May 2008 - Haskell-Cafe

hint / ghc api and reloading modules
by Evan Laforge 01 Jun '08

01 Jun '08

I'm using "hint", but since it's basically a thin wrapper around the GHC API, this is probably a GHC api question too. Maybe this should go to cvs-ghc? Let me know and I'll go subscribe over there. It's my impression from the documentation that I should be able to load a module interpreted, make changes to it, and then reload it. This is, after all what ghci does. It's also my impression that the other imported modules should be loaded as object files, if the .hi and .o exist, since this is also what ghci does. However, if I load a module and run code like so (using hint): GHC.loadModules ["Cmd.LanguageEnviron"] GHC.setTopLevelModules ["Cmd.LanguageEnviron"] GHC.setImports ["Prelude"] cmd_func <- GHC.interpret (mangle_code text) (GHC.as :: LangType) It works fine until I change LanguageEnviron. If I make a change to a function, I don't see my changes in the output, as if the session is only getting partially reset. If I insert a syntax error, then I do see it, so it is recompiling the file in some way. However, if I *rename* the function and call it with the new name, I get a GhcException: During interactive linking, GHCi couldn't find the following symbol: ... etc. So I examined the code in hint for loadModules and the code in ghci/InteractiveUI.hs:/loadModule, and they do look like they're doing basically the same things, except a call to rts_revertCAFs, which I called too just for good measure but it didn't help (I can't find its source anywhere, but the ghci docs imply it's optional, so I suspect it's a red herring). Here's a condensed summary of what hint is doing: -- reset GHC.setContext session [] [] GHC.setTargets session [] GHC.load session GHC.LoadAllTargets -- rts_revertCAFs -- load targets <- mapM (\f -> GHC.guessTarget f Nothing) fs GHC.setTargets session targets GHC.load session GHC.LoadAllTargets -- interpret let expr_typesig = "($expr) :: xyz" expr_val <- GHC.compileExpr session expr_typesig return (GHC.Exts.unsafeCorce# expr_val :: a) -- GHC.compileExpr maybe_stuff <- hscStmt hsc_env ("let __cmCompileExpr = "++expr) ([n],[hv]) <- (unsafeCoerce# hval) :: IO [HValue] return (Just hv) and then ghci does: -- load GHC.setTargets session [] GHC.load session LoadAllTargets targets <- io (mapM (uncurry GHC.guessTarget) files') GHC.setTargets session targets GHC.load session LoadAllTargets rts_revertCAFs putStrLn "Ok, modules loaded: $modules" -- interpret GHC.runStmt session stmt step -- GHC.runStmt Just (ids, hval) <- hscStmt hsc_env' expr coerce hval to (IO [HValue]) and run it carefully So it *looks* like I'm doing basically the same thing as ghci... except obviously I'm not because ghci reloads modules without any trouble. Before I go start trying to make hint even more identical to ghci, is there anything obviously wrong here that I'm doing?

2 1

Parsec, updateState, and failure
by Dimitry Golubovsky 31 May '08

31 May '08

Hi, If a parser which updated user state fails, will such update be reverted? Suppose we have two parsers combined with <|> p = p1 <|> p2 p1 has the following: p1 = try $ do ... -- getting something from the input stream updateState (\st -> ...) -- updated state based on what gotten from the input x <- p3 -- p3 should see updated state and return something updateState (\st -> ...) -- updated state again (if p3 succeeded) return x If p3 fails, p1 fails too (second updateState will not be reached). But what will p2 (tried next) see in the user state? Will it be state after the first updateState, or will failure of p1 roll the update back? Is there any bracket- or try-catch-finally-like mechanism for Parsec? Thanks. -- Dimitry Golubovsky Anywhere on the Web

5 4

Upd: Parsec, updateState, and failure
by Dimitry Golubovsky 31 May '08

31 May '08

Hi, So far, I figured it out like this: updateState ... mbx <- (p3 >>= return . Just) <|> (return Nothing) updateState ... case mbx of Just x -> return x Nothing -> pzero but this seems a bit clumsy - is there a more elegant solution? ---------- Forwarded message ---------- From: Dimitry Golubovsky <golubovsky(a)gmail.com> Date: Sat, May 31, 2008 at 10:12 AM Subject: Parsec, updateState, and failure To: haskell <haskell-cafe(a)haskell.org> Hi, If a parser which updated user state fails, will such update be reverted? Suppose we have two parsers combined with <|> p = p1 <|> p2 p1 has the following: p1 = try $ do ... -- getting something from the input stream updateState (\st -> ...) -- updated state based on what gotten from the input x <- p3 -- p3 should see updated state and return something updateState (\st -> ...) -- updated state again (if p3 succeeded) return x If p3 fails, p1 fails too (second updateState will not be reached). But what will p2 (tried next) see in the user state? Will it be state after the first updateState, or will failure of p1 roll the update back? Is there any bracket- or try-catch-finally-like mechanism for Parsec? Thanks. -- Dimitry Golubovsky Anywhere on the Web -- Dimitry Golubovsky Anywhere on the Web

1 0

is there some book about haskell and data struct and alg?
by smellcode 31 May '08

31 May '08

is there some book about haskell and data struct and alg? i mean data struct and algorithm in haskell i am freshman i want to study haskell with data struct and alg -- Hu Jinpu (nickname: smellcode) Web developer email =~ /(^hujinpu)\(a)(gmail)\.(com)$/ or email.gsub!(/hujinpu/, 'smellcode') http://hujinpu.net

9 10

Re: [Haskell] Programme terminates silently
by Bertram Felgenhauer 31 May '08

31 May '08

(redirecting to haskell-cafe. the haskell mailing list is primarily for announcements) (sorry, Thomas, if you get this message twice) Thomas Bevan wrote: > I've written the programme below. > > The lircLoop should never terminate. Unfortunately it does. Worse, no error > messages are generated. > Not even the final line "Closing down" is printed. > > How is this possible? I couldn't find the Hmpf.Fork module you're using, so I couldn't test this. However, one possibility is that program receives a signal that is neither handled or ignored. This is the case for SIGPIPE [1]. You can strace your program to find out whether that's the case. System.Posix.Process provides facilities for installing your own signal handlers. HTH, Bertram [1] http://hackage.haskell.org/trac/ghc/ticket/1619

1 0

Re: [Haskell-cafe] Data.Tree.Zipper in the standard libraries
by Krasimir Angelov 31 May '08

31 May '08

An updated Zipper version is uploaded here: http://www.haskell.org/haskellwiki/User_talk:Kr.angelov It doesn't use State monad anymore and it returns Maybe. This seems to be the common preference, is it? Feel free to vote against. Should we change Data.Map also? There is another proposal for changes in findMin/findMax so it is better to make this two breaking changes together rather than in a later release. It also allows to pass whole new subtree in insert{Left/Right/Down}. On Sat, May 24, 2008 at 3:41 AM, <ajb(a)spamcop.net> wrote: > G'day all. > > Quoting Don Stewart <dons(a)galois.com>: > >> This is Haskell, we should use Maybe. > > This is Haskell, more abstract is good. > > I do agree, though, that Monad is arguably the wrong abstraction. Something > like this would arguably be better: > > class (Functor f) => Fail f where > return :: a -> f a > fail :: String -> f a > > (And yes, the String is arguably important; Maybe doesn't give you that.) > > Cheers, > Andrew Bromage > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe >

3 3

Fwd: Re: [Haskell-cafe] [Newbie question] -- Looping stdin until condition is met
by Martin Blais 31 May '08

31 May '08

FYI. On Sat, 31 May 2008 02:11:13 +0200, "Daniel Fischer" <daniel.is.fischer(a)web.de> said: > Am Samstag, 31. Mai 2008 02:28 schrieb Martin Blais: > > Allright, this is a definitely a newbie question. > > > > I'm learning Haskell and going through the exercises in the > > beautiful Hutton book, and one of them requires for me to > > write a loop that queries a line from the user (stdin), > > looping until the user enters a valid integer (at least > > that's how I want to implement the interface to the > > exercise). I have tried tons of variants and modifications > > of code, and I can't find the way to implement this. Here is > > what my emacs buffer is at now:: > > > > import Text.Read > > import System.IO > > import qualified Control.Exception as C > > > > getNum :: IO Int > > getNum = do line <- (hGetLine stdin) > > x <- (C.catch (do return (read line :: Int)) (\e -> getNum)) > > return x > > All you need is a little strictness, > x <- (C.catch (return $! read line :: Int) (\e -> getNum)) > works. Another option is using evaluate instead of return. > The problem is that (read line :: Int) is not evaluated until it is > needed, > that is when it's going to be printed, but then it's too late to catch > the > exception. > > Some general remarks: > hGetLine stdin === getLine > do x <- action > return x > is the same as > action > > > > > main = do x <- getNum > > putStr ((show x) ++ "\n") > > print x > > > > Now, I've tried the Prelude's catch, the Control.Exception > > catch, I've tried moving it at the top of getnum, I've tried > > without catch, I've tried a ton of other shtuff, so much > > that I'm starting to think that Emacs is going to run out of > > electrons soon. I asked some half-newbie friends who are > > insanely enthousiastic about Haskell and they can't do it > > either (I'm starting to think that those enthousiastic > > friends are dating a beautiful girl with 3 PhDs, but she has > > a 2-inch thick green and gooey wart smack on her nose and > > they're so blindly in love that they can't admit that she > > does). I've asked some university profs and they sidetrack > > my question by saying I shouldn't do I/O so early. Can > > anyone here restore my faith in the Haskell section of > > humanity? > > > > 1. How do I catch the exception that is raised from "read"? > > By forcing the evaluation. > > > > 2. Where do I find the appropriate information I need in > > order to fix this? I'm probably just not searching in the > > right place. (Yes, I've seen the GHC docs, and it doesn't > > help, maybe I'm missing some background info.) > > Get used to lazy evaluation, the Hutton book should contain a chapter > about > that. > > > > > 3. Please do not tell me I should solve the problem > > differently. Here is the problem I'm trying to solve, and > > nothing else: > > > > "Write a program that reads a line from the user, > > looping the query until the line contains a valid > > integer." > > > > It shouldn't be too hard i think. The best answer would be a > > two-liner code example that'll make me feel even more stupid > > than I already do. > > > > Thanks in advance. > > >

1 0

I/O without monads, using an event loop
by Robin Green 30 May '08

30 May '08

I have been thinking about to what extent you could cleanly do I/O without explicit use of the I/O monad, and without uniqueness types (which are the main alternative to monads in pure functional programming, and are used in the Concurrent Clean programming language). Suppose you have a main event handler function, like this: eventMain :: (Event, SystemState AppState) -> (Command, SystemState AppState) This function could be called over and over in an event loop, until an EndProgram command was received, and the event loop would itself do all the actual I/O (the SystemStates are only in-memory representations of some part of the system state, plus the application's own state). Things like disk I/O could be done with commands which generate events when complete. Interprocess communication could be done in the same way. Then eventMain, and everything called by it, would be referentially-transparent, and yet non-monadic. You could of course build higher-level stuff on top of that. On the other hand, it's quite stateful, because anything you need to remember between events need to be recorded, either in the SystemState or externally (e.g. in a file). I suppose this is the most important disadvantage? Is there any published work or code using this approach, or something like it, in a pure functional language? I'm primarily interested in embedded system and desktop UIs, rather than say web-based systems, although both would be interesting. -- Robin

7 7

Lazy lists simulated by unboxed mutable arrays
by Henning Thielemann 30 May '08

30 May '08

I wonder whether the following idea has been investigated or implemented somewhere: We could simulate a list with strict elements, i.e. data StrictList a = Elem !a (StrictList a) | End by an unboxed array with a cursor to the next element to be evaluated and a function that generates the next element. Whenever an element with an index beyond the cursor is requested, sufficiently many new elements are written to the array and the cursor is advanced. This would still allow the nice tricks for recursive Fibonacci sequence definition. This will obviously save memory, but can we also expect that it is noticeably faster than (StrictList a) ?

6 18

test driving cabal install... cabal install and normal install (runghc Setup) don't mix... two package.conf files...
by Thomas Hartman 30 May '08

30 May '08

After a little drama with zlib, I managed to get cabal-install installed. I then attempted to do cabal install HAppS-Server since this is a module with a lot of dependencies, and in rapid development flux, so perenially painful for me to install. The result is that I managed to install everything up to HAppS-State, which I think is the last dependency, but then seemed to hang indefinitely in the middle of installing HAppS-Server at the end. OK, I thought, then perhaps I can do normal runghc Setup.hs after downloading and unzipping the tar from http://hackage.haskell.org/cgi-bin/hackage-scripts/package/HAppS-Server-0.9… However, this resulted in error thartman@thartman-laptop:~/haskellInstalls/smallInstalls/HAppS-Server-0.9.2.1>runghc Setup.hs configure ... Setup.hs: At least the following dependencies are missing: HAppS-Data >=0.9.2... Strange, because I had just installed that module via cabal-install, and I could load it in ghci with :m +HappS.Data. I then ran ghc-pkg and got this strange result that my packages were broken into two different files. Is this by design? ghc-pkg list /usr/local/lib/ghc-6.8.2/package.conf: Cabal-1.2.3.0, Cabal-1.3.11, Cabal-1.5.2, DeepArrow-0.2, .... /home/thartman/.ghc/i386-linux-6.8.2/package.conf: HAppS-Data-0.9.2.1, HAppS-IxSet-0.9.2.1, HAppS-State-0.9.2.1, I am curious if anybody else is able to install HAppS-Server using cabal install, and whether they can shed any light on the other isuses I raised. Thomas.

5 8