December 2014 - Haskell-Cafe

GHC Heap Profiler time axis
by Anton Dessiatov 19 Dec '14

19 Dec '14

Hello, Haskell-Cafe! For the last 2 days I’m debugging something that looks like a space leak. My program is a rather long-running network service. When I first used heap profiler I was a bit confused with its output that showed that my program was running for only about 0.3-0.4 seconds when I launched it idling for about a minute. That was confusing enough for me to dig into GHC sources trying to figure out the reason behind this and I found that values of the sample times are taken from the mut_user_time function, which returns a time that process spent in user-mode code (outside the kernel) minus time spent on garbage collection. This introduces great amount of unpredictable non-linearity of the heap profile graph x axis, which I personally consider very counterintuitive. So my question is does anybody know why is it done this way? Wouldn’t it be better if x axis would just show a time elapsed since the process started? Kind regards, Anton.

2 1

Re: [Haskell-cafe] cabal-install on openbsd
by Ramakrishnan Muthukrishnan 19 Dec '14

19 Dec '14

On Fri, Dec 19, 2014, at 03:47 PM, Gusoi Constantin - Codrut wrote: > I have the same problem on Ubunutu 14.04, what you need to do is to > use more RAM / SWAP. I have 1 GB of swap. I still get the error. The allocation error is for 1 MegaByte (1048576 bytes). That's what make me really surprised. May be it runs out of memory and fails when it cannot allocate one more megabyte of memory? Ramakrishnan

1 0

cabal-install on openbsd
by Ramakrishnan Muthukrishnan 19 Dec '14

19 Dec '14

Hi, I have a new OpenBSD 5.6 installation on AMD64. The OpenBSD package collection has ghc 7.6.3 and Cabal-install 1.16.0.2. After installing them, I attempted to install a new version of cabal-install using the packaged cabal-install by doing `cabal install cabal-install'. But I get a "ghc: out of memory (requested 1048576 bytes)" error while it is compiling cabal-install 1.20.0.3. I have a ticket created against ghc but I wonder if it is a bug with ghc or cabal-install or OpenBSD or something else. Anyway.. the ticket is here [1]. I can help debug the problem on my OpenBSD box. Thanks Ramakrishnan [1] <https://ghc.haskell.org/trac/ghc/ticket/9904>

1 0

Re: [Haskell-cafe] Typed TH
by oleg＠okmij.org 19 Dec '14

19 Dec '14

Typed meta-programming has a very rich history. For a good overview of various approaches and main problems, please see the Related work section of the following paper http://okmij.org/ftp/tagless-final/TaglessStaged/beyond-journal.pdf There is a special subsection on Typed TH. (The version of that paper with minor revisions will be uploaded in two-three months). The main problems are how to ensure hygiene in the presence of effects, and how to generate polymorphic bindings (let-bindings). Both problems have been essentially solved. (The solution to the second problem has been presented at small meetings so far). The paper explains the combinator library to generate TH code that is ensured to be well-typed, *by construction*. The library permits arbitrary monadic effects. There is no Q monad since it turns out not necessary.

1 0

gracefully invoking `less` using System.Process (and ignoring ctrl-c)
by Guillaume Hoffmann 19 Dec '14

19 Dec '14

Hi, I'm looking at a bug that we should fix for the 2.10 release of darcs. I'd like to make the invocation of the pager (most of the time `less`) as identical to git as possible. Currently darcs has a workaround for the "resource vanished (broken pipe)" problem (that happens when you leave `less` before all of the text has been read by it). The workaround involves creating a temporary file and then passing it to less [1]. This works but it's slow (for big inputs) and ugly IMO. I have a fix proposal [2] that is made of two changes: * use System.Process.createProcess (just to update the code), and * in the "ortryrunning" function (that chains attempts at calling various pagers until one works), ignoring IOError's of the kind ResourceVanished (otherwise we would invoke every pager in the chain if we quit them before they read all their input) The problem is that pressing CTRL-C in `less` now kills everything (darcs+less). Is there a way to fix this? Note that the darcs executable runs inside of the function withSignalHandled [3], which defines its behaviour with respect to various signals. Guillaume [1] http://hub.darcs.net/darcs/darcs-reviewed/browse/src/Darcs/UI/External.hs#4… [2] http://hub.darcs.net/gh/darcs-screened_1/patch/20141210002305-5ef8f [3] http://hub.darcs.net/darcs/darcs-reviewed/browse/src/Darcs/Util/SignalHandl…

3 4

Dockerfile for Yesod
by Jonathan Paugh 18 Dec '14

18 Dec '14

Greetings, all I'm brand new to Docker, and struggled to find resources specific to Yesod. So, once I managed to piece together a Dockerfile, I decided to share it. Hope it helps someone. The Dockerfile is pasted at the bottom. It can be used by dropping it in your Yesod application directory, and replacing "app" with the name of your app. If you're new to Docker like I was, cd to your Yesod app dir, and run $ docker build --tag app-image . $ docker run -d -p 3000:3000 app-image to get started. Your web app should now be available at http://localhost:3000 NB: The approot setting can make a profound and confusing difference if it is mis-configured. I recommend temporarily setting approot to "" so that it will use absolute (site-relative) urls instead of fully specified ones. Happy Holidays, Jon Dockerfile ----------- FROM haskell:7.8 RUN cabal update RUN apt-get update && apt-get install -y libpq5 libpq-dev ADD ./app.cabal /opt/app/app.cabal WORKDIR /opt/app RUN cabal sandbox init RUN cabal install --only-dependencies --max-backjumps=-1 --reorder-goals -j ADD . /opt/app RUN cabal build RUN dist/build/db-import/db-import ENV PATH /opt/app/.cabal-sandbox/bin:$PATH EXPOSE 3000 CMD ["/opt/app/dist/build/app/app", "production", "-p", "3000"]

1 0

Missing `$` like operators in Haskell
by Clinton Mead 18 Dec '14

18 Dec '14

Below are the `$` like operators in Haskell (view in fixed width): --------------------------------------------------------------------- | Function first | Op | Function second | Op | ------------------------------------------------------------------------------- | Plain | (a -> b) -> a -> b | $ | a -> (a -> b) -> b | | | Functor | (a -> b) -> f a -> f b | <$> | f a -> (a -> b) -> f b | | | Applicat| f (a -> b) -> f a -> f b | <*> | f a -> f (a -> b) -> f b | <**> | | Monad | (a -> m b) -> m a -> m b | =<< | m a -> (a -> m b) -> m b | >>= | ------------------------------------------------------------------------------- The "function second" forms I couldn't find for "plain" and "functor". What are the most common operators to used in these places?

9 13

managing local cabal package database
by Carl Eyeinsky 18 Dec '14

18 Dec '14

Dear Cafe readers, I'm wondering what do you guys use as the general method in developing projects using your own private projects? Cabal allows to specify a directory for local packages ('local-repo' in '.cabal/config') with specific structure. Are there any scripts that take a git repo, checkout its tags one by one and put them in some specified directory in the correct way? The crude structure I'm imagining is that I'd like to reuse general purpose stuff I have made myself before, and have it available similar to having it in Hackage. Why I don't want to put my things in the Hackage itself, is that the code often changes fast, and would be little help to others until some later date, when things have calmed down.. (Does this seem a reasonable procedure, or how do you do it?) Cheers, -- Carl Eyeinsky

5 6

Why does Haskell PVP have two values for the major version? "A.B..." and a couple other questions
by Zach Moazeni 18 Dec '14

18 Dec '14

Hello, Forgive me if this is a frequently asked question, I've searched the web and can't find an answer. I originally sent this to the Beginners list and someone said this might get more responses here instead. What is the history that led up to the PVP <https://www.haskell.org/haskellwiki/Package_versioning_policy> specifying two values as the major version? I come from other tools that primarily use http://semver.org/ (or a variant) and I'm confused in what cases I should bump "A" but not "B" (or bump "B" but not "A")? A concrete example: If I make backwards incompatible changes to a package whose latest version is 1.0.x, should the next version be 2.0.x or 1.1.x? What sorts of things should I consider for choosing 2.0 over 1.1 and vice versa? Another question, by far most packages I have encountered either lead with a 0 or a 1 for "A". Does that have some bearing on the long term stability that package users should expect in the future? Finally, if "1.0.x.." is meant to convey a level of long term support, does "B" get rarely used? Since "C" version bumps to include backwards compatible additions, and "D"+ for backwards compatible bug fixes. (I know "D" isn't technically a PVP category, I'm just relating it to the patch version of semver). Thanks for your help! -Zach

14 16

building hmatrix on OSX
by Erik Rantapaa 18 Dec '14

18 Dec '14

I'm trying to build hmatrix-0.16.1.1 on OSX 10.10.1, and I'm running into some undefined C functions and types in src/C/vector-aux.c: - struct random_data - random_r() - initstate_r() A specific error: src/C/vector-aux.c:738:24: error: variable has incomplete type 'struct random_data' struct random_data buffer; On a Linux system they seem to be defined in stdlib.h, but that doesn't seem to be the case on OSX. Is there a way to work around this? FWIW, I'm using GHC for OSX (https://ghcformacosx.github.io/) and I've installed `gsl` via `brew install gsl`.

2 2