November 2016 - Haskell-Cafe

commenting out a line of code
by Johannes Waldmann 15 Nov '16

15 Nov '16

Dear Cafe - Sometimes, a comment can change the meaning of a program. No? Well, but "commenting out" a line of code certainly does. So, can you guess what this program will output? main = do if True then do putStrLn "A" else do -- putStrLn "B" putStrLn "done" I nominate this for inclusion in a hypothetical Haskell equivalent of http://www.javapuzzlers.com/ - J.

6 5

.haddock files
by Patrick Pelletier 15 Nov '16

15 Nov '16

Where can I find documentation about .haddock files? What is in them? Is it possible to generate a .haddock file from source code and then later produce HTML documentation from the .haddock file? Or is that not what it's for? The documentation for Haddock doesn't even seem to mention .haddock files, unless I'm missing something. I tried subscribing to the Haddock mailing list so I could post this question there, but at http://projects.haskell.org/cgi-bin/mailman/subscribe/haddock I got the message: "We're sorry, we hit a bug! Mailman experienced a very low level failure and could not even generate a useful traceback for you. Please report this to the Mailman administrator at this site." Anybody know what's going on with that? Thanks, --Patrick

2 1

bizarre behaviour
by Lawrence Wickert 14 Nov '16

14 Nov '16

Hi, Ok, I am just learning Haskell. So, every day I open a text editor, a pdf file of a Haskell book and a terminal to run ghci. I resize the windows and go on my merry way. Yesterday I decided to install this script https://raw.githubusercontent.com/hotice/webupd8/master/session with the object to save my session windows and restore them the next day. Worked fine although I had to authorize some keyring popup box before it would work. Now I opened Google Chrome and an iHeart radio link. At this point the following appeared at the ghci prompt: Prelude> Vector smash protection is enabled. ghci continues to run fine. If I close and then re-open the iHeart link the message reappears. I am using Ububtu 16.04LTS and I don't use drugs or alcohol. Any ideas as to what is going on? Larry

2 1

Re: [Haskell-cafe] Named arguments to functions?
by Sylvain Henry 14 Nov '16

14 Nov '16

Hi Saurabh, Implicit parameters already look like named arguments: > :set -XImplicitParams > let f = ?x + ?y > let { ?y = 5; ?x = 10 } in f 15 > let g = f where { ?x = 5; ?y = 10 } > g 15 Perhaps we could extend TypeApplications to support the following more natural syntax: > f @(?y = 5) @(?x = 10) which would be syntaxic sugar for the code above. Sylvain On 13/11/2016 14:38, Saurabh Nanda wrote: > Thank you for your reply. > > If you are interested in something like this, you may want to > start a discussion of the design on one of the other lists, or > perhaps write up a proposal for GHC. > > > While I would like the general area of records getting better in > Haskell, I don't think I have the necessary technical chops to lead a > GHC proposal around this. Do you know any proposal/extension in GHC > that is already trying to solve: (a) records, and/or (b) named arguments? > > The proposals currently under discussion are on the `github` URL. > > > Possible to update the prime.haskell.org <http://prime.haskell.org> > website to reflect the current state of affairs? > > -- Saurabh. > > > On 10/11/2016 22:59, Iavor Diatchki wrote: >> Hello, >> >> I would say that this is a too complex of a change to discuss for >> Haskell Prime without implementation or prior experience. If you >> are interested in something like this, you may want to start a >> discussion of the design on one of the other lists, or perhaps write >> up a proposal for GHC. >> >> The proposals currently under discussion are on the `github` URL. >> >> Cheers, >> -Iavor >> >> >> >> >> >> On Thu, Nov 10, 2016 at 2:16 AM, Saurabh Nanda >> <saurabhnanda(a)gmail.com <mailto:saurabhnanda@gmail.com>> wrote: >> >> Hi Everyone, >> >> I was going through >> https://prime.haskell.org/wiki/Process#Proposals >> <https://prime.haskell.org/wiki/Process#Proposals> and noticed >> the following: >> >> > Ideally the language change should be implemented already, so >> that experience gained with the implementation can inform the >> discussion >> >> Is this a hard requirement? I would like to see the ability to >> pass named arguments to functions somehow (most probably through >> anonymous records). I'm unaware of any existing language >> extension that solves this problem. Does this render the >> discussion out-of-bounds from a Haskell Prime perspective? Has >> this already been proposed and considered? >> >> Also, going by what is given on the Wiki, I'm confused about >> where to check the current status of proposals that are being >> considered. Which is the primary link: >> >> * >> https://prime.haskell.org/query?status=new&status=assigned&status=reopened&… >> <https://prime.haskell.org/query?status=new&status=assigned&status=reopened&…> >> * Or, https://github.com/haskell/rfcs >> <https://github.com/haskell/rfcs> >> >> -- Saurabh. >> >> _______________________________________________ >> Haskell-prime mailing list >> Haskell-prime(a)haskell.org <mailto:Haskell-prime@haskell.org> >> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-prime >> <http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-prime> >> >>

1 0

Re: [Haskell-cafe] megaparsec 5.1.2 - missing doc links on hackage
by Mark Karpov 13 Nov '16

13 Nov '16

Hey Charles, I uploaded the docs manually. If you run into this problem again remember that there is usually a version of the docs available on Stackage as well: https://www.stackage.org/nightly-2016-11-13/package/megaparsec-5.1.2 Best, Mark Charles Martin writes: > Just drawing your attention to the missing documentation links for the > current version: > https://hackage.haskell.org/package/megaparsec-5.1.2 > > Cheers - Charles

1 0

Cross compilation
by Jan von Löwenstein 12 Nov '16

12 Nov '16

Hi, I have a Mac OS X workstation and want to compile a cli for Linux (and potentially Windows) in addition to OS X. I could not find anything but https://ghc.haskell.org/trac/ghc/wiki/CrossCompilation. This has an explanation how I can build a cross compiling GHC for a specific combination of build and target system/architecture. Stack has an open issue regarding cross compilation. Is that really all there is out there? Regards Jan

3 4

TH GadtC parameter
by Paolino 12 Nov '16

12 Nov '16

Hello, I couldn't find the evidence for the first parameter of GadtC in Con data type to be a [Name] in place of a simpler Name. https://hackage.haskell.org/package/template-haskell-2.11.0.0/docs/Language… How should I match a normal gadt constructor syntax with that? Thank you Paolino

2 1

GHC API: type class instance check
by J. Stutterheim 11 Nov '16

11 Nov '16

Hi all, I'm writing a core-to-core GHC plugin and have two questions: - Given some t :: Type, how can I check whether there exists an instance of some type class C for t? - How can I check whether some type class C exists, and if so, how can I get a value c (presumably :: Class) of this class? Thanks! - Jurriën

3 4

Announcing Req -- HTTP client library
by Mark Karpov 11 Nov '16

11 Nov '16

Hi, I would like to announce a new HTTP client library called Req, or rather a new API to tried-and-true http-client. The Haddock is available here: https://www.stackage.org/haddock/nightly-2016-11-11/req-0.1.0/Network-HTTP-… It happens to be a complete tutorial on its own, so you can just read through it and it will be clear what it's all about and what kind of API is proposed. Hackage: https://hackage.haskell.org/package/req GitHub: https://github.com/mrkkrp/req Best, Mark

1 0

Getting Haddock docs onto Hackage (using Stack)
by de Buitléir, Amy 11 Nov '16

11 Nov '16

A package I uploaded to Hackage a week ago still has no docs. (It says "Docs pending" and "Build status unknown".) What is the best way to get Haddock documentation for a package onto Hackage if you are using Stack? Once upon a time, it used to be built automatically after the package was uploaded, and if it wasn't, you could find out the reason why by checking the build logs. When that stopped working, I used a script called hackage-docs.sh that I got from somewhere. But now I've migrated from Cabal to Stack, and the old script no longer works. I can modify it to work with Stack, but is there a better way? Should I instead use the "delete documentation and trigger rebuild" button on Hackage? Here's the package in question: http://hackage.haskell.org/package/grid-7.8.7

5 6