Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 329 330 331 332 333 334 335 336 337 338 339 .. 5022
— 502,152 events total
2020-09-30 19:18:49 × wroathe quits (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net) (Ping timeout: 264 seconds)
2020-09-30 19:19:07 × heatsink quits (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net) (Ping timeout: 240 seconds)
2020-09-30 19:19:11 × tbreslein quits (~tbreslein@2a02:8108:140:44f8::901) (Remote host closed the connection)
2020-09-30 19:21:18 × Orbstheorem quits (~roosember@hellendaal.orbstheorem.ch) (Quit: WeeChat 2.3)
2020-09-30 19:21:19 × texasmynsted quits (b9e53b1b@185.229.59.27) (Ping timeout: 245 seconds)
2020-09-30 19:22:24 <p0a> Hello, I have this code: https://paste.tomsmeding.com/Y6afnL1R
2020-09-30 19:22:25 × mu_ quits (~mu@unaffiliated/mu) (Read error: Connection reset by peer)
2020-09-30 19:22:50 → mu_ joins (~mu@unaffiliated/mu)
2020-09-30 19:22:51 <p0a> The error is that you can't have Eq, Show, Read be derived for File', I believe due to HKD. What can I do?
2020-09-30 19:23:27 hackage poly 0.5.0.0 - Polynomials https://hackage.haskell.org/package/poly-0.5.0.0 (Bodigrim)
2020-09-30 19:24:35 → xerox_ joins (~xerox@unaffiliated/xerox)
2020-09-30 19:25:15 <p0a> I think I have to somehow use `deriving' with specific instances but I don't know how to do that
2020-09-30 19:25:31 → heatsink joins (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net)
2020-09-30 19:25:35 <tomsmeding> p0a: the issue is that to derive e.g. Eq for File' f, it is required that Eq (KHD f FilePath), etc. also holds
2020-09-30 19:26:04 <tomsmeding> for simple cases ghc can automatically generate the necessary contexts in the generated instance, but apparently in this case it can't
2020-09-30 19:26:20 <p0a> How can I provide the instances for File? That's the one I want
2020-09-30 19:27:11 <p0a> if I add `deriving' after `type File = File' Identity' I get a parsing error
2020-09-30 19:27:33 <tomsmeding> indeed, adding 'deriving' after a tye synonym declaration is invalid syntax
2020-09-30 19:27:39 <tomsmeding> however, you can do it like this:
2020-09-30 19:27:55 × alp quits (~alp@2a01:e0a:58b:4920:7484:254:28d2:c14c) (Ping timeout: 272 seconds)
2020-09-30 19:28:15 <tomsmeding> https://paste.tomsmeding.com/okCA6gS4
2020-09-30 19:28:32 × AlterEgo- quits (~ladew@124-198-158-163.dynamic.caiway.nl) (Quit: Leaving)
2020-09-30 19:28:32 <tomsmeding> StandaloneDeriving is to allow writing those, well, stand-alone 'deriving' statements
2020-09-30 19:28:58 <tomsmeding> FlexibleInstances and TypeSynonymInstances is to be allowed to write an instance for a type synonym (i.e. File, as opposed to File') in the first place
2020-09-30 19:29:18 → ahmr88 joins (~ahmr88@cpc85006-haye22-2-0-cust131.17-4.cable.virginm.net)
2020-09-30 19:29:21 <p0a> heh thank you
2020-09-30 19:29:26 → tinga joins (~tinga@5-168-208-80-pool.fiber.fcom.ch)
2020-09-30 19:29:27 <p0a> that was tricky
2020-09-30 19:29:48 × mnrmnaugh quits (~mnrmnaugh@unaffiliated/mnrmnaugh) (Remote host closed the connection)
2020-09-30 19:29:49 <tomsmeding> (well really that's only TypeSynonymInstances; FlexibleInstances is then required because when expanding the File type synonym, the instance becomes Eq (File' Identity), and because that Identity is not a type variable, FlexibleInstances is needed)
2020-09-30 19:29:53 × heatsink quits (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net) (Remote host closed the connection)
2020-09-30 19:30:10 → heatsink joins (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net)
2020-09-30 19:30:22 <tomsmeding> generally when you want to derive instances that GHC doesn't do automatically, I try it as a standalone deriving statement
2020-09-30 19:30:37 <tomsmeding> usually that will also not work (though sometimes it will), but then GHC tends to give better error messages
2020-09-30 19:31:04 <p0a> hahah got it
2020-09-30 19:31:06 <tomsmeding> in this case, if you just add StandaloneDeriving and write those stand-alone instances, it will correctly ask you to add the other two extensions
2020-09-30 19:31:12 <p0a> you sort of have to squeeze it out of ghc
2020-09-30 19:31:15 <tomsmeding> :p
2020-09-30 19:31:54 <p0a> I am afraid I am not understanding the extensions very well but I will continue writing code for now
2020-09-30 19:32:07 <tomsmeding> which one?
2020-09-30 19:32:07 <p0a> that is, I do not understand their purpose
2020-09-30 19:32:13 × mu_ quits (~mu@unaffiliated/mu) (Read error: Connection reset by peer)
2020-09-30 19:32:23 <p0a> well let's start with TypeSynonymInstances
2020-09-30 19:32:35 <tomsmeding> without any extensions, haskell only allows you to write an instance for a type of the form T a b c d
2020-09-30 19:32:40 <p0a> Oh wait, that one you explained
2020-09-30 19:32:42 → mu_ joins (~mu@unaffiliated/mu)
2020-09-30 19:32:45 <tomsmeding> where T is a type name, and a b c d ... are type variables
2020-09-30 19:32:47 <p0a> I think the cryptic one is FlexibleInstnaces
2020-09-30 19:32:59 <tomsmeding> yeah
2020-09-30 19:33:20 <tomsmeding> if T is a type synonym, it doesn't match the recipe so you need TypeSynonymInstances (which resolves the name first)
2020-09-30 19:33:39 <tomsmeding> if a b c d, etc, are not type variables but explicitly instantiated to a particular type, then you need FlexibleInstances
2020-09-30 19:33:58 <tomsmeding> for example, you can't write 'instance Show [MyCustomType]' directly
2020-09-30 19:34:13 <tomsmeding> because that's, of course, really 'Show ([] MyCustomType)', and MyCustomType is not a type variable
2020-09-30 19:34:23 <p0a> What is it instead?
2020-09-30 19:34:28 <tomsmeding> well, a type :p
2020-09-30 19:34:29 <tomsmeding> not a variable
2020-09-30 19:34:31 <p0a> right, okay
2020-09-30 19:34:39 <tomsmeding> you _can_ write 'instance Show [a]' because that is 'Show ([] a)' and 'a' _is_ a type variable
2020-09-30 19:34:52 <geekosaur> except it'll overlap
2020-09-30 19:34:58 <tomsmeding> basically, normally you can only create instances for _all_ polymorphic versions of a type
2020-09-30 19:35:09 <tomsmeding> geekosaur: okay yes in this case for Show ([] ...) it will overlap
2020-09-30 19:35:12 <p0a> I see, FlexibleInstances let's you exercise an incision
2020-09-30 19:35:17 <tomsmeding> because there is already Show a => Show [a]
2020-09-30 19:35:28 <tomsmeding> substitute some random other class for Show :p
2020-09-30 19:35:35 <tomsmeding> p0a: kinda :)
2020-09-30 19:36:01 <tomsmeding> once instance contexts start to get complex you might also need UndecidableInstances at some point
2020-09-30 19:36:02 <p0a> Very nice, thank you :) that's an extension I commonly see
2020-09-30 19:36:03 → madjestic joins (~madjestic@86-88-72-244.fixed.kpn.net)
2020-09-30 19:36:20 <tomsmeding> though I don't know the precise rules for that one :)
2020-09-30 19:36:29 <p0a> don't worry it's not on the menu right now
2020-09-30 19:36:31 <geekosaur> "avoid"
2020-09-30 19:36:35 <p0a> lol
2020-09-30 19:36:52 <tomsmeding> well yeah, avoid most things that say "undecidable"
2020-09-30 19:38:25 → wroathe joins (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net)
2020-09-30 19:39:11 × ahmr88 quits (~ahmr88@cpc85006-haye22-2-0-cust131.17-4.cable.virginm.net) (Remote host closed the connection)
2020-09-30 19:39:39 → borne joins (~fritjof@200116b864837e00537d5cc8226f9d9f.dip.versatel-1u1.de)
2020-09-30 19:40:44 <p0a> If I want (Just 3) == 3 and Nothing == 3 to be True, how can I implement this? Say I have a :: Maybe Int and b :: Int, and wish to have the above comparisons
2020-09-30 19:41:09 <p0a> I'm going to look into Control.Monad.Maybe
2020-09-30 19:41:11 <tomsmeding> :t (==)
2020-09-30 19:41:13 <lambdabot> Eq a => a -> a -> Bool
2020-09-30 19:41:15 <p0a> Or Maybe Data.Maybe
2020-09-30 19:41:23 <tomsmeding> Maybe Int does not unify with Int
2020-09-30 19:41:30 → Orbstheorem joins (~roosember@hellendaal.orbstheorem.ch)
2020-09-30 19:41:40 → kindaro joins (1f08d381@h31-8-211-129.dyn.bashtel.ru)
2020-09-30 19:41:42 <p0a> Sure, it doesn't have to be with == , I'm just curious of neat source code to do that
2020-09-30 19:41:49 <{abby}> > maybe 3 id (Just 3)
2020-09-30 19:41:50 <p0a> my code would have if elses
2020-09-30 19:41:51 <lambdabot> 3
2020-09-30 19:41:57 <{abby}> > maybe 3 id Nothing
2020-09-30 19:42:00 <lambdabot> 3
2020-09-30 19:42:12 × mu_ quits (~mu@unaffiliated/mu) (Read error: Connection reset by peer)
2020-09-30 19:42:18 <tomsmeding> > (fromMaybe 3 (Just 5), fromMaybe 3 Nothing)
2020-09-30 19:42:20 <lambdabot> (5,3)
2020-09-30 19:42:22 <{abby}> maybe default id = fromMaybe default, but I'm not sure if fromMaybe is in scope (import it from Data.Maybe)
2020-09-30 19:42:27 <tomsmeding> that's what fromMaybe is for
2020-09-30 19:42:42 → mu_ joins (~mu@unaffiliated/mu)
2020-09-30 19:42:44 <tomsmeding> ah yes
2020-09-30 19:43:10 <tomsmeding> p0a: surely it would be a case expression on the Maybe value, though, not an actual if-then-else?
2020-09-30 19:43:29 <p0a> Right tomsmeding , but you can also use isJust and isNothing
2020-09-30 19:43:30 <{abby}> tomsmeding: better safe than sorry :p I'm never sure what's in lambdabot's scope
←Prev  Next→
Page 1 .. 329 330 331 332 333 334 335 336 337 338 339 .. 5022

All times are in UTC.