- Haskell-Cafe - Haskell.org

how to organize a parallel parser/compiler
by Maurizio Vitale 27 May '15

27 May '15

hello haskell-cafe, suppose you have a compiler pipeline roughly similar to (from write you a Haskell <http://dev.stephendiehl.com/fun/007_path.html> ): modl :: FilePath -> L.Text -> CompilerM () modl fname = parseP fname >=> dataP >=> groupP >=> renameP >=> desugarP >=> inferP >=> evalP and suppose you have a language where you can compile multiple files and compilation must be made _as if_ files were concatenated together in the order given on the command line (there might be side effects that affect parsing of subsequent files); still in the average case there're no dependencies and files could be compiled in parallel (or some of the dependencies can be handled at the AST level as one can prove they don't affect parsing but only aspects of the AST that can be patched). What are good strategies for dealing with this and rerun some of the parseP functions (and in the simplest solution the complete pipeline after it). The strategy I have in mind now doesn't mix well with the above pipeline, so I'd like to see if there're alternative solutions. Basically, what I have in mind is: - each parse function gets a file, something to watch on for the result of the previous parse (let's say an MVar, or some variation of speculation <https://hackage.haskell.org/package/speculation> , not sure yet) and an input environment (same for everybody). It produces an AST + what he's sensitive to (e.g. what would have affected the parsing) and what he generates that the next guy must be sensitive to. - before producing the 'what it generates part', it must be sure to have completed a valid parse, so he'll wait on the input MVar to know that the previous files have been parsed properly and that whatever side effects it caised wouldn't affect parsing. This wait will be done on a thread that doesn't count towards the parallelism limit as it is presumably cheap and we don't really serialize the parsing. [a similar question related to the above pipeline would be how do I fit in any kind of global transformation pass] Is there any library or other ideas on how to combine multiple pipelines where we want to run them with maximum parallelism but the outcome of (even a partial step of) one pipeline can invalidate and require to rerun others? Thanks for any idea, Maurizio

1 0

ZeroMQ Thread Safety
by David Turner 27 May '15

27 May '15

Hi all, According to the ZeroMQ docs[1]: "Individual ØMQ sockets are not thread safe except in the case where full memory barriers are issued when migrating a socket from one thread to another." The Haskell ZeroMQ binding [2] is a fairly thin wrapper around the C API, and in particular doesn't seem to do much to enforce thread safety as described above. My question: do I need to use bound threads to call functions on sockets to ensure that all the foreign calls happen on the same OS thread, or is it the case that two consecutive foreign calls in a single Haskell thread have a "full memory barrier" between them if they end up being called from different OS threads? Thanks in advance, David [1] http://api.zeromq.org/4-0:zmq [2] http://hackage.haskell.org/package/zeromq4-haskell

1 0

Account for the Haskell Wiki
by Alois Cochard 27 May '15

27 May '15

Hi, I would like a user account to access the Haskell wiki. Desired username: 'aloiscochard' Cheers -- *Λ\ois* http://twitter.com/aloiscochard http://github.com/aloiscochard

1 0

ClassyPrelude vs. Haskell.Language.Interpreter
by Mike Meyer 26 May '15

26 May '15

No answer on -beginners, so I'm trying -cafe. I'm trying to run interpreted code via ClassyPrelude, and getting some results that make me suspect a bug in the Prelude's type system. Or maybe the interpreter. Anyway, here's a bit of code that works as expected: {-# LANGUAGE NoImplicitPrelude #-} import ClassyPrelude import Language.Haskell.Interpreter main :: IO () main = do fun <- runInterpreter $ makeFun "reverse" case fun of Left e -> print e Right f -> readFile "/etc/motd" >>= hPut stdout . f makeFun expr = do set [languageExtensions := [NoImplicitPrelude]] setImportsQ [("ClassyPrelude", Nothing)] interpret expr (as :: Text -> Text) I don't think I can simplify this any further. It works as expected, and also works as expected, and prints out the contents of /etc/motd reversed. However, if you change the type signature in the last line from Text -> Text to LText -> Ltext (to get lazy text), you get no output. But if you change the function in the first line after main from "reverse" to "id", it works. So far, it might be an issue with lazy IO. However, change the type signature in the last line to LText -> Text. In this case, there is no output for either value of the expression. I expect an error in this case, as neither id nor reverse should be able to have the type LText -> Text! So, is there something I missed in either ClassyPrelude or the Interpreter? Or is this a subtle interaction, in which case can someone suggest a workaround? Or have I found a bug in one of the two?

3 3

Re: [Haskell-cafe] Installing ghc-mod on Windows
by Anakreontas 26 May '15

26 May '15

I think cabal reported errors regarding the header files where in fact the problem was that it could not find a c compiler. I tried to install QuickCheck and the primitives package required by QuickCheck failed to install with errors related to header files. After installing gcc, primitives and QuickCheck installed without errors. On Tuesday, May 26, 2015 at 11:18:30 AM UTC+1, Anakreontas wrote: > > ... continued from previous message > .. even with command cabal install --extra-include-dirs=include cabal > still reports an error. > The header file is useful for older ghc versions that the one I have so > removing the dependency on the header resolved the problem. > > On Tue, May 26, 2015 at 11:16 AM, Anakreontas Mentis < > anakreonmejdi(a)gmail.com> wrote: > >> I installed the MinGHC distribution but it did not solve the problem. >> >> To install ghc-mod I had to: >> 1) cabal install ghc-mod >> to install what cabal can build without errors. >> 2) Modify the cabal file of ansiterminal according to Roman's advice. It >> installed successfully. >> 3) Modify the cabal file of lifted-base where I removed the Include-dirs >> and Includes directives and all references to inlinable.h from the source >> files. >> cabal reports that it can not find the header file or it is damaged. >> The file exists in the include directory but even with >> >> On Fri, May 22, 2015 at 7:58 PM, Michal Antkiewicz < >> mantkiew(a)gsd.uwaterloo.ca> wrote: >> >>> There used to be a big warning "don't download just the ghc. Use the >>> platform. " before. >>> >>> Nowadays, it's best to get MinGHC from >>> >>> https://www.haskell.org/downloads/windows >>> >>> this distribution is best, it'll allow you build the network package and >>> other things like the ansi-terminal, etc. >>> >>> Get the 32bit installer. >>> >>> Michał >>> >>> On Fri, May 22, 2015 at 2:10 PM, <anakreonmejdi(a)gmail.com> wrote: >>> >>>> Hello. >>>> I don't know the MinGHC release. To install ghc I downloaded a windows >>>> install program from https://www.haskell.org/ghc/. >>>> >>>> >>>> Michal Antkiewicz <mantkiew(a)gsd.uwaterloo.ca> writes: >>>> >>>> > Hi, >>>> > >>>> > are you using msys and MinGHC? >>>> > >>>> > For now, I clone ghc-mod @ master and just cabal install it (with >>>> MinGHC >>>> > 0.7.10 and cabal 1.22) within msys. It has not been released yet. >>>> > >>>> > Michał >>>> > >>>> > On Fri, May 22, 2015 at 12:09 PM, Roman Cheplyaka <roma(a)ro-che.info> >>>> wrote: >>>> > >>>> >> On 22/05/15 18:06, Anakreontas wrote: >>>> >> > I tried to instal ghc-mod on Windows with cabal. Two dependencies >>>> fail >>>> >> > to compile because of missing native libraries. >>>> >> > ansi-terminal requires kernel32 and >>>> >> > lifted-base requires inlinable.h. >>>> >> > >>>> >> > Have you solved a similar problem? In Linux it would be easy to >>>> install >>>> >> > the native libraries but in Windows I do not know which programs I >>>> >> > should install and where to find them. >>>> >> >>>> >> You should be able to install ansi-terminal by patching its cabal >>>> file, >>>> >> see https://github.com/feuerbach/ansi-terminal/issues/6 >>>> >> >>>> >> Once again, I'm asking Windows users and developers for their input >>>> on >>>> >> this issue. >>>> >> >>>> >> Roman >>>> >> >>>> >> >>>> >> _______________________________________________ >>>> >> Haskell-Cafe mailing list >>>> >> Haskell-Cafe(a)haskell.org >>>> >> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe >>>> >> >>>> >> >>>> >>>> -- >>>> >>> >>> >> >

1 0

aeson-utils: aeson-0.9 compatibility package
by Adam Bergmark 25 May '15

25 May '15

Hi folks, aeson-0.9 was released today. It changes the behavior of `decode` and `eitherDecode`: They now both allow atomic values on the top level. Previously the behavior was inconsistent, `encode 1 = "1"` but `decode "1" :: Maybe Int = Nothing`. aeson-utils[1] has the functions `decodeV` and `eitherDecodeV` that work like aeson-0.9's `decode` and `eitherDecode` respectively, also with older aeson versions. If you want to stay backwards compatible without having varying semantics based on dependencies, check it out! [1] http://hackage.haskell.org/package/aeson-utils Regards, Adam

1 0

Any recommendations on how to interact with the running state of a program?
by Anatoly Yakovenko 25 May '15

25 May '15

I am playing around with some deep learning algorithms and I need a way to periodically read the intermediate result and change some of the system variables. Any ideas on how to best approach this problem? I was thinking of something basic, like having the program periodically read and write from some files. Thanks, Anatoly

1 0

New Functional Programming Job Opportunities
by Functional Jobs 25 May '15

25 May '15

Here are some functional programming job opportunities that were posted recently: Clojure/Erlang backend engineer/architect at Zoomo http://functionaljobs.com/jobs/8831-clojure-erlang-backend-engineer-archite… Cheers, Sean Murphy FunctionalJobs.com

1 0

ANN: Leksah 0.15
by Hamish Mackenzie 25 May '15

25 May '15

Hi Everyone, The new leksah package is in Hackage and the installers for Windows and OS X are available for download at: https://github.com/leksah/leksah/wiki/download (remember to choose the one that matches your GHC version) Nw in this release: * Support for GHC 7.10.1. * Uses Gtk+ 3.16 on Windows and OS X. * GtkInspector works on all platforms (Ctrl+Shift+I). * Updated build instructions https://github.com/leksah/leksah (OS X and Windows builds are fairly easy now). * Underlining of doctest failures. * Lots of bug fixes. Things that are new since 0.14.0: * Support for using GHCJS. https://www.youtube.com/watch?v=zQnExdDL63c * GtkInspector https://wiki.gnome.org/Projects/GTK%2B/Inspector. * Web Inspector pane (right click on output pane to bring it up. https://trac.webkit.org/wiki/WebInspector * Cabal commands executed by Leksah show up in the log pane. I remain optimistic about the future of Leksah. GTK+ has improved hugely in the last couple of years. Also WebKitGTK+ and GHCJS mean integrating tools into Leksah will be easier than ever. A great many people contribute (mostly indirectly) to Leksah, it is much appreciated and I hope that number will grow. As always, please help! Hop on #leksah or file and issue at https://github.com/leksah/leksah/issues, even if it is just to vent your displeasure with Leksah (every bug report helps). Here are some fun ideas you might like to try working on if you are looking for a way to contribute to Leksah: * Try using Leksah and fix the thing that annoys you most about it. * Figure out if/how ide-backend can be used in Leksah. * Create an XMonad like manager for <div> based panes. Ideally it would work for both GTK+ and GHCJS. * Migrate existing UI components from Gtk+ to GHCJS. * Anything GHCJS based and then add support in Leksah for it. * GHCJS based scripting (make it work for Leksah like elisp for Emacs). * Improve Yi integration. What we have is unusably incomplete. * Add GHCJS backend for Yi (replacing the Pango one). * Improve JavaScript based editor support (currently only CodeMirror is supported and not very well). * Add support for using SoH and/or FPView within Leksah. If you are interested in any of these please let me know and I will do what I can to help you get started. Hamish PS. I am kind of sad Eclipse FP has stalled for now and I hope it restarts soon. Integrating Haskell into existing IDEs (especially those with large user bases) will help sell Haskell to the world. I wish JP well and I am sure he will find something interesting and challenging to work on next.

1 0

Increasing parallelism in a frequency counter
by Andreas Pauley 25 May '15

25 May '15

Hi everyone, I've started playing with some parallel programming in Haskell, mostly by applying the first few examples from Simon Marlow's book to my own toy problem: https://github.com/apauley/parallel-frequency I'm quite pleased with the speedup I have so far, the fastest parallel implementation is just over twice as fast as its sequential counterpart. But looking at the threadscope images from the generated event logs I can see that there are still significant portions of my program that runs on only one core. Is anyone interested to see if we can increase the parallelism in this frequency counter? Kind regards, Andreas -- https://keybase.io/apauley http://za.linkedin.com/in/apauley http://www.meetup.com/lambda-luminaries http://www.meetup.com/Bitcoins-Baby-Bitcoins GPG Public Key: https://keybase.io/apauley/key.asc

3 3