Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 .. 5022
— 502,152 events total
2020-11-16 06:33:49 <moet> instance Num 'A where ... fails because `'A :: T` but `Num :: * -> Constraint`
2020-11-16 06:34:04 × towel quits (~towel@unaffiliated/towel) (Quit: ZNC - https://znc.in)
2020-11-16 06:34:05 × {abby} quits (~{abby}@unaffiliated/demhydraz) (Quit: Ping timeout (120 seconds))
2020-11-16 06:34:11 → {abby}2 joins (~{abby}@unaffiliated/demhydraz)
2020-11-16 06:34:38 × guest1116 quits (~user@49.5.6.87) (Ping timeout: 265 seconds)
2020-11-16 06:35:02 → towel joins (~towel@unaffiliated/towel)
2020-11-16 06:35:07 × pavonia quits (~user@unaffiliated/siracusa) (Ping timeout: 260 seconds)
2020-11-16 06:35:09 pavonia_ is now known as pavonia
2020-11-16 06:38:47 <moet> nevermind.. the problem was a stuck type family upstream ..
2020-11-16 06:39:43 → coot_ joins (~coot@37.30.49.253.nat.umts.dynamic.t-mobile.pl)
2020-11-16 06:40:43 × coot quits (~coot@37.30.49.253.nat.umts.dynamic.t-mobile.pl) (Ping timeout: 256 seconds)
2020-11-16 06:40:44 coot_ is now known as coot
2020-11-16 06:41:12 × Sanchayan quits (~Sanchayan@171.61.102.109) (Quit: leaving)
2020-11-16 06:41:26 <moet> now i have questions about stuck type families if anybody has their polykinds hat on
2020-11-16 06:41:32 → quarters joins (~quarters@unaffiliated/quarters)
2020-11-16 06:41:40 <moet> otherwise i'll just figure it out
2020-11-16 06:43:05 <quarters> hello. I'm following http://learnyouahaskell.com/syntax-in-functions and noticed that type declarations are shown above every function. Is this something that can be replicated in the ghci repl? when I try it, I get <interactive>:18:1: error:
2020-11-16 06:43:05 <quarters> • No instance for (Show (Integer -> String))
2020-11-16 06:43:05 <quarters> arising from a use of ‘print’
2020-11-16 06:43:05 <quarters> (maybe you haven't applied a function to enough arguments?)
2020-11-16 06:43:06 <quarters> • In a stmt of an interactive GHCi command: print it
2020-11-16 06:43:30 → o1lo01ol1o joins (~o1lo01ol1@bl11-140-216.dsl.telepac.pt)
2020-11-16 06:43:30 × matthew- quits (~matthew@smtp.wellquite.org) (Ping timeout: 272 seconds)
2020-11-16 06:43:50 → matthew- joins (~matthew@smtp.wellquite.org)
2020-11-16 06:47:12 <moet> quarters: yes, there are two ways https://termbin.com/zngt
2020-11-16 06:47:47 <moet> quarters: one is to type a one-liner lambda on the rhs of a let, parenthesize the whole lambda, and then give it a type
2020-11-16 06:47:52 × merijn quits (~merijn@83-160-49-249.ip.xs4all.nl) (Ping timeout: 256 seconds)
2020-11-16 06:48:05 × o1lo01ol1o quits (~o1lo01ol1@bl11-140-216.dsl.telepac.pt) (Ping timeout: 256 seconds)
2020-11-16 06:48:14 <moet> quarters: the other is to write a multi-line-function just as you would in a file, by preceding it with :{ and then following it with :}
2020-11-16 06:48:24 <moet> see the paste at https://termbin.com/zngt
2020-11-16 06:48:37 → Sanchayan joins (~Sanchayan@171.61.102.109)
2020-11-16 06:49:34 <quarters> moet: awesome. thank you!
2020-11-16 06:50:01 × falafel_ quits (~falafel@50.216.77.242) (Ping timeout: 264 seconds)
2020-11-16 06:52:58 <moet> does making a type-family have a result of kind `k` automatically make the type family "stuck"? is there a way to inspect the result?
2020-11-16 06:55:03 <Axman6> :kind! should show you it in as evaluated for as possible
2020-11-16 06:55:19 → justanotheruser joins (~justanoth@unaffiliated/justanotheruser)
2020-11-16 06:55:32 <Axman6> (that's probably not a precide definition of what it does, but :kind! is probably what you're after)
2020-11-16 06:56:02 → avn joins (~avn@78-56-108-78.static.zebra.lt)
2020-11-16 06:56:22 → santa_1 joins (~santa_@185.163.110.116)
2020-11-16 06:58:19 <moet> Axman6: hrm.. that tool does help me to identify the problem but my question is more about what the meaning of a polymorphic typeclass is
2020-11-16 06:58:37 <moet> Axman6: if you have a sec, take a look at this minimal example: https://termbin.com/n79h
2020-11-16 06:59:13 <moet> in the example, i've made a type family that seems to be useless, because it is stuck due to its polymorphic result
2020-11-16 07:00:48 <dminuoso> moet: No, typeclasses can have instances for non-* types as well. See Functor
2020-11-16 07:01:12 <dminuoso> Oh, scrolled up again.
2020-11-16 07:01:52 <Axman6> this is definitely beyond my knowledge of these things
2020-11-16 07:02:07 × wroathe quits (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net) (Quit: leaving)
2020-11-16 07:02:51 <moet> dminuoso: heh, yeah, i realized the issue with that typeclass was actually the stuck type family not resolving to the type which matched the typeclass :)
2020-11-16 07:03:09 <moet> Axman6: thanks for taking a look anyhow.. it's beyond my knowledge too!
2020-11-16 07:03:17 <dminuoso> Im surprised this typechecks at all
2020-11-16 07:03:37 <dminuoso> Or kindchecks, rather
2020-11-16 07:04:10 <moet> dminuoso: it does admittedly require like 4 extensions
2020-11-16 07:04:27 <keltono> try `:t funcName`
2020-11-16 07:04:35 <keltono> whoops, wrong channel
2020-11-16 07:04:41 <dminuoso> moet: which ones?
2020-11-16 07:04:48 <dminuoso> TypeFamilies and PolyKinds presumably
2020-11-16 07:04:50 <dminuoso> Which else?
2020-11-16 07:04:56 <moet> DataKinds and KindSignatures
2020-11-16 07:05:08 <dminuoso> Ah, well those make sense too
2020-11-16 07:05:12 <moet> Though DataKinds is probably only because i used tuple on RHS
2020-11-16 07:05:15 <dminuoso> Right
2020-11-16 07:05:25 <dminuoso> You could alternatively say
2020-11-16 07:05:32 <dminuoso> % type family TF (a :: *) :: k
2020-11-16 07:05:32 <yahb> dminuoso:
2020-11-16 07:05:40 <dminuoso> % type instance TF Bool = Type
2020-11-16 07:05:40 <yahb> dminuoso: ; <interactive>:102:25: error:; Ambiguous occurrence `Type'; It could refer to; either `Language.Haskell.TH.Type', imported from `Language.Haskell.TH' (and originally defined in `Language.Haskell.TH.Syntax'); or `Data.Kind.Type', imported from `Data.Kind' (and originally defined in `GHC.Types')
2020-11-16 07:05:57 <dminuoso> % type instance TF Bool = Data.Kind.Type
2020-11-16 07:05:57 <yahb> dminuoso:
2020-11-16 07:06:02 <dminuoso> % type instance TF Int = Data.Kind.Type -> Data.Kind.Type
2020-11-16 07:06:03 <yahb> dminuoso:
2020-11-16 07:07:42 <moet> % :kind! TF Bool
2020-11-16 07:07:43 <yahb> moet: k; = TF Bool
2020-11-16 07:07:47 <moet> % :kind! TF Int
2020-11-16 07:07:47 <yahb> moet: k; = TF Int
2020-11-16 07:07:52 <moet> also stuck?
2020-11-16 07:08:06 <dminuoso> I dont think "stuck" is the right terminology here
2020-11-16 07:08:19 <dminuoso> % :kind! TF Int
2020-11-16 07:08:19 <yahb> dminuoso: forall {k}. k; = TF Int
2020-11-16 07:08:33 <dminuoso> (I added :set -fprint-explicit-foralls in a query)
2020-11-16 07:08:46 <dminuoso> So it returns a polymorphic kind
2020-11-16 07:08:58 <moet> yes, i think it's because k cannot be both * and (* -> *) at the sametime
2020-11-16 07:09:04 <dminuoso> The confusing thing is why the tf instances type check at all, because `Type` is certainly *not* polymorphic
2020-11-16 07:09:46 <dminuoso> But then again, I only know how you can have polymorphic kinds in negative positions (say like Const)
2020-11-16 07:09:54 <dminuoso> Or *polymorphic types rather
2020-11-16 07:10:18 <dminuoso> Let's say we did
2020-11-16 07:10:29 <dminuoso> % data C (forall k. t :: k) = C
2020-11-16 07:10:30 <yahb> dminuoso: ; <interactive>:111:9: error:; Unexpected type `forall k. t :: k'; In the data declaration for `C'; A data declaration should have form; data C a = ...
2020-11-16 07:11:02 <dminuoso> Oh that's not allowed?
2020-11-16 07:11:37 <glguy> % :kind! TF Bool :: *
2020-11-16 07:11:37 <yahb> glguy: *; = *
2020-11-16 07:11:48 <glguy> % :kind! [TF Bool]
2020-11-16 07:11:48 <yahb> glguy: *; = [*]
2020-11-16 07:12:11 <moet> % :kind! [TF Int]
2020-11-16 07:12:11 <yahb> moet: *; = [* -> *]
2020-11-16 07:12:15 <moet> oh, wow..
2020-11-16 07:12:17 <dminuoso> Is it possible this tyfam crosses universes and returns a proper kind?
2020-11-16 07:12:29 <dminuoso> (Sort of TypeInType)
2020-11-16 07:12:50 <dminuoso> No this is still weird
2020-11-16 07:12:51 <glguy> You have to know the desired result kind to know which TF instance to use
2020-11-16 07:13:14 <glguy> % type instance TF Bool = (->)
2020-11-16 07:13:14 <yahb> glguy:
←Prev  Next→
Page 1 .. 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 .. 5022

All times are in UTC.