/macbeth/blob/master/src/Macbeth/Wx/Game.hs

1 0

When are MVars better than STM?
by Thomas Koster 28 Jan '16

28 Jan '16

Hi friends, Using Criterion, I have been running benchmarks to measure the relative performance of STM and MVars for some simple transactions that I expect will be typical in my application. I am using GHC 7.10.2 and libraries as at Stackage LTS 3.2. I have found that STM is faster than MVars in all my benchmarks, without exception. This seems to go against accepted wisdom [1][2][3]. I have not included my source code here to save space, but if you suspect that I am using MVars incorrectly, just say so and I will post my source code separately. I have two questions: 1. When are MVars faster than STM? If the answer is "never", then when are MVars "better" than STM? (Choose your own definition of "better".) 2. When given two capabilities (+RTS -N2), MVars are suddenly an order of magnitude slower than with just one capability. Why? For those who want details: My benchmark forks four Haskell threads. Each thread repeats a transaction that increments a shared counter many, many times. These transactions must be serialized. The counter is therefore highly contended. One version uses an MVar to store the counter in the obvious way. The other version uses a TVar instead. By the way, simply using "atomic-primops" to increment the counter won't do because the increment operation is actually a mock substitute for a more complex operation. I use the counter for my benchmarks because the real operation needs much more memory and I don't want the additional, unpredictable GC cost to affect my measurements. Typical measurements are: 1 capability, using MVar: 37.30 ms 1 capability, using TVar: 24.88 ms 2 capabilities, using MVar: 1.564 s 2 capabilities, using TVar: 80.09 ms 4 capabilities, using MVar: 2.890 s 4 capabilities, using TVar: 207.8 ms Notice that the MVar version suddenly slows by an order of magnitude when run with more than one capability. Why is this so? (This is question 2.) Despite the absolute time elapsed, I realize that the CPU usage characteristics of the two versions are also quite different. I realize that the MVar version interlocks the four threads so that only one capability is ever busy at a time, irrespective of the number of capabilities available, whereas the STM version allows up to four capabilities to be busy at once. However, I believe that the additional parallel transactions in the STM version would be mostly wasted, destined to be retried. Unless I am mistaken, this assumption appears to be consistent with the observation that the STM version with -N1 is the fastest of all. Despite all this wasted work by the thundering herd, the total CPU time (i.e. my power bill) for the STM version is still less than for the MVar version, because the MVar version so dramatically slow. Paradoxically, MVars seem to be the wrong tool for this job. So when are MVars faster than STM? (This is question 1.) [1] https://stackoverflow.com/questions/15439966/when-why-use-an-mvar-over-a-tv… [2] https://www.reddit.com/r/haskell/comments/39ef3y/ioref_vs_mvar_vs_tvar_vs_t… [3] https://mail.haskell.org/pipermail/haskell-cafe/2014-January/112158.html Thanks, Thomas Koster

5 9

Hackage hoq?
by Vasili I. Galchin 28 Jan '16

28 Jan '16

Hello, I successfully built hoq. When I run th hoq executable, I get ">" prompts but am not sure what to do. I read some of the code and I tried putting .hoq on the command line, That didn't work. Has anyone tried hoq?? Vasily

1 0

[Ann] A new, fun, easy book/tute on Haskell
by Julian Leviston 27 Jan '16

27 Jan '16

[Ann] A new, fun, easy book/tute on Haskell http://www.happylearnhaskelltutorial.com Just released this two days ago, roughly. We've released 7 out of 20 sections on the site so far, and the book is up on leanpub ($8 or $4 - $20 set your own price), at about 95% done. We don't want to put it to 100% until we've had at least 20 or more technical people read it and give us feedback, and until we've done a thorough revision for corrections. We felt like there's a need for an updated education resource that's both fun and for super beginners who have never programmed before. Essentially a slightly less drawing-focussed and slightly more technically advanced version of the 1980's Usborne Computer Guides http://www.usborne.com/catalogue/feature-page/computer-and-coding-books.aspx something that appeals to people who are more visually focussed as we are. Our plan is to keep expanding it into more volumes if there's enough interest and paid support. We're also very interested in hearing critical criticism and or praise! :) Here is a rough breakdown of our take on approaching education: 1. Tackling reading first, then slowly introducing writing when enough examples have been seen to increase confidence, because reading & writing are separate skills 2. Many small, fun examples for each thing to keep interest high: fun examples helps with motivation 3. Gradual, partial introduction of topics, in context at first: graded, less to take in at once, using the writing phase to soldify understanding, will possibly add a revision phase later 4. Not so much theory introduced before the practical has been introduced (examples-first), which gives a concrete context for the theory 5. Pictures. Some of these are visual aides, which are useful as explanations: a picture tells 1000 words, so they say 6. No assumption of previous programming experience. Almost every other guide available assume some programming 7. Smaller sections because completing small things gives a real sense of achievement, which increases motivation 8. Will be tackling how to deconstruct problems using both top down and bottom up approaches: most guides don't tackle this in a simple or basics-first way 9. It's free to read and online, so able to be discussed in public - some guides are, others aren't 10. Not so math focussed including "mathy jargon". We keep away from terms like catamorphism, lambda calculus, monad, until we need them, and these are/will be only introduced when appropriate amounts of concrete practical knowledge are present through repeated exposure to examples, so that it's obvious what is meant Thanks all! Julian - (of GetContented)

1 0

.cabal: API-compatible library versions
by Imants Cekusins 27 Jan '16

27 Jan '16

Stack may already do something like this, I don't know. Anyway, here is an idea. Currently .cabal lists version or version range. What if package info included lists of API-compatible versions of this package? Let's say one library requires v10 of packageA. Another library requires v15 of the same packageA. This is a hypothetical scenario :-P However v10, v15 are API-compatible as far as package maintainer knows. Package info specifies v10 and v15 as API-compatible too. What if it were possible to issue: cabal install packageA .. and then if v10, v15 are compatible, v15 is returned without warnings. If v10 and v15 are not compatible, cabal would warn. Basically instead of library clients needing to test multiple library versions or requiring 1 exact version, deps could be specified as 1 version (with which development took place) and cabal (with hints from package maintainers) would figure this out. It might even be possible to fine-tune it to check modules & symbols actually used by library consumer app. ?

3 6

Local types
by M Farkas-Dyck 26 Jan '16

26 Jan '16

I often wish to be able to define local types and instances, so, for example: import Data.Set as Set ordNubBy :: (a -> a -> Ordering) -> [a] -> [a] ordNubBy cmp = go Set.empty where newtype T = T a instance Ord T where T x `compare` T y = cmp x y go _ [] = [] go ys (x:xs) = bool (x:) id (T x ∈ ys) $ go (Set.insert (T x) ys) xs The notion is that type and instance declarations would become legal in `let` or `where` bonds; they would effectively declare new types and instances in that scope for each use of that term; and such local types and instances could never leave the scope they were defined in, by enforcement, so consistency of instances would be preserved. I know some related work which proposes local instances [0] but none i know proposes also local types. I'm seeking some feedback here before i formalize further and potentially start making implements. Thoughts? Questions? Better ideas? Critical flaws? [0] http://okmij.org/ftp/Haskell/TypeClass.html#local

4 8

[Template Haskell]
by Dominik Bollmann 26 Jan '16

26 Jan '16

Hi all, I'm just getting my feet wet with template haskell, and I tried to write a tmap function which maps a function over the ith component of an n-tuple (which uses a slightly different approach than the given version on the TH wiki): -- | Selects the ith component of an n-tuple tsel :: Int -> Int -> ExpQ -- n-tuple a -> a tsel i n = [| \t -> $(caseE [| t |] [alt]) |] where alt = match (tupP pats) body [] pats = map varP xs xs = [ mkName ("x" ++ show k) | k <- [1..n] ] body = normalB . varE $ xs !! (i-1) -- | Maps a function over the ith component of an n-tuple tmap :: Int -> Int -> ExpQ -- :: (a -> b) -> n-tuple -> n-tuple tmap i n = do f <- newName "f" t <- newName "t" lamE [varP f, varP t] $ [| let prefix = map extract [1..(i-1)] new = $f ($(tsel i n) $t) suffix = map extract [(i+1)..n] extract k = $(tsel k n) t in tupE $ prefix ++ [new] ++ suffix |] However, this code results in the following error: Sandbox.hs:26:29: Stage error: ‘k’ is bound at stage 2 but used at stage 1 … In the splice: $(tsel k n) In the Template Haskell quotation [| let prefix = map extract [1 .. (i - 1)] new = $f ($(tsel i n) ($t)) suffix = map extract [(i + 1) .. n] extract k = $(tsel k n) $t in tupE $ prefix ++ [new] ++ suffix |] Compilation failed. Could anyone explain to me what stage 2 and stage 1 refer to, and further, what the logical flaw in the above snippet is? What exactly is wrong with line `extract k = $(tsel k n) $t' ? Thanks! Dominik.

2 1

Re: [Haskell-cafe] Local types
by Johannes Waldmann 26 Jan '16

26 Jan '16

> I often wish to be able to define local types and instances ... https://mail.haskell.org/pipermail/haskell-cafe/2014-October/116291.html

3 2