- Haskell-Cafe - Haskell.org

hackage compile failure with QuickCheck 2.5
by Levent Erkok 17 Jul '12

17 Jul '12

[This message is more appropriate for a hackage mailing list I presume, but that doesn't seem to exist. Let me know if there's a better place to send it.] I'm having a hackage compile failure for a newly uplodaded package that has a QuickCheck 2.5 dependence. The error message is: [13 of 13] Compiling Test.QuickCheck.All ( Test/QuickCheck/All.hs, dist/build/Test/QuickCheck/All.o ) Test/QuickCheck/All.hs:15:1: Bad interface file: /usr/local/tmp/archive/install/lib/template-haskell-2.6.0.0/ghc-7.4.1/Language/Haskell/TH.hi Something is amiss; requested module template-haskell-2.6.0.0:Language.Haskell.TH differs from name found in the interface file template-haskell:Language.Haskell.TH The full log file is at (search for "Something is a miss" in it): http://hackage.haskell.org/packages/archive/sbv/2.2/logs/failure/ghc-7.4 Needless to say, I don't see this problem when I compile this package at home with the same compiler (ghc 7.4.1) as hackage is using; also Hackage has a successfully compiled QuickCheck 2.5 package. Could it be something related to the particular cabal/ghc installation on the hackage server? In particular, I don't understand why it picks template-haskell 2.6.0.0 when there's a newer version (2.7.0.0). As far as I can see, QuickCheck doesn't put an upper limit on its template haskell version dependency. I'd appreciate any pointers with this. (Googling and questions on the #haskell irc channel didn't help much, unfortunately.) -Levent.

1 0

Haskell's type inference considered harmful [Re: [Haskell] A riddle...]
by Andreas Abel 17 Jul '12

17 Jul '12

[Is Haskell-Cafe the right place to discuss DESIGN issues of the Haskell language? This message is not about a concrete problem...] Congratulations to all that solved the riddle so quickly. (First answer in only 8 Minutes!) Now to the point of the exercise: Shocking realizations. 1. Haskell's type inference is NON-COMPOSITIONAL! In the riddle below, I am defining two things f ("rgbliste") and g ("farbliste"). Even though they are not strongly connected, but g comes after f in the definition order, the code of g influences the type of f. THAT'S WRONG! :-( It gets worse: 2. DEAD CODE changes a program's result! Yes indeed. Dead code ("farbliste") in the source cannot be removed safely. And: 3. Haskell "casts" 181935629 to Word16 value 7693 WITHOUT WARNING! Needless to say, these are serious problems for leading-edge strongly typed language. The experts are probably aware of this. But I awoke harshly from a sweet dream. Issue 3 is quite harmful and reminds me more of C than of a strongly typed language. Fixing this is not trivial. The problem is that all numeric constants have type Num a => a, and of course, type variable a can be instantiated to Int Word16 or whatever. Numeric constants need to inhabit subclasses of Num, e.g., Word16, Int and Integer /classes/ to make this behavior go away. Update: There is nothing wrong with Word16 computing modulo 2^16, but a literal like 181935629 should not get type Word16 without warning. Java asks for explicit casts in byte i = (byte) 1234876; but mindlessly also in byte i = (byte) 123; which is not a great solution either. Issues 1/2 go away with {-# LANGUAGE NoMonomorphismRestriction #-} (thx to Christian who pointed me to this). I wonder why this is not on by default?! There seems to be an unresolved dispute about this... http://hackage.haskell.org/trac/haskell-prime/wiki/NoMonomorphismRestriction But even with MonomorphismRestriction, issue 1/2 are avoidable with a more refined approach to type meta variables in the Haskell type inference: After type-checking a strongly connected component, type meta variables created for this part of the Haskell program should not be instantiated in the (topologically) later defined parts of the program. YOU DON'T WANT THE USE OF A FUNCTION INFLUENCE ITS DEFINITION! There are two alternatives what to do with remaining meta-variables of a SCC: 1. With monomorphism restriction: instantiate to the "best type". In our example below, Haskell chooses Integer. 2. Without monomorphism restriction: generalize. Choice 1 will give an error in the program below, but i rather have an error than a silent change of behavior of my program! Cheers, Andreas P.S.: In Agda, in the situation of unsolved metas in an "SCC" (a mutual block), these metas are just frozen, triggering typing errors later (since Agda cannot generalize). The user has to go back an insert more type annotations. But this is safe, whereas silent late instantiation breaks compositionality. On 16.07.2012 17:25, Andreas Abel wrote: > Today a student came to me with a piece of code that worked it executed > by itself, but produced different result in the context of his larger > problem. We cut down the example to the following: > >> import Graphics.UI.Gtk >> >> -- should produce [(26471,0,65535),... >> rgbliste = >> (map (\ i -> >> let rb = 5 * (mod (mod 181935629 (4534+i)) 100)-250+128 in >> let gb = 5 * (mod (mod 128872693 (5148+i)) 100)-250+128 in >> let bb = 5 * (mod (mod 140302469 (7578+i)) 100)-250+128 in >> let r = min 255 $ max 0 rb in >> let g = min 255 $ max 0 gb in >> let b = min 255 $ max 0 bb in >> (r*257,g*257,b*257)) [0..]) >> >> --farbliste = map (\ (r,g,b) -> Color r g b) rgbliste >> >> main :: IO () >> main = do >> print $ head rgbliste > > If you run it, it prints (26471,0,65535). > If you uncomment the def. of farbliste, it prints (44461,65535,65535). > > I was surprised. What is going on? > > Cheers, > Andreas > -- Andreas Abel <>< Du bist der geliebte Mensch. Theoretical Computer Science, University of Munich Oettingenstr. 67, D-80538 Munich, GERMANY andreas.abel(a)ifi.lmu.de http://www2.tcs.ifi.lmu.de/~abel/

3 2

Call to arms: lambda-case is stuck and needs your help
by Mikhail Vorozhtsov 17 Jul '12

17 Jul '12

Hi. After 21 months of occasional arguing the lambda-case proposal(s) is in danger of being buried under its own trac ticket comments. We need fresh blood to finally reach an agreement on the syntax. Read the wiki page[1], take a look at the ticket[2], vote and comment on the proposals! P.S. I'm CC-ing Cafe to attract more people, but please keep the discussion to the GHC Users list. [1] http://hackage.haskell.org/trac/ghc/wiki/LambdasVsPatternMatching [2] http://hackage.haskell.org/trac/ghc/ticket/4359

4 4

Re: [Haskell-cafe] What is the difference between runhaskell and compile?
by Magicloud Magiclouds 17 Jul '12

17 Jul '12

OK. Finally this problem totally prevents my project from working. By some guy's suggestion, I did a packet capturing, which really showed a difference. The rwdc file is for a ldap that at our IDC, which did not work with compiled binary. The rodc file is for a ldap that at local, which worked all the time. Each file contains two parts of records, at 3.* sec and 6/7.* sec. The 3.* part is for runhaskell way. The 6/7.* part is for compiled binary way. The big difference here is: the failed one sent unbindRequest before got bindResponse success. Just a clue, I have no idea what is going on.... On Fri, Jun 1, 2012 at 5:17 PM, Chris Dornan <chris(a)chrisdornan.com> wrote: > No problem -- I have attached the executable too. > > Chris > > -----Original Message----- > From: Magicloud Magiclouds [mailto:magicloud.magiclouds@gmail.com] > Sent: 01 June 2012 02:01 > To: Chris Dornan > Subject: Re: [Haskell-cafe] What is the difference between runhaskell and compile? > > I see. > Although I am using a UTF-8 environment, but all string I was using were ASCIIs. > If it is not too much trouble, may I have your LDAP compiled library files? I think since the Main.o is identical, maybe the libraries are different. > > On Thu, May 31, 2012 at 2:09 PM, Chris Dornan <chris(a)chrisdornan.com> wrote: >> I wouldn't pay too much attention to the fact that the hash tags are >> different -- they are very unlikely to be the same across two >> installations. (For example, my GHC installations are relocated under >> /usr/hs, immediately changing the contents of all of the .conf files >> and therefore the hashes.) >> >> The object files are more interesting, and they are identical! >> >> At the moment I can only explain what we are seeing by assuming that some change has affected the GHC 7.4.1 runtime system -- but only for compiled programs. >> >> Are you using non-ASCII characters in your LDAP passwords? I am thinking that maybe the compiled 7.4.1 could be using a different LOCALE from everything else, leading in some way to the passwords being encoded differently en route to the LDAP server. I don't know whether this could actually happen, but it is the kind of mechanism that seem most likely to be causing the problem. >> >> Did you try to running the variants that read address, account and passwords from standard input? Did it show the same perverse behaviour? >> >> Chris >> >> >> >> >> >> -----Original Message----- >> From: Magicloud Magiclouds [mailto:magicloud.magiclouds@gmail.com] >> Sent: 31 May 2012 02:47 >> To: Chris Dornan >> Subject: Re: [Haskell-cafe] What is the difference between runhaskell and compile? >> >> Hi, >> I have compared the files. The .o-es are exactly the same. The .hi-s are different. >> The interface hash, flag hash, and "import -/ LDAP-0.6.6:LDAP.Constants xxxxx(hash)" are different. >> I do not know why only this module (LDAP.Constants) has a different hash. I am not quite familiar with ghc compiling process. May I know your conclusion? >> >> On Wed, May 30, 2012 at 3:50 PM, Chris Dornan <chris(a)chrisdornan.com> wrote: >>> Did you carry out a standard (optimised) build to get ghc-7.4.1? >>> >>> Are you going to try the justhub rpms or compare the object/header files I sent. >>> >>> Chris >>> >>> -----Original Message----- >>> From: Magicloud Magiclouds [mailto:magicloud.magiclouds@gmail.com] >>> Sent: 30 May 2012 04:07 >>> To: Chris Dornan >>> Cc: Haskell-Cafe >>> Subject: Re: [Haskell-cafe] What is the difference between runhaskell and compile? >>> >>> A little information. >>> I did not notice the gcc/binutils versions. But in CentOS, the ghc >>> 7.2.2/7.4.1 were all compiled myself with all default configurations. >>> >>> On Tue, May 29, 2012 at 10:54 PM, Chris Dornan <chris(a)chrisdornan.com> wrote: >>>> On 29 May 2012 02:21, Magicloud Magiclouds >>>> <magicloud.magiclouds(a)gmail.com> wrote: >>>>> Interesting. I have this code tested in Debian unstable/stable, >>>>> CentOS 6.1, all 64 bit, with two different version of libldap2. >>>>> At first, Debian-s were installed with 7.4.1, CentOS with 7.2.2. >>>>> Only in CentOS the code connected after compiled. >>>>> Then I removed 7.4.1 from Debian stable and installed 7.2.2. The code worked. >>>>> At last, I installed 7.4.1 in CentOS. The code did not work. >>>>> >>>>> Could you send the .hi/.o to me, so maybe I could find out the >>>>> different? Also the exact original source. >>>>> Thank you. >>>> >>>> Interesting indeed! I am guessing that you are using the GHC-7.4.1 >>>> bindist from haskell.org. >>>> >>>> I will try and find some time to marshal the source code and >>>> intermediate files (am on the road -- will need to collect it from >>>> base, make it generic etc.). >>>> >>>> You might also like to try the http://justhub.org ghc-7.4.1-hub on >>>> your CentOS-6.1 node. It is a separate build from the haskell.org >>>> bindist and comes with it's own in-board gcc (4.6.1) and binutils >>>> (2.21) used for the build. It should work for you. >>>> >>>> (You could also try ghc-7.4.2-RC1-hub.) >>>> >>>> Chris >>> >>> >>> >>> -- >>> 竹密岂妨流水过 >>> 山高哪阻野云飞 >>> >>> And for G+, please use magiclouds#gmail.com. >>> >> >> >> >> -- >> 竹密岂妨流水过 >> 山高哪阻野云飞 >> >> And for G+, please use magiclouds#gmail.com. >> > > > > -- > 竹密岂妨流水过 > 山高哪阻野云飞 > > And for G+, please use magiclouds#gmail.com. -- 竹密岂妨流水过山高哪阻野云飞 And for G+, please use magiclouds#gmail.com.

2 6

Adding to the haskell wiki
by Nathan Hüsken 16 Jul '12

16 Jul '12

Hey, I have been a passive member of the haskell community for quite some time now. I have benefit from the community a lot through the mailing list, the haskell webpage and other channels. Thanks for all that! Now I am trying to give a little bit back, and I started a blog about writing javascript games in haskell using UHC and haste. I am not an expert in haskell nor javascript, but I am "collecting" information from various sources and put it together in the blog so that others with a similar goal my have a smoother start. Writing a blog is one thing, but of course I also want people who are interested in the topic to find it. Never being an active member of the community, I wonder if I should do. I guess I should add it to the haskellwiki. But I am unsure where. It certainly falls in the "games" category, but the page only lists game projects. The blog can be found here: http://jshaskell.blogspot.de/ Regards, Nathan

3 2

Re: [Haskell-cafe] [Haskell] A riddle...
by Vo Minh Thu 16 Jul '12

16 Jul '12

Oh the mailing list was haskell@ instead of café. Sorry. (Now cc'ing café.) Andreas, please post this kind of mail to the haskell-cafe instead of the haskell mailing list. 2012/7/16 Vo Minh Thu <noteed(a)gmail.com>: > It seems like the infered type (and thus bounds) is different when you > force the result to be a Color or not. Just give explicit type > signatures and conversion functions. > > Cheers, > Thu > > 2012/7/16 Andreas Abel <andreas.abel(a)ifi.lmu.de>: >> Today a student came to me with a piece of code that worked it executed by >> itself, but produced different result in the context of his larger problem. >> We cut down the example to the following: >> >>> import Graphics.UI.Gtk >>> >>> -- should produce [(26471,0,65535),... >>> rgbliste = >>> (map (\ i -> >>> let rb = 5 * (mod (mod 181935629 (4534+i)) 100)-250+128 in >>> let gb = 5 * (mod (mod 128872693 (5148+i)) 100)-250+128 in >>> let bb = 5 * (mod (mod 140302469 (7578+i)) 100)-250+128 in >>> let r = min 255 $ max 0 rb in >>> let g = min 255 $ max 0 gb in >>> let b = min 255 $ max 0 bb in >>> (r*257,g*257,b*257)) [0..]) >>> >>> --farbliste = map (\ (r,g,b) -> Color r g b) rgbliste >>> >>> main :: IO () >>> main = do >>> print $ head rgbliste >> >> >> If you run it, it prints (26471,0,65535). >> If you uncomment the def. of farbliste, it prints (44461,65535,65535). >> >> I was surprised. What is going on? >> >> Cheers, >> Andreas >> >> -- >> Andreas Abel <>< Du bist der geliebte Mensch. >> >> Theoretical Computer Science, University of Munich >> Oettingenstr. 67, D-80538 Munich, GERMANY >> >> andreas.abel(a)ifi.lmu.de >> http://www2.tcs.ifi.lmu.de/~abel/ >> >> >> _______________________________________________ >> Haskell mailing list >> Haskell(a)haskell.org >> http://www.haskell.org/mailman/listinfo/haskell

2 1

lazy boxed array and builder?
by Evan Laforge 16 Jul '12

16 Jul '12

he recent discussion of whether storablevector should be deprecated in favor of vector reminded me: vector supports storable arrays, but it doesn't support lazy arrays. While storablevector has lazy arrays and a builder, it doesn't support boxed types (it would be become misnamed if it did!). So it seems the niche of boxed lazy arrays is unfilled. And if vector grew a lazy variant it could subsume storablevector and we could consolidate array libraries a little more. It seems a lot of Writer monad type stuff (logging etc.) could be made more efficient with a boxed lazy array builder, at the cost of rougher grained laziness. Does such a thing already exist? I'd check hackage, but it's down... http://www.downforeveryoneorjustme.com/http://hackage.haskell.org/

5 4

Getting a segmentation fault when starting/stopping the RTS, from C, several times.
by Captain Freako 16 Jul '12

16 Jul '12

Hi experts, Should I expect the following C code to run to completion, or am I trying to do something that was never intended? Thanks, -db C code: 1 #include <stdio.h> 2 #include <stdlib.h> 3 #include "HsFFI.h" 4 5 int main() 6 { 7 int argc = 1, i; 8 char* argv[] = {"ghcDll", NULL}; // argv must end with NULL 9 char** args = argv; 10 11 for(i=0; i<10; i++) { 12 fprintf(stderr, "Starting up the RTS...\n"); 13 hs_init(&argc, &args); 14 printf("Shutting down the RTS...\n"); 15 hs_exit(); 16 } 17 } This is what I get, when I compile and run it: dbanas@dbanas-eeepc:~/prj/haskell/amitool$ make test_rts rm -f test_rts ghc -o test_rts -lHSrts -lm -lffi -lrt test_rts.o dbanas@dbanas-eeepc:~/prj/haskell/amitool$ ./test_rts Starting up the RTS... Shutting down the RTS... Starting up the RTS... Shutting down the RTS... Segmentation fault

4 4

Plain lambda inside banana brackets in the arrow notation
by Tsuyoshi Ito 15 Jul '12

15 Jul '12

Hello, In a program, I have an arrow MyArr and a combinator called repeat of the following type: repeat :: Int -> (Int -> MyArr e a) -> MyArr e a My problem is that the code becomes messy when I use this combinator inside the arrow notation, and I am looking for a way to write the code in a more readable way. To explain the problem, first consider the following combinator repeat', which is less general than repeat: repeat' :: Int -> MyArr (e, Int) a -> MyArr e a repeat' n f = repeat n g where g i = arr (\e -> (e, i)) >>> f Combinator repeat' is nice to use in the arrow notation, thanks to banana brackets and the interpretation of lambda: test1 :: MyArr [Double] String test1 = proc xs -> do let y = func1 xs z <- job1 -< xs (|(repeat' 100) (\i -> job2 -< xs !! i + y + z)|) -- func1 :: [Double] -> Double -- job1 :: MyArr [Double] Double -- job2 :: MyArr Double String However, in my program, I often have to use repeat instead of repeat' like: test2 :: MyArr [Double] String test2 = proc xs -> do let y = func1 xs z <- job1 -< xs repeat 100 (\i -> proc (xs, y, z) -> job3 (i * 2) -< xs !! i + y + z) -< (xs, y, z) -- job3 :: Int -> MyArr Double String Note that variable i is used as an argument to function job3 outside MyArr, and this cannot be done with repeat'. The code for test2 looks messy to me because I have to write “(xs, y, z)”, that is, the list of variables used inside the subcomputation explicitly (and twice). It does not seem possible to use banana brackets here because the type of the subcomputation does not meet the requirements stated in http://www.haskell.org/ghc/docs/7.4.2/html/users_guide/arrow-notation.html#…. How can I use combinators like repeat, which takes a plain function as an argument, in the arrow notation in a more readable way? Or am I trying to do an impossible thing? Best regards, Tsuyoshi

3 9

I can see lists "working" for matrix algorithms if the access patterns are mostly from the start or end of rows and columns.
by KC 15 Jul '12

15 Jul '12

Are there any such matrix algorithms that operate mainly from the start or end or rows and columns? Note: If the access pattern starts from the end of a row then one stores the row in reverse order in a list. -- -- Regards, KC

1 0