Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 330 331 332 333 334 335 336 337 338 339 340 .. 5022
— 502,152 events total
2020-09-30 19:43:40 <tomsmeding> {abby}: fair :p
2020-09-30 19:43:51 <kindaro> In Template Haskell, how can I convert an `Exp` to a `Pat`? I have a `Type` and I want to pattern match on its literal name as given by `Type.Reflection.tyConName`.
2020-09-30 19:44:01 <tomsmeding> p0a: isJust and fromJust? while there are legitimate use cases for fromJust, think thrice when using it
2020-09-30 19:44:12 <p0a> (x == \x -> fromMaybe x y) is then what I'd do I think
2020-09-30 19:44:19 <p0a> thank you
2020-09-30 19:44:21 <tomsmeding> because usually you're better off with a case expression or an existing combinator
2020-09-30 19:44:23 <kindaro> So, I can construct an `Exp` with a quasi quote, but it seems I cannot construct a `Pat`.
2020-09-30 19:44:32 <tomsmeding> less partiality more better
2020-09-30 19:44:41 <p0a> right, that is true
2020-09-30 19:45:12 <kindaro> For example, this would not work: `getTypeNameLiteral subject = (tyConName ∘ typeRepTyCon) (typeRep @ $(return subject))`
2020-09-30 19:45:54 <kindaro> This compiles but gives me a `Q Exp`: `getTypeNameLiteral subject = [|(tyConName ∘ typeRepTyCon) (typeRep @ $(return subject))|]`
2020-09-30 19:46:09 × filwisher quits (~filwisher@cpc76738-dals23-2-0-cust186.20-2.cable.virginm.net) (Ping timeout: 260 seconds)
2020-09-30 19:46:27 <kindaro> This does not compile: `getTypeNameLiteral subject = [p|(tyConName ∘ typeRepTyCon) (typeRep @ $(return subject))|]`
2020-09-30 19:46:36 <tomsmeding> what kind of pattern would you want it to be
2020-09-30 19:46:47 <tomsmeding> string? data constructor?
2020-09-30 19:46:58 <tomsmeding> can't you create the Pat value yourself
2020-09-30 19:47:01 <tomsmeding> https://hackage.haskell.org/package/template-haskell-2.16.0.0/docs/Language-Haskell-TH.html
2020-09-30 19:47:43 <kindaro> For `Type` = `(ConT GHC.Types.Int)`, I would like `"Int"`, and so on.
2020-09-30 19:48:19 <kindaro> I cannot create a `Pat` value because I have no idea what string `tyConName` would produce.
2020-09-30 19:48:32 <kindaro> I have to first run `tyConName` somehow.
2020-09-30 19:49:04 <tomsmeding> oh your problem is to get the TyCon to pass to tyConName?
2020-09-30 19:49:10 <tomsmeding> from what do you want to get it
2020-09-30 19:49:18 <tomsmeding> a value of that type?
2020-09-30 19:49:19 <kindaro> From `Type`.
2020-09-30 19:49:50 <kindaro> This `Type`: https://hackage.haskell.org/package/template-haskell-2.16.0.0/docs/Language-Haskell-TH.html#t:Type
2020-09-30 19:50:01 × ADG1089 quits (uid455466@gateway/web/irccloud.com/x-wfsqzgfoscwuqfvc) (Quit: Connection closed for inactivity)
2020-09-30 19:50:05 × JoelMcCracken[m] quits (joelmccrac@gateway/shell/matrix.org/x-fkrjmkjfhjvsypdi) (Quit: authenticating)
2020-09-30 19:50:12 → JoelMcCracken[m] joins (joelmccrac@gateway/shell/matrix.org/x-nmxpzgkyreszwsoo)
2020-09-30 19:50:15 <kindaro> So, I first have to convert this `Type` to an actual expression, at which point it becomes incompatible with `Pat`.
2020-09-30 19:50:25 <tomsmeding> but, a type can have many constructors, right?
2020-09-30 19:50:32 <tomsmeding> data A = A | B
2020-09-30 19:50:39 <kindaro> It has one type constructor.
2020-09-30 19:50:42 <tomsmeding> if the Type is A, do you want the TyCon for A or for B?
2020-09-30 19:50:43 <tomsmeding> ah
2020-09-30 19:50:43 <kindaro> In your case `A`.
2020-09-30 19:50:50 → nobody239419341 joins (5e7be5a0@94.123.229.160)
2020-09-30 19:51:01 <tomsmeding> can't you use generics
2020-09-30 19:51:18 <kindaro> Like how?
2020-09-30 19:51:54 → conal joins (~conal@64.71.133.70)
2020-09-30 19:51:57 <kindaro> As you may remember, my problem can be put as constructing a large case expression.
2020-09-30 19:52:06 <tomsmeding> right
2020-09-30 19:52:13 <kindaro> The way to do that is via Template Haskell.
2020-09-30 19:52:13 × mu_ quits (~mu@unaffiliated/mu) (Read error: Connection reset by peer)
2020-09-30 19:52:16 <aldessa> by the way for my earlier problem, the solution was 'stack ls --dependencies --no-include-base --license' to see all the licenses
2020-09-30 19:52:27 hackage deriving-compat 0.5.10 - Backports of GHC deriving extensions https://hackage.haskell.org/package/deriving-compat-0.5.10 (ryanglscott)
2020-09-30 19:52:29 × ransom quits (~c4264035@undergraduate-jvossen-9690.mines.edu) (Quit: My MacBook has gone to sleep. ZZZzzz…)
2020-09-30 19:52:41 → mu_ joins (~mu@unaffiliated/mu)
2020-09-30 19:52:44 × mirrorbird quits (~psutcliff@2a00:801:3fb:53b9:b09b:cd4e:4e4a:44a9) (Ping timeout: 244 seconds)
2020-09-30 19:52:49 <kindaro> I made some progress, but it is unclear how I can associate a literal pattern with a `Type`.
2020-09-30 19:52:51 <nobody239419341> hello can someone help me understand this function signrature `palindrome = reverse >>= (==)` I'm beginner to haskell
2020-09-30 19:53:45 <Cheery> :t reverse >>= (==)
2020-09-30 19:53:45 <sm[m]> aldessa: stack ls dependencies --no-include-base --license ... very cool, thanks
2020-09-30 19:53:46 <lambdabot> Eq a => [a] -> Bool
2020-09-30 19:53:50 <tomsmeding> kindaro: I know how to write a function 'magic :: Generic a => String' that produces the name of the first type constructor of 'a'; would that be sufficient for you?
2020-09-30 19:53:53 <monochrom> It is not for beginners. Please don't worry about that one. Please learn and play with basic examples first.
2020-09-30 19:54:09 <p0a> nobody239419341: I believe that is not a signature. I think signatures refer to the types after ::
2020-09-30 19:54:12 <monochrom> Whoever showed you "reverse >>= (==)" did not respect that you are a beginner.
2020-09-30 19:54:13 <tomsmeding> kindaro: I'm not familiar enough with TH to know whether generics play nicely with it
2020-09-30 19:54:44 <nobody239419341> I was self learning from https://wiki.haskell.org/99_questions/Solutions/6 here and would like to understand that solution
2020-09-30 19:54:46 <kindaro> I expect that they do not.
2020-09-30 19:54:54 <nobody239419341> I wrote first solution myself without looking to the solutions page
2020-09-30 19:56:04 <nobody239419341> I know how would I look to the function signatures by ` :t ` in ghci
2020-09-30 19:56:06 <Cheery> :t (>>=) reverse
2020-09-30 19:56:08 <lambdabot> ([a] -> [a] -> b) -> [a] -> b
2020-09-30 19:56:14 <Cheery> :t (==)
2020-09-30 19:56:16 <lambdabot> Eq a => a -> a -> Bool
2020-09-30 19:56:58 <tomsmeding> kindaro: 'tyConName (typeRepTyCon (typeRep @T))' is of type String
2020-09-30 19:57:01 sm[m] discovers a few weirdo packages using BSD-2 (megaparsec & text)
2020-09-30 19:57:09 <tomsmeding> not sure where Exp comes into the discussion
2020-09-30 19:57:32 <tomsmeding> also I now realise that you're talking about the _type_ constructor, not a _data_ constructor :p
2020-09-30 19:58:16 × ddellacosta quits (~dd@86.106.121.168) (Quit: WeeChat 2.8)
2020-09-30 19:58:52 tomsmeding is surprised about the amount of reflection that ghc gives you with Type.Reflection, Data.Typeable and templatehaskell
2020-09-30 19:59:14 <kindaro> nobody239419341  This is some advanced Haskell, I have been doing it for some years but it took me effort to understand how your function works. You may wish to consult the Monad instance for functions. https://hackage.haskell.org/package/base-4.14.0.0/docs/src/GHC.Base.html#line-979
2020-09-30 19:59:16 <Cheery> tomsmeding: any short introduction to Type Reflection you'd recommend?
2020-09-30 19:59:28 <Cheery> or data typeable, those sound interesting.
2020-09-30 20:00:26 × Quarl quits (~Quarl@94.191.138.174.mobile.tre.se) (Read error: Connection reset by peer)
2020-09-30 20:00:30 <tomsmeding> Cheery: I learned about the existence of Type.Reflection when kindaro mentioned it
2020-09-30 20:00:37 <tomsmeding> I'm afraid I'm not the right person to ask :p
2020-09-30 20:00:39 <Cheery> nobody239419341: that implementation for checking palindrome is vicious.
2020-09-30 20:00:41 <nobody239419341> kindaro thank you I will look to it
2020-09-30 20:00:43 <kindaro> nobody239419341  As you may see, it does exactly what you need it to do. _«Why»_ is another question. It is fair to say that I never ever used the monad instance for functions.
2020-09-30 20:01:01 <p0a> Oh so that's what it is using. Wow, tricky
2020-09-30 20:01:37 <Cheery> functions make a valid monad with their last argument.. it's a bit difficult to understand but it's valid.
2020-09-30 20:01:38 <kindaro> Seeing that `Reader` is really a function, this may give some intuition.
2020-09-30 20:01:45 <sm[m]> ...and one weirdo using ISC (glguy :). Interesting, I didn't know about ISC
2020-09-30 20:01:58 hackage linear 1.21.2 - Linear Algebra https://hackage.haskell.org/package/linear-1.21.2 (ryanglscott)
2020-09-30 20:02:09 <sm[m]> https://www.gnu.org/licenses/license-list.en.html#ISC
2020-09-30 20:02:10 × mu_ quits (~mu@unaffiliated/mu) (Read error: Connection reset by peer)
2020-09-30 20:02:19 <Cheery> nobody239419341: [a] -> [a] is the type signature for reverse.
2020-09-30 20:02:31 <p0a> Cheery: I think this explains some of it? https://mjoldfield.com/atelier/2014/07/monads-fn.html
2020-09-30 20:02:33 <nobody239419341> yes
2020-09-30 20:02:34 <p0a> Or am I wrong?
2020-09-30 20:02:42 → mu_ joins (~mu@unaffiliated/mu)
2020-09-30 20:02:52 <tomsmeding> sm[m]: what about https://hackage.haskell.org/package/dependent-map
2020-09-30 20:02:58 × eyenx quits (~eyenxeyen@unaffiliated/eye/x-1653358) (Quit: killed)
2020-09-30 20:03:05 → xcmw joins (~textual@dyn-72-33-0-89.uwnet.wisc.edu)
2020-09-30 20:03:07 <Cheery> nobody239419341: now imagine that's something like F [a] where F x = [a] -> x
2020-09-30 20:03:50 <nobody239419341> what does it mean to apply the infix operator `->` for function definition
2020-09-30 20:04:04 → ddellacosta joins (~dd@86.106.121.168)
2020-09-30 20:04:26 <kindaro> tomsmeding, so the story goes: I get the list of `Type` values by examining `reifyInstances`. Now I must use `tyConName` on the corresponding types. But this correspondence is not trivial to establish.
←Prev  Next→
Page 1 .. 330 331 332 333 334 335 336 337 338 339 340 .. 5022

All times are in UTC.