- Haskell-Cafe - Haskell.org

Re: [Haskell-cafe] [Newbie] Why or why not haskell ?
by Donn Cave 11 Dec '05

11 Dec '05

Quoth Tomasz Zielonka <tomasz.zielonka(a)gmail.com>: ... | What I fear about the future of Haskell, is that we will have so many | libraries FFIying to C, that our programs will crash as often as | programs written in C. | | FFI is necessary, but IMO it shouldn't be used when it doesn't have | to be. I spent some time recently trying to make a trivial text analysis run fast enough to use it on large data sets. I eventually achieved a satisfactory result - twice as fast as simple awk and Python versions. Of course I used the FPS packed string library functions, and I spent a lot of time profiling. One thing I found that way, is that FPS.words iterates over the string byte by byte, in Haskell, but there's a function (breakSpace?) that calls C to find the next space - and it's a lot faster. On the other extreme, I have found it surprisingly painless to call OpenLDAP functions from Haskell, which means I don't have to write my own LDAP protocol in Haskell, but also SASL, SSL, GSSAPI and Kerberos. The effort to do that would be enormous and eventually misguided, not a good thing for Haskell programmers in any way. So -- well, I actually agree with you in principle, but I'm just thinking it depends a lot on when we decide we `have to' use FFI, and in the end the advantages will very often be too compelling. Even for basic data manipulation without Haskell overhead, like FPS. Donn Cave, donn(a)drizzle.com

1 0

Re: [Haskell-cafe] syscall, sigpause and EINTR on Mac OSX
by Donn Cave 11 Dec '05

11 Dec '05

Quoth Joel Reymont <joelr1(a)gmail.com>: ... | My program is currently stuck here. The man pages say that sigpause | will only terminate by being interrupted and EINTR will be the errno. | EINTR is signal 2, the same one that I'm trapping and the one sent | when ^C is pressed. I sure haven't been able to follow all of the ensuing discussion on this, but in case it helps -- EINTR is not a signal, it's an error status from a function that was interrupted by a signal. Just going on this error, the signal may have been SIGINT - or any other signal, such as for example SIGALRM, which could be generated by a timer function (see alarm, setitimer), or SIGCHLD, which could be generated on exit of a child process under certain circumstances. Etc. Obviously the use of signals in this way (timers, for example) has the potential to make your application unsound. It's particularly insidious when it occurs in library functions. Now if you actually have observed that your SIGINT handler was entered at this point, then please ignore me. Donn Cave, donn(a)drizzle.com

2 1

Detecting Cycles in Datastructures
by Tom Hawkins 11 Dec '05

11 Dec '05

In a pure language, is it possible to detect cycles in recursive data structures? For example, is it possible to determine that "cyclic" has a loop? ... data Expr = Constant Int | Addition Expr Expr cyclic :: Expr cyclic = Addition (Constant 1) cyclic Or phased differently, is it possible to make "Expr" an instance of "Eq" such that cyclic == cyclic is smart enough to avoid a recursive decent? -Tom

14 32

GHCi and threads, was Connecting to a running process (REPL)
by Rohan Drape 10 Dec '05

10 Dec '05

On Tue Dec 6 15:01:45 EST 2005 Tomasz Zielonka wrote: > Unfortunately it seems that forkIO'ed threads are freezed when GHCi is > waiting for command-line input. I bet it would be possible to let > the threads work in the background. I think the current behaviour is > caused by using readline, which is a foreign library. Or it is by > design? I also think this is unfortunate, and would like to know if this can be worked around (ie. I could happily live without readline if that is the issue and it can be switched off). Regards, Rohan

1 0

HXT : need to retrieve a part of a document by a XPath expression
by Simon Odou 10 Dec '05

10 Dec '05

Hi everyone, I am using the HXT (Haskell XML Toolbox) library to parse XML documents. And I would like to be able to retrieve a part of a document by a XPath expression. I saw the "getXPath" function which nearly does the work but not exactly: if you have the xpath "//grandfather/father/son", calling getXPath on the "grandfather" node gives the "son" subtree; if you have the xpath "father/son", calling getXPath on the "father" node gives the "son" subtree. I would like to get the "son" subtree when I call getXPath-like on the son. Something like: if you have the xpath "father/son" or "//grandfather/father/son", calling getXPath-like on the "son" node gives the "son" subtree. I understand why getXPath is done like that. But is there an easy way to get the previous result? Thanks -- Simon Odou simon.odou(a)gmail.com

1 0

Can't Haskell catch up with Clean's uniqueness typing?
by haskell-cafe.mail.zooloo＠xoxy.net 10 Dec '05

10 Dec '05

Hi all, being occupied with learning both languages, I'm getting curious if Haskell couldn't achieve most of the performance gains resulting from uniqueness typing in Clean by *automatically* determining the reference count of arguments wherever possible and subsequently allowing them to be physically replaced immediately by (the corresponding part of) the function's result. Are there any principal obstacles, or *could* this be done, or *is* this even done already, e. g. in ghc? Regards, zooloo -- No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.1.362 / Virus Database: 267.13.12/192 - Release Date: 05.12.2005

18 31

What is a Boxed Array?
by John Velman 09 Dec '05

09 Dec '05

I've tried google and google scholar, wikipedia, and planetMath. Can't find a description. Can someone point me to a freely available reference? Thanks, John Velman

5 5

mapOfmap & composition
by raptor 09 Dec '05

09 Dec '05

hi, Is there a function like "map (map f) list" i.e. map of map ? And second why when I try do do function composition i.e. : newf = map . f But this works : newf x = map (f x) thanx

4 3

Monad applied in Java
by Ben.Yu＠combined.com 09 Dec '05

09 Dec '05

My apology to those who don't care about Java. Please trash this note. I recently published two open source Java projects that are based on the Monad computation model learned from Haskell community. 1. A port for Parsec. This library is called Jparsec. It implements monadic parser combinator in Java. Not nearly as concise and clean because of the limitation of the Java language, but it implements the same idea. And it looks like the only monadic parser combinator lib in Java so far, according to my google homework. ;--) url: http://docs.codehaus.org/display/JPARSEC/home 2. A light-weight Dependency Injection container that's built around the Monad model. If you see <sequence id="mybean"> <bean var="a1" class="A1"/> <bean var="a2" class="A2"/> <method class="X" name="create" args="$a1,$a2"/> </sequence> You'll know what I'm talking about --- do-notation. With a bunch of "bind" calls under the hood, it is equivalent as saying: do a1 = bean "A1" a2 = bean "A2" method "X" "create" [a1,a2] url: http://docs.codehaus.org/display/YAN/home This is the only ioc container that uses Monad model in Java. Just FYI. Any comment is welcome. Ben.

1 0

preffered 'map'
by raptor 09 Dec '05

09 Dec '05

hi, I imported : import Data.Map as Map but now anywhere when I want ot use "map" it complains for name clashes, so I have to specifiy Prelude.map all the time. Is there a way to specify that i mean Prelude not Data 'map' (but not fqn) I use Hugs, 'cause error messages are more understandable. tia.

5 4

how to nicely implement phantom type coersion?
by David Roundy 09 Dec '05

09 Dec '05

Hello all, I've been working on a sort of static consistency-checking framework for darcs patches using GADTs and phantom existential types, and am looking for a bit of advice on whether I might be able to create a class to avoid some tedious coding overhead. The basic idea is that a patch will have type (Patch a b) where "a" and "b" are phantom types. Sequential patches will have identical ending and beginning types. So that a sequential pair, for example, can be written as data Sequential a c where Sequential :: Patch a b -> Patch b c -> Sequential a c The trickiness is that we need to be able to check for equality of two patches, and if they are truly equal, then we know that their ending states are also equal. We do this with a couple of operators: (=\/=) :: Patch a b -> Patch a c -> Maybe (EqContext b c) (=/\=) :: Patch a z -> Patch b z -> Maybe (EqContext a b) data EqContext a b = EqContext { coerce_start :: Patch a z -> Patch b z, coerce_end :: Patch z a -> Patch z b, backwards_coerce_start :: Patch b z -> Patch a z, backwards_coerce_end :: Patch z b -> Patch z a } where we use the EqContext to encapsulate unsafeCoerce so that it can only be used safely. The problem is that it's tedious figuring out whether to use coerce_start or coerce_end, or backwards_coerce_end, etc. Of course, the huge advantage is that you can't use these functions to write buggy code (at least in the sense of breaking the static enforcement of patch ordering). So I'm wondering if there might be some sort of multi-parameter class one could define that would ease the use of an EqContext. For example I can imagine a (somewhat lame) class class EqTypes a b where safe_coerce_end :: Patch x a -> Patch x b safe_coerce_start :: Patch x a -> Patch x b safe_coerce_start = invert $ safe_coerce_end $ invert p (note that invert :: Patch a b -> Patch b a) then we could perhaps write some template haskell to define a couple of instances of EqTypes from a data type EqContext... except that I suspect that this is isn't possible, since the EqContext only exists at runtime, so unless one could define classes dynamically at runtime I'm out of luck. So is there some other approach that I can use for easily coercing phantom types based on runtime checks? Any suggestions? What I'd really love would be to be able to create a function that is polymorphic in kind rather than in type such as (inventing a rather poor syntax off the top of my head): data EqContext a b = EqContext { safe_coerce :: f(a,b) -> f(b,a) } where f(a,b) is a function of two types that returns a type, so the value of f(a,b) might be (Patch a b) or (Patch x b) or something like that. But I'm not sure if this is possible in Haskell, and if it is, then it definitely requires some sort of tricky extension that I'm not familiar with... -- David Roundy http://www.darcs.net

7 7

RE: [Haskell-cafe] STM reference?
by Simon Peyton-Jones 09 Dec '05

09 Dec '05

GHC now has a Wiki http://cvs.haskell.org/trac/ghc/wiki, so I've started a new page to give pointers to material on concurrent programming using GHC. (The link is near the bottom.) Please add to it. Simon | -----Original Message----- | From: haskell-cafe-bounces(a)haskell.org [mailto:haskell-cafe-bounces@haskell.org] On Behalf Of | Bulat Ziganshin | Sent: 09 December 2005 07:07 | To: Scherrer, Chad | Cc: haskell-cafe(a)haskell.org | Subject: Re: [Haskell-cafe] STM reference? | | Hello Chad, | | Friday, December 09, 2005, 3:09:29 AM, you wrote: | | SC> Can anyone suggest some references to learn how to program using GHC | SC> with threads? I've read a little bit about STM, but it's still pretty | SC> mysterious to me. Is this the best approach to take? I've never used | SC> threads in any language, but monads are fairly comfortable for me. | | STM is just additional instrument, you can completely avoid it (as we | had do until 6.4 release). Below is a list of papers related to | imperative programming in Haskell. first contains pragmatic | introduction in this world, including all concurrency stuffs, second | is about an real high-concurrency application designed by GHC team. | other papers is of more theoretical (and more strict) character | | Tackling the awkward squad: monadic input/output, concurrency, exceptions, and foreign-language | calls in Haskell | [http://research.microsoft.com/Users/simonpj/papers/marktoberdorf/markto berdorf.ps.gz] | Writing High-Performance Server Applications in Haskell, Case Study: A Haskell Web Server, | [http://www.haskell.org/~simonmar/papers/web-server.ps.gz] | [http://www.haskell.org/ghc/docs/papers/concurrent-haskell.ps.gz] | [http://www.haskell.org/~simonmar/papers/conc-ffi.pdf] | The Concurrent Haskell Foreign Function Interface | [http://www.haskell.org/ghc/docs/papers/threads.ps.gz] | Asynchronous Exceptions in Haskell [http://www.haskell.org/~simonmar/papers/async.ps.gz] | [http://research.microsoft.com/~simonpj/papers/stm/stm.pdf] | [http://www.haskell.org/ghc/docs/papers/except_ps.gz] | Imperative Functional Programming [http://www.haskell.org/ghc/docs/papers/imperative.ps.gz] | | | | -- | Best regards, | Bulat mailto:bulatz@HotPOP.com | | | | _______________________________________________ | Haskell-Cafe mailing list | Haskell-Cafe(a)haskell.org | http://www.haskell.org/mailman/listinfo/haskell-cafe

1 0

STM reference?
by Scherrer, Chad 09 Dec '05

09 Dec '05

Hi, Can anyone suggest some references to learn how to program using GHC with threads? I've read a little bit about STM, but it's still pretty mysterious to me. Is this the best approach to take? I've never used threads in any language, but monads are fairly comfortable for me. Thanks! Chad Scherrer Computational Mathematics Group Pacific Northwest National Laboratory "Time flies like an arrow; fruit flies like a banana." -- Groucho Marx

2 1

busting up a structured binary file
by Brian McQueen 08 Dec '05

08 Dec '05

Can someone point me to a sample so I can get started? This will be my first Haskkel project, after reading about it for some months now. I need to extract text from a structed binary file. Its a local database for a commercial app of proprietary structure, though the structure has been determined. So I need to move along a number of bytes, take a few chunks of so many more bytes and grab the string and move on to the next record, dumping the contents - ultimately dumping it to XML. It'll need to run on windows so I guess I had better use Hugs. So a sample and any advice would be an excellent start for me, and would be greatly appreciated. Thanks Brian McQueen

2 1

Indenting issue with Haskell mode
by Joel Reymont 08 Dec '05

08 Dec '05

I believe the following is not indented correctly. It only happens with (x:xs) when you use nested cases, I think. case foo of [] -> case bar of [] -> return () (x:xs) -> -- http://wagerlabs.com/

2 1

RE: [Haskell] fptools mirror in darcs ready for testing
by Simon Marlow 08 Dec '05

08 Dec '05

On 07 December 2005 13:06, John Goerzen wrote: > On Wed, Dec 07, 2005 at 11:02:29AM +0000, Duncan Coutts wrote: >>> Also please note that these repos are READ ONLY for now. Nobody >>> will be accepting darcs patches until Simon (or someone) gives the >>> word. >> >> When that does happen, it'd be great if the repos specified a default >> 'darcs send' email address so that people do not have to guess where >> to send their patches. > > Yes, they will. Simon is thinking of having that be the appropriate > cvs-%2A(a)haskell.org list. Right now, this isn't happening because there > is no good place to send them. I think the cvs-* lists are fine. > Speaking of which: how long do we want to keep this interim > arrangement before switching to darcs completely? I'm not sure yet! So far I haven't used it in anger enough, though as Duncan points out it's hard to really use it until we can commit darcs changes using darcs. Certainly performance of the --partial tree seems good enough, though I don't like that I can't see the history for individual files. I can't get browsing to work using Trac: with the full darcs repository it takes too long to do anything (like 10 seconds to bring up a directory), and with the --partial one it can't browse propely, presumably because there isn't per-file history. Cheers, Simon

3 2

syntactic sugar for comonads
by Scherrer, Chad 08 Dec '05

08 Dec '05

Hi, I'm wondering about a syntactic sugar for comonads. These are still very new to me, but it seems like their usage will become much more common once manipulations are more convenient (I believe this was the case with monads and arrows, correct?). Here's an example for motivation. Consider the data type data Stream a = a :< Stream a and suppose we want to write a function movingWindow :: Stream a -> Stream [a] so that, for example, movingWindow 3 (1 :< 2 :< 3 :< ...) evaluates to [1,2,3] :< [2,3,4] :< [3,4,5] :< ... Recognizing Stream as a comonad, this is pretty easy: movingWindow n s = s =>> (take n . toList) toList :: Stream a -> [a] toList (x :< xs) = x : toList xs If the second argument of (=>>) is written in lambda form, this comes out as movingWindow n s = s =>> \x -> take n $ toList x This looks analogous to the way do is translated for monads, so a sugary alternative could be movingWindow n s = do x -< s take n $ toList x Note that the presence of "-<" tells us we're in a comonad, rather than a monad. I'm not at all stuck on this, but I think it would be good to get the ball rolling. What do you think? Chad Scherrer Computational Mathematics Group Pacific Northwest National Laboratory "Time flies like an arrow; fruit flies like a banana." -- Groucho Marx

2 1

Re: [Haskell-cafe] Can't Haskell catch up with Clean's uniqueness typing?
by Wolfgang Jeltsch 08 Dec '05

08 Dec '05

Am Donnerstag, 8. Dezember 2005 04:00 schrieb Jan-Willem Maessen: > On Dec 7, 2005, at 9:58 AM, Wolfgang Jeltsch wrote: > > Am Mittwoch, 7. Dezember 2005 14:21 schrieb Jan-Willem Maessen: > >> [...] > >> > >> The principle obstacles are the same as for any reference counting > >> scheme: It imposes more run-time overhead than GC does, unless the data > >> structures involved are large. > > > > Why? I think the point with uniqueness typing/analysis is that > > this is done at *compile-time*. > > But the email asked if there were some other way of doing this using > run-time techniques in the absence of such a type system (or at > least, that was the way I understood it, and the spirit of my answer). I thought that the original question was about using some kind of uniqueness type system at an intermediate stage during compiling. Haskell would still have no uniqueness types but the compiler would infer uniqueness types internally and use the uniqueness information it gets from this. > -Jan > [...] Best wishes, Wolfgang

2 1

Learning Haskell
by Jimmie Houchin 08 Dec '05

08 Dec '05

Hello, I am exploring the possibility of learning Haskell. I am not a professional programmer, nor do I have a CS or Math degree. I do play and program with Python, Smalltalk (Squeak), Javascript, explored Erlang some. I do not have a strong math background. Is lack of strong math background a major hindrance to learning Haskell? Also, I understand Haskell's benefits for programming larger projects. But how does it do on programming in the small? ie: using Haskell where I might have used Python for scripting? Thanks for any help and wisdom. Jimmie Houchin

14 25

String to Hash ?
by raptor 08 Dec '05

08 Dec '05

I have made a simple Split function, here is how it works : str = "k1=v1|k2=v2|k3=v3" > map (sSplit '=') (sSplit '|' str) [["k1","v1"],["k2","v2"],["k3","v3"]] Now what I want to do is to return listOfPairs (instead listOflists) so that I can give this as parameter to Data.Map.fromList and build hash-like structure. What would be the easiest way ? PS. Can anyone point me to a source/link where I can see more examples of using Hash-like structures ?

3 3

Mixing C++ and Haskell, OpenSSL thread safety, and using mmap
by Branimir Maksimovic 08 Dec '05

08 Dec '05

I wrotted some messages on fa.haskell newsgroup but then I've figured that people actually read mailing lists :) So this is digest of mine recent messages on this newsgroup. First I want to say about OpenSSL thread safety. It is not thread safe by default. Who wants to import and use OpenSLL functions with FFI, have to set locking hooks for it, or else spurious crashes with useless stack trace will result. Higher level of concurrency, more likely crash will happen. Since this hooks are called by C, it is easer to setup mutexes in C module, then call Haskell main function from it. details: http://www.openssl.org/docs/crypto/threads.html Second I've successfully linked Haskell and C++ with threaded run time. Well, I'm planning to use Haskell in C++ programs and to call Haskell functions from C++ and vice versa. So far I've tried some small test programs and it seems works ok. If I launch threads from C++ then call Haskell I have to link with -threaded flag as Haskell run time complains about entering functions unsafelly without it. Also I have another question about hs_add_root. Is this neccessary? I've tried with and without call to this function and everything seems to work same way. That means I don't notice any difference. I ask this because if for example I have lot of Haskell modules do I need to call hs_add_root for every and each one? This is just a small example that works both on linux and windows, (I have different C++ versions but Haskell module is same). Haskell makes and frees array of pointers to CStrings (char**) from argument that is CString(char*). Error checking and handling is intentionally left out for now. I would appreciate any critics about it, as I've started to learn language one month ago, but have professional experience since 92 in other languages, specially C++. -- Haskell module module MakeWords where import Foreign import Foreign.C.String import Foreign.Storable import Foreign.Ptr import Foreign.Marshal.Array foreign export ccall makeWords :: CString -> IO (Ptr CString) foreign export ccall freeWords :: Ptr CString -> IO () makeWords :: CString -> IO (Ptr CString) makeWords cs = do let lst = words $ unsafePerformIO $ peekCString cs p <- mallocArray0 $ length lst makeWords' lst p return (p) makeWords' :: [String] -> Ptr CString -> IO () makeWords' [] p = do poke p nullPtr makeWords' (s:strs) p = do poke p $ unsafePerformIO $ newCString s makeWords' strs (plusPtr p $ sizeOf p) freeWords :: Ptr CString -> IO () freeWords p = do freeWords' p free p freeWords' :: Ptr CString -> IO () freeWords' p = if nullPtr /= (unsafePerformIO $ peek p) then do free $ unsafePerformIO $ peek p freeWords' $ plusPtr p $ sizeOf p else return () -- end Haskell module // C++, windows version #include <HsFFI.h> #ifdef __GLASGOW_HASKELL__ #include "makeWords_stub.h" #endif #include <iostream> #include <ostream> #include <windows.h> using namespace std; class Mutex{ public: Mutex(){ InitializeCriticalSection(&s); } ~Mutex(){ DeleteCriticalSection(&s); } void acquire(){ EnterCriticalSection(&s); } void release(){ LeaveCriticalSection(&s); } private: Mutex(const Mutex&); Mutex& operator=(const Mutex&); CRITICAL_SECTION s; }m; DWORD WINAPI tf(LPVOID arg) { for(int i=0;i<100000;++i) { const char** p = (const char**)makeWords(arg?arg:(void*)"boring default"); const char** q = p; if(!(i%10000)) { m.acquire(); while(*q){ cout<<*q++<<' '; } cout<<endl; m.release(); } freeWords(p); } return 0; } #ifdef __GLASGOW_HASKELL__ extern "C" void __stginit_MakeWords(); #endif #ifndef NUM_THREADS #define NUM_THREADS 4 #endif int main(int argc,char* argv[]) { hs_init(&argc,&argv); #if defined(__GLASGOW_HASKELL__) && defined(STGINIT_NEDEED) hs_add_root(__stginit_MakeWords); // is this neccessary? // seems works without it, // both on linux and windows #endif DWORD t[NUM_THREADS]; HANDLE h[NUM_THREADS]; char *args[NUM_THREADS] = { "abc def", "one two three", "house on the beach", "end of the line", }; for(int i =0; i<NUM_THREADS;++i) { h[i]=CreateThread(NULL,0,tf,args[i],0,&t[i]); } DWORD waitall = WaitForMultipleObjects(NUM_THREADS,h,TRUE,INFINITE); for(int i =0; i<NUM_THREADS;++i) CloseHandle(h[i]); hs_exit(); } // end C++ If someone is interested, it is possible to use g++ to link but -u flag symbols have to be obtained by -v3 flag to ghc first. Third, can I use mmap in Haskell? I see that there is MArray whitch is base for StorableArray, but I need MMappedArray which will use mmap on linux or CreateFileMapping and MapViewOfFile on windows instead of malloc. So I would appreciate if someone knows if there is such class already, but it isn;t a problem to reuse StorableArray code and create new persistent Array based on this. Greetings, Bane. _________________________________________________________________ Express yourself instantly with MSN Messenger! Download today it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/

2 3

Re: Verbosity of imperative code
by Scherrer, Chad 07 Dec '05

07 Dec '05

> On Tue, Dec 06, 2005 at 10:58:45PM +0300, Bulat Ziganshin wrote: > > the third-priority problem is language itself. in particular, i hate > > Haskell school of imperative manipulations: > > > > x' <- readIORef x > > y' <- readIORef y > > writeIORef z (x'*y') Here's a way to make some of this less messy: class PlusEq a b m | a -> m where (+=) :: a -> b -> m () instance (Num a) => PlusEq (IORef a) a IO where xRef += y = do x <- readIORef xRef writeIORef xRef (x + y) instance (Num a) => PlusEq (IORef a) (IORef a) IO where xRef += yRef = do y <- readIORef yRef xRef += y Then instead of do x <- xRef y <- yRef writeIORef yRef (x + y) you can just say y += x. I've started on an InPlace module to do things like this in general. -Chad Scherrer

1 0

Connecting to a running process (REPL)
by Joel Reymont 07 Dec '05

07 Dec '05

Is there a good standard way of supplying a read-eval prompt in a program? I would like to a running process with something ghci-like to be able to inspect the state and possibly modify it. The running process would be heavily multi-threaded. Thanks, Joel -- http://wagerlabs.com/

4 4

What's a thread doing / Erlang-style processes / Message passing
by Joel Reymont 07 Dec '05

07 Dec '05

I'm finding myself in dire need of monitoring the Haskell runtime. More precisely, I would like to know what each that I launch is doing at a given point in time. Is there a way to obtain this information now? I would be fine with knowing if a thread is blocking on a foreign call, MVar or something like that. The question is prompted by my current project and the issues I'm facing now. I believe I minimized the amount of garbage that I'm generating by moving to unboxed arrays from lists (thanks dcoutts) but I still have memory utilization issues. I currently have each thread associated with a TChan and I'm going to try to abstract that today by creating a special type of a thread object that is associated with two mailboxes (in and out). When starting this thread you would supply the event loop to read from the inbox and another one to write to the mailbox. I would also add stats to the TChan mailboxes so that I know the number of messages pending in each mailbox and can monitor it. This mirrors my current architecture where I have each poker bot as three threads: #1 reading messages from a socket and posting to #3, #2 reading messages sent by #3 and writing to the socket, #3 reading messages sent by #1, processing them and posting to #2. I suppose I'm trying to implement Erlang-like processes where each process has a mailbox for incoming messages and can send messages to any other process. In Erlang you can also check how many messages are pending to each process, etc. I don't think implementing message passing on top of exceptions is a good idea but please correct me if I'm wrong. In Erlang you are tasked with implementing the message loop yourself and retrieve messages by using a "receive" construct where you can pattern-match on the type of message inside. It seems that custom messages would need to be implemented on top of Dynamic but is there a way to pattern-match on that? I have messages implemented like this now but is there a better abstraction? data Event = Enter | Exit | Quit | Timeout String | Connected | Disconnected | Error String | Cmd Command | Custom Dynamic -- can't pattern-match on this? deriving Show Last but not least, to be able to send messages to any thread I would need to keep those around in some sort of a table. I would need to create records and keep the thread id, the mailbox and possibly some sort of a per-thread string so that threads can update me on their doings. Do you have any suggestions? Thanks, Joel -- http://wagerlabs.com/

2 5

RE: [Haskell-cafe] Re: :t main
by Simon Peyton-Jones 07 Dec '05

07 Dec '05

I've not been following this thread, but I wanted to check: you do know about Tarmo Uustalu's stuff about comonads, don't you? http://www.cs.helsinki.fi/u/ekarttun/comonad/ summarises (link to "The essence of dataflow programming" at the bottom) Simon | -----Original Message----- | From: haskell-cafe-bounces(a)haskell.org [mailto:haskell-cafe-bounces@haskell.org] On Behalf Of | David Menendez | Sent: 06 December 2005 07:13 | To: Scherrer, Chad | Cc: haskell-cafe(a)haskell.org | Subject: RE: [Haskell-cafe] Re: :t main | | Scherrer, Chad writes: | | > > From: Cale Gibbard [mailto:cgibbard@gmail.com] | > > See: | > > http://haskell.org/pipermail/haskell-cafe/2003-January/003794.html | > > | > > The OI comonad as previously envisioned breaks referential | > > transparency. I/O just doesn't seem to be something which one | > > can easily do comonadically, since once coeval/extract is | > > applied, you're back to plain values, and there's no | > > imposition of sequencing. | > > | > > - Cale | > | > Hmm, I hadn't seen that. The asymmetry is pretty frustrating in that | > case. | > | > After poking around a bit more, I (re)discovered some discussion of | > this on Hawiki: | > http://www.haskell.org/hawiki/CoMonad | > | > After reading Dave Menendez's comments, I'm wondering... | > If we consider | > IO a = Realworld -> (Realworld, a) | > then wouldn't we dually have something like | > OI a = (Realworld, Realworld -> a)? | > | > Could this be what screws things up? Right now it seems like OI is | > acting like it has the same type as IO, with a different name. | | Aside from the product comonad and the exponent (reader) monad, I don't | think we'll find much overlap between monad and comonad functionality. | It's been said that monads describe effects that propagate outwards, | whereas comonads describe effects that propagate inwards. | | My guess is that comonadic IO would look more like dataflow programming. | -- | David Menendez <zednenem(a)psualum.com> | "In this house, we obey the laws | <http://www.eyrie.org/~zednenem> | of thermodynamics!" | _______________________________________________ | Haskell-Cafe mailing list | Haskell-Cafe(a)haskell.org | http://www.haskell.org/mailman/listinfo/haskell-cafe

2 1

New runtime crash, not sure what to make of it
by Joel Reymont 07 Dec '05

07 Dec '05

Program compiled with -threaded on Mac OSX. Crash happens when waiting in select, I think. Program received signal EXC_BAD_ACCESS, Could not access memory. Reason: KERN_PROTECTION_FAILURE at address: 0x00000000 [Switching to process 832 thread 0x2303] 0x0011ce20 in stg_enter_info () (gdb) where #0 0x0011ce20 in stg_enter_info () #1 0x00170e7c in StgRunIsImplementedInAssembler () #2 0x001247a4 in schedule () #3 0x00124468 in taskStart () #4 0x0017f800 in startProcWrapper () #5 0x9002b200 in _pthread_body () Haskell stack trace: (gdb) p16 $r22 0x130a87c: 0x130a898 0x130a878: 0x13162c <selw_info> 0x130a874: 0x2a 0x130a870: 0x1305000 0x130a86c: 0x1b6aec <sdG0_info> 0x130a868: 0x130a884 0x130a864: 0x20e174 <GHCziIOBase_bindIO_closure> 0x130a860: 0x20e17c <GHCziIOBase_returnIO_closure> 0x130a85c: 0x1315b0 <selK_info> 0x130a858: 0x130a878 0x130a854: 0x130a86c 0x130a850: 0x20e174 <GHCziIOBase_bindIO_closure> 0x130a84c: 0x10002 <s7uY_dflt+38> 0x130a848: 0x1174c0 <stg_PAP_info> 0x130a844: 0x170114 <stg_stop_thread_info> 0x130a840: 0x170208 <stg_noforceIO_info> 0x130a83c: 0x118bd0 <stg_ap_v_info> -- http://wagerlabs.com/

1 1

RE: [Haskell-cafe] STM and `orElse` on a few thousand TMVars
by Simon Marlow 07 Dec '05

07 Dec '05

On 06 December 2005 10:02, Joel Reymont wrote: > Is it efficient to wait on a few thousand TMVars? That depends... while waiting your thread consumes no CPU at all (of course), and simply waiting on all those TVars only imposes a small constant overhead on anyone updating one of the TVars (a TMVar is just a TVar). However, checking all those TVars is O(n), and this happens twice each time you wake up: once to find which TVar was modified, and once again to go back to sleep on all of them. I'm not sure what you're using all those TMVars for, but it sounds like it might be better to multiplex them all into a single channel, or something. > How would I write something like that, assuming that my TMVars were > in a list. perhaps something along the lines of foldr orElse (return ()) . map takeTMVar it depends what you want to return. Cheers, Simon

5 5

Retrieving the caught signal within a handler
by Joel Reymont 07 Dec '05

07 Dec '05

Is there a way to retrieve the signal within the signal handler? I would like to know the signal that I caught. Thanks, Joel -- http://wagerlabs.com/

3 3

Number of outstanding messages in Chan or TChan
by Joel Reymont 06 Dec '05

06 Dec '05

Is there a way to check the number of outstanding messages in a Chan or TChan? I suspect my problem is related to not reading messages fast enough and I'd like to troubleshoot this by scanning my channels once every few seconds and dumping statistics. Thanks, Joel -- http://wagerlabs.com/

3 5

RE: [Haskell-cafe] Re: :t main
by Scherrer, Chad 06 Dec '05

06 Dec '05

> From: Simon Peyton-Jones [mailto:simonpj@microsoft.com] > > I've not been following this thread, but I wanted to check: > you do know about Tarmo Uustalu's stuff about comonads, don't you? > > http://www.cs.helsinki.fi/u/ekarttun/comonad/ summarises > (link to "The essence of dataflow programming" at the bottom) > > Simon > Thanks, Simon. I've seen it, but I haven't fully digested it yet. Dataflow programming is new to me, and it's not immediately obvious how it ties in with IO in general. I'll take another look, though. Chad Scherrer Computational Mathematics Group Pacific Northwest National Laboratory "Time flies like an arrow; fruit flies like a banana." -- Groucho Marx

1 0

Re: [Haskell-cafe] Learning Haskell
by Donn Cave 06 Dec '05

06 Dec '05

Quoth Lemmih <lemmih(a)gmail.com>: ... | You might wanna have a look at Don's FastPackedString library: | http://www.cse.unsw.edu.au/~dons/fps.html Thanks, that's actually what got me thinking about trying this again. | > | I am also hoping to use it for web development. | > | > Wonder how `links' is coming along. (Hope they figured out a better | > name, anyway.) | | Got an url for the project? http://homepages.inf.ed.ac.uk/wadler/links/ Having looked it up, of course I looked at it. Please forget I asked. Donn Cave, donn(a)drizzle.com

1 0

RE: TChan implementation: Why TVarList
by Simon Peyton-Jones 06 Dec '05

06 Dec '05

The mutable cell is in the tail. A [TVar a] would be quite different. You can read about a very similar impl (based on MVars) in the original "Concurrent Haskell" paper (on my papers page) S | -----Original Message----- | From: Joel Reymont [mailto:joelr1@gmail.com] | Sent: 06 December 2005 11:28 | To: Simon Peyton-Jones | Cc: Haskell Cafe | Subject: TChan implementation: Why TVarList | | Simon, | | Why did you guys implement TChan on top of your own TVarList instead | of a regular list? | | -- | 'TChan' is an abstract type representing an unbounded FIFO channel. | data TChan a = TChan (TVar (TVarList a)) (TVar (TVarList a)) | | type TVarList a = TVar (TList a) | data TList a = TNil | TCons a (TVarList a) | | Thanks, Joel | | -- | http://wagerlabs.com/ | | | |

3 2

RE: [Haskell-cafe] What's a thread doing / Erlang-style processes /Message passing
by Simon Marlow 06 Dec '05

06 Dec '05

On 06 December 2005 10:28, Joel Reymont wrote: > I'm finding myself in dire need of monitoring the Haskell runtime. > More precisely, I would like to know what each that I launch is doing > at a given point in time. > > Is there a way to obtain this information now? I would be fine with > knowing if a thread is blocking on a foreign call, MVar or something > like that. If you compile with -debug, and run with +RTS -Ds you get a scheduler trace which might help. Cheers, Simon

1 0

RE: [Haskell-cafe] Signal 2 and -threaded
by Simon Marlow 06 Dec '05

06 Dec '05

On 06 December 2005 10:59, Joel Reymont wrote: > I'm seeing some funky behavior in my program compiled with -threaded. > It runs for a little while and then catches signal 2 and quits. This > does not happen when -threaded is not used. > > I'm compiling my library with -threaded so that is always constant. > The above applies to my main module and another module that is part > of it. > > Any ideas? I don't know any reason your program should be getting signal 2 (SIGINT, I assume), unless the program sends itself SIGINT or someone else does. Maybe on your system a process gets SIGINT when it exceeds some resource limit? Have you tried strace or equivalent (truss?)? Cheers, Simon

1 0

RE: [Haskell-cafe] Re: TChan implementation: Why TVarList
by Simon Peyton-Jones 06 Dec '05

06 Dec '05

sounds as if you need a priority queue, so you can say "give me the message with the earliest time" but otherwise yes. Simon | -----Original Message----- | From: haskell-cafe-bounces(a)haskell.org [mailto:haskell-cafe-bounces@haskell.org] On Behalf Of Joel | Reymont | Sent: 06 December 2005 11:46 | To: Simon Peyton-Jones | Cc: Haskell Cafe | Subject: [Haskell-cafe] Re: TChan implementation: Why TVarList | | Well, I meant more like TVar [a] but I see that you are pulling from | the front and appending to the rear. | | I need to implement a mailbox where messages can be pulled out based | on a predicate or in order of arrival. I'm thinking of using a Map | keyed on ClockTime. | | Do you have any suggestions? | | Thanks, Joel | | On Dec 6, 2005, at 11:31 AM, Simon Peyton-Jones wrote: | | > The mutable cell is in the tail. A [TVar a] would be quite different. | > You can read about a very similar impl (based on MVars) in the | > original | > "Concurrent Haskell" paper (on my papers page) | | -- | http://wagerlabs.com/ | | | | | | _______________________________________________ | Haskell-Cafe mailing list | Haskell-Cafe(a)haskell.org | http://www.haskell.org/mailman/listinfo/haskell-cafe

2 1

TChan implementation: Why TVarList
by Joel Reymont 06 Dec '05

06 Dec '05

Simon, Why did you guys implement TChan on top of your own TVarList instead of a regular list? -- | 'TChan' is an abstract type representing an unbounded FIFO channel. data TChan a = TChan (TVar (TVarList a)) (TVar (TVarList a)) type TVarList a = TVar (TList a) data TList a = TNil | TCons a (TVarList a) Thanks, Joel -- http://wagerlabs.com/

1 0

Signal 2 and -threaded
by Joel Reymont 06 Dec '05

06 Dec '05

I'm seeing some funky behavior in my program compiled with -threaded. It runs for a little while and then catches signal 2 and quits. This does not happen when -threaded is not used. I'm compiling my library with -threaded so that is always constant. The above applies to my main module and another module that is part of it. Any ideas? I decided to compile with -threaded because otherwise killThread seems to block when a thread is in a foreign call. Thanks, Joel -- http://wagerlabs.com/

1 0

RE: [Haskell-cafe] GADT question
by Simon Peyton-Jones 06 Dec '05

06 Dec '05

| I would like the type system to distinguish the three types and enforce | their properties. so far, I have come up with the following: | | | data Sigma | data Tau | data Rho | | data Type a where | ForAll :: [Id] -> Type Rho -> Type Sigma | Fun :: Type a -> Type a -> Type a | TyCon :: TyCon -> Type Tau | TyVar :: TyVar -> Type Tau It looks as if you want a kind of subtyping relationship, so that Tau < Rho < Sigma. The standard way to achieve that is using polymorphism (see, for example, papers about FFI to object-oriented languages) data Sigma a = S a data Rho a = R a data Tau = Tau data Type a where | ForAll :: [Id] -> Type (Sigma (Rho a)) -> Type (Sigma b) | Fun :: Type a -> Type a -> Type a | TyCon :: TyCon -> Type (Sigma (Rho Tau)) | TyVar :: TyVar -> Type (Sigma (Rho Tau)) I'm not sure if this would work. Simon

1 0

STM and `orElse` on a few thousand TMVars
by Joel Reymont 06 Dec '05

06 Dec '05

Is it efficient to wait on a few thousand TMVars? How would I write something like that, assuming that my TMVars were in a list. Thanks, Joel -- http://wagerlabs.com/

1 0

RE: [Haskell-cafe] Re: :t main
by Scherrer, Chad 06 Dec '05

06 Dec '05

> From: Cale Gibbard [mailto:cgibbard@gmail.com] > See: > http://haskell.org/pipermail/haskell-cafe/2003-January/003794.html > > The OI comonad as previously envisioned breaks referential > transparency. I/O just doesn't seem to be something which one > can easily do comonadically, since once coeval/extract is > applied, you're back to plain values, and there's no > imposition of sequencing. > > - Cale Hmm, I hadn't seen that. The asymmetry is pretty frustrating in that case. After poking around a bit more, I (re)discovered some discussion of this on Hawiki: http://www.haskell.org/hawiki/CoMonad After reading Dave Menendez's comments, I'm wondering... If we consider IO a = Realworld -> (Realworld, a) then wouldn't we dually have something like OI a = (Realworld, Realworld -> a)? Could this be what screws things up? Right now it seems like OI is acting like it has the same type as IO, with a different name. Chad Scherrer Computational Mathematics Group Pacific Northwest National Laboratory "Time flies like an arrow; fruit flies like a banana." -- Groucho Marx

2 1

Re: [Haskell-cafe] Learning Haskell
by Donn Cave 06 Dec '05

06 Dec '05

Quoth Jimmie Houchin <jhouchin(a)cableone.net>: ... | Haskell looks like a very interesting language. I am only so-so with | Python and I thought that maybe if instead of spending sufficient time | to get proficient with Python, I could invest a similar time (more or | less) and get reasonably (pragmatically speaking) proficient with | Haskell. I know I may never understand the theory and maths behind the | design, but I believe I can appreciate the design and be a decent user | of the tools it provides. I'm coming from a similar background, Python early adopter (well, they do have something in common! Just something sort of trivial.) and no great math background. I find Haskell MUCH more mentally rigorous. I don't think that's just because I have previously been living in "imperative" land, it's because Haskell gets far more abstract. I have lately been thinking I might see if I can achieve a clear understanding of StateT. Not in any theoretical or mathematical sense, just want to understand how it goes together where honestly right now it seems kind of like magic. Not complaining, there could be worse things than magic like that, but I'm just saying our mathematical ignorance doesn't excuse us from some mental heavy lifting that doesn't compare to anything in Python. Then there are performance issues with Haskell that don't come up in other languages. I have tried to do things like analyze big text files with Haskell and given up, because I could not figure out how to get within an order of magnitude of reasonable completion time. I still think it's possible, though maybe not with standard String type, but anyway it's something to look out for. | I am also hoping to use it for web development. Wonder how `links' is coming along. (Hope they figured out a better name, anyway.) Donn

2 1

GADT question
by oleg＠pobox.com 06 Dec '05

06 Dec '05

John Meacham wrote: > data Type = ForAll [TyVar] Rho -- Forall type > | Fun Type Type -- Function type > | TyCon TyCon -- Type constants > | TyVar TyVar -- Always bound by a ForAll > with the three synonyms and their constraints being > type Sigma = Type > type Rho = Type -- No top-level ForAll > type Tau = Type -- No ForAlls anywhere > I would like the type system to distinguish the three types and enforce > their properties. Perhaps the enclosed code might be of help. Tests: test1 = TyCon "Int" *Forall> :t test1 test1 :: Type FANil test2 = Forall ["a"] (Fun (TyVar "a") (TyVar "a")) *Forall> :t test2 test2 :: Type FATop -- Inferred type! test3 = Fun test2 test1 *Forall> :t test3 test3 :: Type FASomewhere test3' = Fun test1 test1 *Forall> :t test3' test3' :: Type FANil test4 = Forall ["b"] test3 *Forall> :t test4 test4 :: Type FATop test4' = Forall ["b"] test2 -- Error! No instance for (RhoC FATop) arising from use of `Forall' at /tmp/m.hs:43:9-14 That is, in test4, the second argument of Forall is not of the type rho! Typeclasses would work too. I'm lacking a good example to show off this machinery... {-# OPTIONS -fglasgow-exts #-} module Forall where data FANil data FATop data FASomewhere type Id = String type Tau = Type FANil data Type a where Forall :: (RhoC a) => [Id] -> Type a -> Type FATop Fun :: (FAMerge a b c) => Type a -> Type b -> Type c TyCon :: Id -> Type FANil TyVar :: Id -> Type FANil type Sigma a = Type a --type (RhoC a) => Rho a = Type a -- I wish that worked... class RhoC a instance RhoC FANil instance RhoC FASomewhere class FAMerge a b c | a b -> c instance FAMerge FANil FANil FANil instance FAMerge FANil FASomewhere FASomewhere instance FAMerge FANil FATop FASomewhere instance FAMerge FATop a FASomewhere instance FAMerge FASomewhere a FASomewhere test1 = TyCon "Int" test2 = Forall ["a"] (Fun (TyVar "a") (TyVar "a")) test3 = Fun test2 test1 test3' = Fun test1 test1 -- I guess we need to hide the data constructors of Type and use -- functions like rho, fun, tycon, etc. to construct the values of -- Type rho :: (RhoC a) => Type a -> Type a rho = id test4 = Forall ["b"] test3 -- The following is error! --test4' = Forall ["b"] test2

1 0

GADT question
by John Meacham 06 Dec '05

06 Dec '05

I have been trying to learn GADTs but feel I don't quite grok them yet. I am playing with an implementation of SPJ's algorithm from http://research.microsoft.com/Users/simonpj/papers/higher-rank/index.htm in it, he has a data type that represents a type, but with several special cases that can't easily be expressed in haskell proper. I was thinking GADTs might help but havn't been able to come up with something good and was hoping someone could point me in the right direction or let me know what I am trying to do is impossible/improbable. right now, it looks like so data Type = ForAll [TyVar] Rho -- Forall type | Fun Type Type -- Function type | TyCon TyCon -- Type constants | TyVar TyVar -- Always bound by a ForAll with the three synonyms and their constraints being type Sigma = Type type Rho = Type -- No top-level ForAll type Tau = Type -- No ForAlls anywhere I would like the type system to distinguish the three types and enforce their properties. so far, I have come up with the following: data Sigma data Tau data Rho data Type a where ForAll :: [Id] -> Type Rho -> Type Sigma Fun :: Type a -> Type a -> Type a TyCon :: TyCon -> Type Tau TyVar :: TyVar -> Type Tau but this is obviously not right, it can't create Rho types at all, Fun should change a Sigma into a Rho but leave Tau the same, Fun should be able to be applied to different types infering the correct one. and a function that accepts a Rho should be able to accept a Tau... (though explicit upconversions would be okay, meaning I need to create TyCons for example with type 'Type Rho' in the upconverter) anyone have any ideas on how to make this work? is this even possible to express with GADTs? John -- John Meacham - ⑆repetae.net⑆john⑈

1 0

RE: [Haskell-cafe] Re: Invalid object in isRetainer
by Simon Marlow 05 Dec '05

05 Dec '05

On 05 December 2005 15:29, Joel Reymont wrote: > So how do you profile for memory leaks then? Ordinary heap profiling (-hc) should work fine with STM. The more exotic kinds of heap profiling haven't been updated yet, so unfortunately they're broken with STM right now. Cheers, Simon

1 0

RE: Invalid object in isRetainer
by Simon Marlow 05 Dec '05

05 Dec '05

On 04 December 2005 17:02, Joel Reymont wrote: > ./randomplay +RTS -hr > ... <- some time passes and then > > randomplay: internal error: Invalid object in isRetainer(): 67 > Please report this as a bug to glasgow-haskell-bugs(a)haskell.org, > or http://www.sourceforge.net/projects/ghc/ > > The combination of this and the failure with -hc -hbdrag,void is > preventing me from optimizing the program and it has a huge space > leak somewhere. Retainer profiling (-hr) doesn't currently work with STM. It is a bug, but I briefly looked into it and it's not a trivial fix so I left it. I'll file a ticket for it. Cheers, Simon

2 1

RE: Wierd activation record found on stack
by Simon Marlow 05 Dec '05

05 Dec '05

On 04 December 2005 16:50, Joel Reymont wrote: > To add a little bit of detail (thanks to dons for the tips), the > signal is being caught in one of the threads launched by the main > thread. Simply calling exitWith (ExitFailure 1) from this thread > causes the error below. I changed my code to send messages to running > threads asking them to quit and this made the error go away. > > On Dec 4, 2005, at 12:27 PM, Joel Reymont wrote: > >> randomplay: internal error: scavenge_stack: weird activation record >> found on stack: 0 Please report this as a bug to >> glasgow-haskell-bugs(a)haskell.org, or >> http://www.sourceforge.net/projects/ghc/ >> >> I installed my own signal handlers with the intent of having my >> profile reports written out. The program runs forever and otherwise >> generates no profile reports when I interrupt it. This was fixed recently, I believe: revision 1.276 date: 2005-11-08 10:44:22 +0000; author: simonmar; state: Exp; lines: +1 -2 raiseAsync: fix bug that can cause a scavenge_stack panic for a thread that has just been killed. The fix will be in 6.4.2, or you can grab a STABLE snapshot from the haskell.org. Cheers, Simon

1 0

RE: [Haskell-cafe] STM, orElse and timed read from a channel
by Simon Peyton-Jones 05 Dec '05

05 Dec '05

| > I think you are really asking for some way to create top-level | > transactional variables? | | Yes. Perhaps it would be easier to do it cleanly than arbitrary | top-level IO bindings? It turns out to be easy to provide newTVarIO :: a -> IO (TVar a) which you can call from inside 'unsafePerformIO'. That means you can allocate top-level TVars without fuss. The whole question of top-level IO remains open, but this will address the immediate question very nicely. I've also added a SourceForge bug to remind us to fail more tidily if you do atomically inside atomically. Simon

2 1

RE: [Haskell-cafe] STM, orElse and timed read from a channel
by Simon Peyton-Jones 05 Dec '05

05 Dec '05

Can you provide a small test case that demonstrates the crash? I think it's pretty unlikely that we'll support nested atomically (via unsafePerformIO) anytime soon. UnsafePerformIO is, well, unsafe. I think you are really asking for some way to create top-level transactional variables? Simon | -----Original Message----- | From: haskell-cafe-bounces(a)haskell.org [mailto:haskell-cafe-bounces@haskell.org] On Behalf Of | Tomasz Zielonka | Sent: 04 December 2005 18:32 | To: Simon Marlow | Cc: Haskell Cafe | Subject: Re: [Haskell-cafe] STM, orElse and timed read from a channel | | On Sat, Dec 03, 2005 at 10:35:54PM +0100, Tomasz Zielonka wrote: | > so I could export a waitUntil function with type | > | > waitUntil :: Time -> STM () | > | > but I tripped on something that was reported before, namely that STM | > transactions can't be nested (as a result of unsafePerformIO or | > unsafeInterleaveIO). Is there a plan to support such scenario? | | OK, it can be worked around by running the atomically block in a new | transaction. But still it would be nice if the program didn't segfault, | but cause an exception for example. | | Updated modules attached. | | Best regards | Tomasz | | -- | I am searching for a programmer who is good at least in some of | [Haskell, ML, C++, Linux, FreeBSD, math] for work in Warsaw, Poland

2 1

STM, Concurrent Haskell and network clients (long, code)
by Joel Reymont 05 Dec '05

05 Dec '05

Folks, After two months with Haskell I think I'm starting to get the proper spirit. Inspired by Simon PJ's recent comments about lots of threads not hurting Haskell I introduced STM into my network client. I would like to run my architecture by you to see if there are any improvements to be made. I would like the code to be as bullet-proof as possible. I'm also jumping through some hoops to make sure none of the 3 threads (main, socket reader and socket writer) block indefinitely. I'm getting a "thread blocked indefinitely" exception now and I think this is because the main thread is not getting notified when the socket threads die. My network client has two threads. #1 reads network packets, converts them to events and posts them to a TChan using a function passed to it by the main thread. #2 blocks on a second TChan and waits for events to be sent out through the socket. The main thread creates a TChan to read events from, passes a closure that writes to the TChan to the network client and uses a network client function to send events to the network client. All in all it works out very nicely as there's no networking, FFI, etc. in the network client logic. This is an improvement from what I had before. The function that starts the poker client has this signature: startPokerClient :: (Event -> IO ()) -- fun to send events to the main thread -> HostInfo -- (host, port) -> AuthInfo -- (user, pass, etc.) -> IO PokerClient where data PokerClient = PC !MaybeThread !MaybeThread !(TChan Command) The main network client thread calls this fun to deliver events to the network client: writePokerClient :: PokerClient -> Command -> IO () writePokerClient (PC _ _ chan) cmd = atomically $ writeTChan chan cmd Main thread can stop the network client using stopPokerClient :: PokerClient -> IO () stopPokerClient (PC tmv2 tmv1 _) = do putStrLn $ "stopPokerClient invoked" maybeKillThread tmv2 maybeKillThread tmv1 The abbreviated startPokerClient looks like this.. {-# NOINLINE lock #-} lock :: MVar () lock = unsafePerformIO $ newMVar () startPokerClient post (host, port) auth = do h <- withMVar lock -- getHostByName is not thread-safe $ \_ -> connectTo host (PortNumber $ fromIntegral port) `catchError` (\e -> fail $ "Failed to connect to server: " ++ show e) -- kick off the SSL handshake ssl <- startSSL doSSLHandshake h ssl ... -- we need one thread to always kill the other (tmv1 :: MaybeThread) <- atomically $ newEmptyTMVar (tmv2 :: MaybeThread) <- atomically $ newEmptyTMVar (tmvssl :: MaybeSSL) <- atomically $ newTMVar (Just ssl) -- thread to read from the socket and post commands tid1 <- forkIO $ writeLoop post h ssl `finally` finalize tmv1 tmv2 h tmvssl post -- thread to write to grab commands generated -- by the main thread and write them to the socket chan <- atomically $ newTChan tid2 <- forkIO $ readLoop h chan `finally` finalize tmv2 tmv1 h tmvssl post -- we have both threads now atomically $! do putTMVar tmv1 (Just tid1) putTMVar tmv2 (Just tid2) return $! PC tmv2 tmv1 chan Interestingly enough, I had to use $! otherwise startPokerClient was sitting there for a while, due to due to lazy evaluation I assume. I'm using the following finalizer for my socket reader/writer threads to make sure that if one thread dies the other one is taken with it and that the main thread is notified. I don't have a Maybe around the socket itself since I don't think closing a socket twice has any bad side effects. I put it in a try, though, for good measure. finalize :: MaybeThread -> MaybeThread -> Handle -> MaybeSSL -> (Event -> IO ()) -> IO () finalize me buddy h ssl post = do -- make sure we are not killed again -- killing a dead thread is apparently -- not free of side effects in GHC 6.4.1 atomically $ putTMVar me Nothing -- kill the other thread maybeKillThread buddy maybeFreeSSL ssl try $ hClose h post $ StopScript -- notify main thread return () The rest of the relevant code is below. Please let me know if it can be simplified or made more robust. maybeFreeSSL :: MaybeSSL -> IO () maybeFreeSSL tmv = do putStrLn $ "maybeFreeSSL invoked" mssl <- atomically $ swapTMVar tmv Nothing case mssl of Nothing -> return () Just (ssl, _, _) -> do sslFree ssl sslFree ssl maybeKillThread :: MaybeThread -> IO () maybeKillThread tmv = do putStrLn $ "maybeKillThread invoked " mtid <- atomically $ swapTMVar tmv Nothing case mtid of Nothing -> return () Just tid -> killThread tid --- Socket reader writeLoop :: (Event -> IO ()) -> Handle -> (SSL, BIO, BIO) -> IO () writeLoop post h ssl = do cmd <- read h ssl post $ Cmd cmd writeLoop post h ssl --- Socket writer readLoop :: Handle -> TChan Command -> IO () readLoop h chan = do cmd <- atomically $ readTChan chan write h cmd readLoop h chan --- SSL, etc. not used now but will be used to decrypt later read :: Handle -> (SSL, BIO, BIO) -> IO Command read h _ = do bytes <- hGet h 4 let (size', _) = unpickle endian32 bytes size = fromIntegral $ size' - 4 if size > 0 then do packet <- hGet h size when (size /= length packet) $ fail "Packet size mismatch" unstuff packet else do fail $ "read: size is 0: bytes: " ++ show bytes unstuff :: [Word8] -> IO Command unstuff packet = do let (kind, packet') = unpickle puCmdType packet (cmd', _) = unpickle (puCommand kind) packet' case cmd' of InvalidCommand -> do fail $ "unstuff: Cannot parse: " ++ show packet SrvCompressedCommands sz data' -> do data'' <- uncompress data' $ fromIntegral sz unstuff $ drop 4 data'' _ -> return $! cmd' write :: Handle -> Command -> IO () write h cmd = do let kind = cmdType cmd bytes' = pickle puCmdType kind bytes = bytes' ++ pickle (puCommand kind) cmd when (length bytes <= 0) $ fail $ "write: Could not pack: " ++ show cmd let size = 4 + length bytes size' = pickle endian32 $ fromIntegral size when (length size' <= 0) $ fail $ "write: Could not pack: " ++ show size hPut h (size' ++ bytes) hFlush h -- http://wagerlabs.com/

2 3

Problem with continuations and typing
by tpledger＠ihug.co.nz 04 Dec '05

04 Dec '05

Jerzy Karczmarczuk wrote: : | zeros fc sc = sc 0 zeros | | fails to compile as well. *I do not ask why, I know*. | | But I would like to continue this exercice along these lines, without too much | exotism (no monads, yet...), for my students. Do you have any simple work-around? | Introduce some algebraic constructors? Perhaps higher-rank polymorphism could do | something (but then I would have to explain it to my folk...) : How about this for a non-exotic algebraic type? > newtype G a b = G{ unG :: b -> (a -> G a b -> b) -> b } > glist g = unG g [] (\b g' -> b : glist g') > zeros = G (\no yes -> yes 0 zeros) > disj g1 g2 = G (\no yes -> unG g1 (unG g2 no yes) > (\b g1' -> yes b (disj g1' g2))) I haven't had much practice with continuations, so don't know whether I've just lost some generality there. But it does support *some* avoidance of higher-rank polymorphism, through the use of good old partial application. For example, the type of the state variable s doesn't leak into the result type of unfold: > unfold f s = G (\no yes -> case f s of > Nothing -> no > Just (s', b) -> yes b (unfold f s')) HTH, Tom

1 0