- Haskell-Cafe - Haskell.org

LIA bindings?
by ARJANEN Loïc Jean David 13 Jun '18

13 Jun '18

Dear Café, I saw in the report that the standard libraries’ operations don’t conform to LIA and that such operations should be done in an external library. Does such a library exist ? Regards,ARJANEN Loïc

1 0

Re: [Haskell-cafe] Announce: Haskell Platform 8.4.3
by Johannes Waldmann 13 Jun '18

13 Jun '18

> Haskell Platform 8.4.3 Nice. Now I'm wondering - is there an estimated release date for stackage LTS for ghc-8.4.3? That's a genuine question. I know stackage has (a lot) more packages (platform: less than 100?, last LTS: 2474?) so it involves more interfaces (to software, to humans) that can break in many more ways. - J.W.

2 1

Disable default-to-DeriveAnyClass warning with DerivingStrategies?
by Theodore Lief Gannon 11 Jun '18

11 Jun '18

Is there a flag to disable this warning? There's no mystery or surprise to the default, but if I want all three of GeneralizedNewtypeDeriving, DeriveAnyClass and DerivingStrategies enabled, I'm currently stuck with a warning for e.g. deriving (Generic, ToJSON). The only way to use strategy keywords is on standalone deriving lines, which seems unnecessarily verbose.

1 0

how to prevent 'stack' from installing 'ghc'?
by Johannes Waldmann 10 Jun '18

10 Jun '18

Dear Cafe, 'stack' tries to download and install a 'ghc' binary in case 'ghc' in $PATH does not match what is needed for the resolver given in 'stack.yaml' How can I switch this off completely and reliably? I much rather 'stack' just fails when it detects the mismatch. putting 'system-ghc: true' in 'stack.yaml' and '~/.stack/config.yaml' apparently is not enough. 'stack' does warn before downloading, but sometimes I'm not seeing the warning, e.g., because the whole thing is triggered silently from using 'intero'. I do have several ghc executables (built from source) in /opt/ghc/ghc-$version/bin/ghc and I just need to set PATH accordingly. It would be even better if I could tell 'stack' (once and for all) about these locations. I think I understand why 'stack' does what it does (reproducible builds) but I really want to minimize the number of other people's binaries on my machine. I do use the auto-download feature for easier CI builds - so I agree it can be useful. - J.W.

2 1

removing a cookie via ghcjs
by Michael Litchard 08 Jun '18

08 Jun '18

Using ghcjs, I am able to read a cookie using getCookie. There's no corresponding remove/deleteCookie that I can find. Could I get some assistance in how to remove the same cookie I can see using getCookie?

3 3

TyDe 2018: SUBMISSION DEADLINE EXTENDED TO JUNE 13th.
by Niki Vazou 07 Jun '18

07 Jun '18

Hey all, the TyDe submission deadline got extended for one week! CALL FOR CONTRIBUTIONS: DEADLINE EXTENDED TO JUNE 13th. TyDe 2018: Type-Driven Development https://icfp18.sigplan.org/track/tyde-2018 23-29 Sep, 2018, St. Louis, Missouri (co-located with ICFP) ============================================== *Call for Contributions* We welcome all contributions, both theoretical and practical, on a range of topics including: dependently typed programming; generic programming; design and implementation of programming languages, exploiting types in novel ways; exploiting typed data, data dependent data, or type providers; static and dynamic analyses of typed programs; tools, IDEs, or testing tools exploiting type information; pearls, being elegant, instructive examples of types used in the derivation, calculation, or construction of programs. *Important Dates* **Extended paper submission deadline: Wednesday, June 13, 2018** Extended abstract deadline: Wednesday, June 13, 2018 Author notification: Friday, June 29, 2018 Deadline for camera ready version: August 5, 2018 Workshop: Thursday, September 27, 2018 *Program Committee* - Guillaume Allais, Radboud University Nijmegen, Netherlands - Zena M. Ariola, University of Oregon, USA - David Darais, University of Vermont, USA - Richard Eisenberg, Bryn Mawr College, USA (co-chair) - Jennifer Hackett, University of Nottingham, UK - Shin-ya Katsumata, National Institute of Informatics, Japan - Daan Leijen, Microsoft Research, USA - Shin-Cheng Mu, Academia Sinica, Taiwan - Dominic Orchard, University of Kent, UK - Peter-Michael Osera, Grinnell College, USA - Zoe Paraskevopoulou, Princeton University, USA - Alberto Pardo, Universidad de la Republica, Uruguay - Matthieu Sozeau, University of Paris Diderot, Paris 7, France - Niki Vazou, University of Maryland, USA (co-chair) *Submission details* Submissions should fall into one of two categories: - Regular research papers (12 pages) - Extended abstracts (2 pages) The bibliography will not be counted against the page limits for either category. Regular research papers are expected to present novel and interesting research results, and will be included in the formal proceedings. Extended abstracts should report work in progress that the authors would like to present at the workshop. Extended abstracts will be distributed to workshop attendees but will not be published in the formal proceedings. We welcome submissions from PC members (with the exception of the two co-chairs), but these submissions will be held to a higher standard. Submission is handled through HotCRP: https://tyde18.hotcrp.com/ All submissions should be in portable document format (PDF) and formatted using the ACM SIGPLAN style guidelines: http://www.sigplan.org/Resources/Author/ Note that the ACM SIGPLAN style guidelines have changed from previous years! In particular, submissions should use the new ‘acmart’ format and the two-column ‘sigplan’ subformat (not to be confused with the one-column ‘acmlarge’ subformat!). Extended abstracts must be submitted with the label ‘Extended abstract’ clearly in the title. Best, Niki Vazou

1 0

Usage of `pure`
by PY 07 Jun '18

07 Jun '18

Hello, Dear Cafe! Consider, I have big application with monad and applicative related code. What problems I can hit if I will replace all `return`s with `pure` (in `do`-blocks, anywhere). So, `return` will be totally eliminated. === Best regards, Paul

3 2

Parsing a matrix table with Data.Frames
by Tyson Whitehead 06 Jun '18

06 Jun '18

I have a case where some of my data is stored in CSV matrix form like this example of a table telling whether person i has pet j ,cat,dog,goldfish bill,true,true,false sue,false,true,true fred,false,false,true I can't see any obvious way to read this sort of thing in with the Data.Frames package. Would love to be told differently. Thanks! -Tyson

2 2

Wanted: New Admin for Planet.Haskell.Org
by Gershom B 05 Jun '18

05 Jun '18

Dear all — the planet.haskell.org server is a low-maintenance feed-aggregator that uses the planet venus software to pull together feeds of haskell-related content. I think it is a great resource, and I’ve used it for years. For some time the planet(a)haskell.org address hadn’t been well-monitored, meaning that requests for adding new rss feeds had languished. I’ve gone through and tried to act on the outstanding requests, and will be doing so for the time being. However, I have a lot on my plate, and would like to pass this responsibility on to a volunteer. It basically means, on the occasional recipt of an email, logging into the server, updating a config file, running a command, and logging back out again. If you’re interested in helping out with this, please contact me. Cheers, Gershom

2 1

Re: [Haskell-cafe] No Enum for (,), no Enum or Bounded for Either
by Olaf Klinke 04 Jun '18

04 Jun '18

>Also note that what you're talking about is a special type of objects, >namely > > type BoundedEnum a = (Bounded a, Enum a) -- using ConstraintKinds > >(I'm sure the mathematicians have a better name for this) Indeed the Prelude states no laws for Bounded. But maybe we all agree on this: forall a. fromEnum minBound <= fromEnum a <= fromEnum maxBound With this law and the rule fromEnum.toEnum=id (for those integers that this type's toEnum accepts) we can say that (Bounded a,Enum a) is the class of all types which have a distinguished finite subset. If one assumes further the law toEnum.fromEnum=id (which is violated e.g. by Double) then (Bounded a,Enum a) are precisely the constructively finite types, that is, the types for which one can exhibit an isomorphism with a finite cardinal. Note that in constructive mathematics, a set may be known to be finite in the sense that the assumption it is infinite is absurd, yet one can not name an isomorphism with a finite cardinal. [*] As Jon Fairbairn pointed out, (Int,Int) can be given an Enum instance. The same is true for every type that is build from types that already have an Enum instance using only sum and product. >This is all true, but I was implicitly assuming that Bounded and Enum ought >to agree with Ord. Consider this: From every type's Enum instance you can derive an Ord instance: compare = compare `Data.Function.on` fromEnum -- Olaf [*] In the text [1] by Andrej Bauer I found this funny theorem: Excluded middle is equivalent to the statement that subsets of finite sets are finite. [1] http://www.ams.org/journals/bull/2017-54-03/S0273-0979-2016-01556-4/S0273-0…

2 1