March 2014 - Beginners

RPC library / generic application level protocol
by Emanuel Koczwara 17 Mar '14

17 Mar '14

Hi, I have an application in c++/qt (https://gitorious.org/micropos) Direct connection with PostgreSQL is a big security hole, so I'm trying to redesign my architecture. I want to add an application server between PostgreSQL and client/admin and I want to write that application in Haskell. The core concept of my new architecture is a custom protocol. With that, it would be possible with little effort to create web and mobile clients, possibly using different technologies. Totally custom protocol would be interesting, but writing a parser for c ++ _and_ python _and_ java (and haskell of course) will be too much for me at the moment. XML-RPC looks too heavy, so I ended up with JSON-RPC ('id' attribute is ommited): * authentication example: | > { method: "auth-login", params: { name: <name>, passwd: <passwd> } } | < { result: <status-object> } | > { method: "auth-logout", params: null } | < { result: <status-object> } * uploading/downloading binary data/files example: | > { method: "fs-upload", params: <binary-data> } | { result: <status-object> } | > { method: "fs-download", params: <binary-data-id> } | < { result: <status-object> } * low level generic database methods: | > { method: "db-create", params: { type: <type>, object: <object> } } | < { result: <status-object> } | > { method: "db-update", params: { type: <type>, object: <object> } } | < { result: <status-object> } | > { method: "db-delete", params: { type: <type>, id: <id> } } | < { result: <status-object> } | > { method: "db-list", params: | { type: <type>, | filter-by: [ <predicate-objects> ], | order-by: [ <ordering-objects> ] } } | < { result: [ <ids> ] } | > { method: "db-show", params: | { type: <type>, | ids: [ <ids> ], | fields: [ <fields-options-objects> ] } } | < { result: [ <objects> ] } | > { method: "db-list-show", params: | { type: <type>, | filter-by: [ <predicate-objects> ], | order-by: [ <ordering-objects> ], | fields: [ <fields-options-objects> ] } } | < { result: [ <objects> ] } * low level generic relations: | > { method: "db-insert", params: | { group-type: <type>, | group-id: <id>, | member-type: <type>, | members: [ <ids> ] } } | < { result: <status-object> } | > { method: "db-remove", params: | { group-type: <type>, | group-id: <id>, | member-type: <type>, | members: [ <ids> ] } } | < { result: <status-object> } Example usage (this should add a record in 'client' table): | > { method: "db-create", params: | { type: client, | object:{ name:"John Doe", address: "...", ... } } } | < { result: <status-object> } This 'parsing engine' should be generated from some kind of 'routes description' like in django, rails or yesod. For example something like this: | auth-login (user-name :: String) (user-password :: String) | auth-logout | auth-lock | auth-unlock (user-name :: String) (user-code :: String) | fs-upload (binary-data :: BinaryData) | fs-download (binary-data-id :: BinaryDataId) | db-create (type :: String) (object :: a) | db-update (type :: String) (object :: a) | db-delete (type :: String) (id :: String) The user should supply all data definitions and conversion functions to guarantee type corectness in case of a match, for example: class FromJson a where fromJson :: String -> Maybe a data Client = Client { name :: String, address :: String, ... } instance FromJason Client where fromJson = ... This looks very generic. Am I reinventing the wheel? Which package should I use for this task? msgpack-rpc looks interesting. Other packages have very low download rate. Or maybe there is some kind of generic appliaction server (but I'm looking for tcp/ip level, no http)? Any hints are very welcome. Thanks, Emanuel

1 0

Thames Valley functional meetup
by Courtney Robinson 16 Mar '14

16 Mar '14

Hello all, For anyone in the Thames valley area (UK) interested in a more local functional meetup. We've just started one. It's available at http://www.meetup.com/Thames-Valley-Functional-Programming-Meetup/ We'll be covering Haskell and a host of other functional languages in the talks. Already got the first one lined up for Haskell, Scala and OCaml (to be confirmed). See http://www.meetup.com/Thames-Valley-Functional-Programming-Meetup/events/17… Hope to see you there. -- Courtney Robinson courtney(a)crlog.info http://crlog.info 07535691628 (No private #s)

2 1

Gnome's FLOSS Outreach Program for Women and Google Summer of Code Opportunities
by Kẏra 15 Mar '14

15 Mar '14

Are there any participating organizations in OPW or GSoC providing opportunities to gain Haskell experience? It would be great to see more Haskell projects participate in this

2 3

scotty + ffi segfault
by Miro Karpis 15 Mar '14

15 Mar '14

Hi, please can you help me with following? I'm trying to combine scotty + ffi but it returns 'Segmentation fault/access violation in generated code'. The thing is, that when I run only the ffi setmodulestring function, everything works fine. When I call it from scotty I get the segfault. I have hardcoded the 'param' and 'value' parameters in setmodulestring function just to be sure that the input is correct. scotty code: main = scotty 3000 $ do middleware logStdoutDev post "/setstringtest" $ do let param = "FilePath" let value = "C:/dev/misc/haskell/services/FM" result <- liftIO $ FM.setmodulestring param value text "done" ffi code: setmodulestring :: String -> String -> IO CInt setmodulestring param value = do cParam <- newCString param cValue <- newCString value let cParamLength = fromIntegral $ length param ::CInt cValueLength = fromIntegral $ length value ::CInt setVarInArray = (-1)::CInt result <- c_setmodulestring cParam cParamLength cValue cValueLength setVarInArray return result Any comments/ideas more than appreciated. Cheers, Miro

2 2

library for lattice data structure
by James Toll 13 Mar '14

13 Mar '14

Hi, I have a general question regarding data structures, maybe functional data structures, in Haskell and I’m hoping for some advice. I would like to implement the binomial option pricing model in Haskell. https://en.wikipedia.org/wiki/Binomial_options_pricing_model Given its lattice structure, I originally thought of trying a recursive implementation, but because it’s recombining, the interior nodes would be calculated twice. Although the implementation might be elegant, it doesn’t seem inefficient. Therefore I’ve been searching through Hackage for a data structure library that might lend itself to this problem. There are a ton of libraries and as I don’t have any experience with them, I was hoping someone could provide some advice to narrow the field. Some of the packages I've found so far are: lattices package http://hackage.haskell.org/package/lattices From the name I thought this might be perfect, but the documentation is very terse and there doesn’t appear to be any other documentation so I’m not even sure how to get started with constructing a lattice with this package. matrix package https://hackage.haskell.org/package/matrix This package seems fairly straightforward, and a matrix would work (it’s probably what I would use in R), but it doesn’t seem ideal. There are other more sophisticated packages with matrices like hmatrix and repa but they seem like overkill unless this package is just slow. vector package https://hackage.haskell.org/package/vector Vectors might also work for a data structure, but a matrix just seems like it would be easier. I thought this example implementation in Haskell was intriguing but it doesn’t sound as though it’s very quick. http://www.thulasidas.com/2009-03/a-new-kind-of-binomial-tree.htm Any advice on an appropriate data structure in Haskell for the problem? The solution doesn’t necessarily have to include a package. Thanks James

4 7

Questions too basic for this mailing list?
by Kim-Ee Yeoh 11 Mar '14

11 Mar '14

One of the attributes of the haskell community that I take pride in is the balance between democracy and upholding standards. High technical standards such as correctness, ease-of-reasoning, brevity are valued, as well as high community standards such as humor, empathy, addressing the other with dignity. All the above, coupled with an openness to the world-at-large, is the kit and kaboodle of the accidental haskeller. So when the response to a newcomer boils down to "take your basic questions elsewhere", I wonder about these values and I wonder about the community's openness. I for one regret losing the opportunity of exploring very interesting questions about syntax, even elementary syntax, a common obstacle faced by many. Consider how Haskell syntax was designed in a rare confluence of events requiring tremendous buy-in (and patience, and empathy, and openness) from a superficially similar but actually a very diverse group with vigorous, opposing opinions. In an environment that turns inimical to such basic, foundational values -- whither progress? I feel the loss deeply. -- Kim-Ee

1 0

MultiParamTypeClasses confusion
by Derek McLoughlin 10 Mar '14

10 Mar '14

Hi, In Chapter 6 of "Beginning Haskell" by Apress there's a couple of classes introduced for vectors and things that can be vectorized (not related to Data.Vector) {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE MultiParamTypeClasses #-} class Vector v where distance :: v -> v -> Double instance Vector (Double, Double) where distance (a,b) (c,d) = sqrt $ (c - a) * (c - a) + (d - b) * (d - b) class Vectorizable e v where toVector :: e -> v instance Vectorizable (Double, Double) (Double, Double) where toVector = id x = 1.0 :: Double y = 10.0 :: Double While I understand how to use the "distance" function: ghci> distance (x, x) (y, y) 12.727922061357855 ... and I can see how "toVector" is used in their code ghci> distance (x, x) $ toVector (y, y) 12.727922061357855 I don't understand why this doesn't work: ghci> let z = toVector (y, y) interactive>:32:9: No instance for (Vectorizable (Double, Double) v0) arising from a use of `toVector' The type variable `v0' is ambiguous Possible fix: add a type signature that fixes these type variable(s) Note: there is a potential instance available: instance Vectorizable (Double, Double) (Double, Double) -- Defined at vector.hs:13:10 Possible fix: add an instance declaration for (Vectorizable (Double, Double) v0) In the expression: toVector (y, y) In an equation for `z': z = toVector (y, y) It seems odd that "toVector" works when used as an argument to "distance" but not when used in a let expression. Can anyone explain? Derek.

2 2

Using GHCi, import submodule that needs to import another submodule?
by James Toll 07 Mar '14

07 Mar '14

Hi, I am trying to figure out how to import one submodule from a hierarchical module in GHCi. For example, at the end of Chapter 6 of LYAHFGG (page 106-107), there's an example of a hierarchical module. Or at the bottom of this page, http://learnyouahaskell.com/modules If I recreate that module on my system and try to import, this is the output: [~/Geometry]$ ls Cube.hs Cuboid.hs Sphere.hs [~/Geometry]$ ghci GHCi, version 7.6.3: http://www.haskell.org/ghc/ :? for help Loading package ghc-prim ... linking ... done. Loading package integer-gmp ... linking ... done. Loading package base ... linking ... done. Prelude> import Geometry.Sphere <no location info>: Could not find module `Geometry.Sphere' It is not a module in the current program, or in any known package. Prelude> :set -iGeometry Prelude> import Geometry.Sphere <no location info>: Could not find module `Geometry.Sphere' It is not a module in the current program, or in any known package. Prelude> :load Sphere [1 of 1] Compiling Geometry.Sphere ( Sphere.hs, interpreted ) Ok, modules loaded: Geometry.Sphere. *Geometry.Sphere> :load Cube Cube.hs:6:18: Could not find module `Geometry.Cuboid' Use -v to see a list of the files searched for. Failed, modules loaded: none. Prelude> :load Cuboid [1 of 1] Compiling Geometry.Cuboid ( Cuboid.hs, interpreted ) Ok, modules loaded: Geometry.Cuboid. *Geometry.Cuboid> :load Cube Cube.hs:6:18: Could not find module `Geometry.Cuboid' Use -v to see a list of the files searched for. Failed, modules loaded: none. Prelude> So I can't just "import Geometry.Sphere", and I tried setting -i, but that doesn't seem to help. Instead I have to ":load Sphere". Fine, but then when I try to ":load Cube", it fails because of the line in Cube.hs that imports Geometry.Cuboid. Even if I load Geometry.Cuboid first and then try to load Geometry.Cube, it still fails. So, in GHCi, how do I successfully import a submodule that needs to import some other submodule in the module? Thanks, James

4 9

Combining IO and Either function to "EitherT e IO a"
by Nathan Hüsken 07 Mar '14

07 Mar '14

Hey, I have a function |func1 :: IO String | and another: |func2 :: String -> Either String String | and I want to combine them, giving the output of the first as the input as the second. |func3 :: IO (Either String String) func3 = do tmp <- func1 return (func2 tmp) | Ok, possible. But I rather would like a result of type "EitherT String IO String". So how can I combine these function in a smart way, to get the needed result? Thanks! Nathan

6 11

Good way of combining functions of type IO (Maybe a)
by Nathan Hüsken 04 Mar '14

04 Mar '14

Hey, I want to write a function, which is basically a concatenation of functions of type "IO (Maybe a)". When they all where of type "Maybe a", no Problem I would simple use the Maybe monad. |func :: Maybe c func = do a <- f1 b <- d2 a ... | but now they are of type "IO (Maybe a)". Is there some way of combing these in a similar smart way? Thanks! Nathan

4 4