Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 872 873 874 875 876 877 878 879 880 881 882 .. 5022
— 502,152 events total
2020-10-26 15:32:14 <thblt> tomsmeding: isn't this a compile error?
2020-10-26 15:32:34 <tomsmeding> not if they were actually the same symbol in the beginning
2020-10-26 15:32:44 <thblt> Ha, haven't thought of that.
2020-10-26 15:32:52 <tomsmeding> e.g. Foo exports foo, A and B re-export foo from Foo, and Main imports A and B and uses foo
2020-10-26 15:33:14 tomsmeding now wonderse what ghc does if you make both the A and the B import of foo explicit
2020-10-26 15:33:22 <thblt> And ghc is happy with that? As long as the symbol's definition isn't ambiguous?
2020-10-26 15:33:27 <tomsmeding> yup
2020-10-26 15:33:37 → texasmynsted joins (~texasmyns@99.96.221.112)
2020-10-26 15:33:39 <thblt> Makes sense.
2020-10-26 15:34:44 <tomsmeding> if Main has 'import A (foo)' 'import B (foo)', ghc is completely fine
2020-10-26 15:34:57 <tomsmeding> I guess that's one way to make the output of our imaginary tool unambiguous?
2020-10-26 15:35:41 <thblt> This tool pretends to do that https://github.com/serokell/importify
2020-10-26 15:35:44 <tomsmeding> there is already a tool to prune unused imports (https://hackage.haskell.org/package/fix-imports), but it doesn't convert implicit imports (import A) to explicit ones (import A (foo))
2020-10-26 15:35:57 <thblt> but archived
2020-10-26 15:36:13 × Ivan__1 quits (~yudin@193.137.102.254) (Ping timeout: 265 seconds)
2020-10-26 15:36:49 <tomsmeding> > In the future, we plan for Importify to be able to: [...] Convert imports between implicit and explicit, [...]
2020-10-26 15:36:50 <lambdabot> <hint>:1:14: error: <hint>:1:14: error: parse error on input ‘,’
2020-10-26 15:36:51 <tomsmeding> it doesn't ;)
2020-10-26 15:36:57 <thblt> Ha yes, read too fast.
2020-10-26 15:37:19 → __monty__ joins (~toonn@unaffiliated/toonn)
2020-10-26 15:37:33 → fendor_ joins (~fendor@91.141.2.240.wireless.dyn.drei.com)
2020-10-26 15:38:23 <thblt> OK, I have another stupid question. I wrote a simple number guessing game using monad transformers, and it should work I think, but… I have no idea how to *start* the game. Here's the code https://paste.thb.lt/1603726423.lhs.html
2020-10-26 15:38:26 × bartemius quits (~bartemius@109-252-20-20.nat.spd-mgts.ru) (Remote host closed the connection)
2020-10-26 15:38:34 <thblt> The question is: what's the type of the thing that takes a GConf, inits a GState and actually instantiates a Game?
2020-10-26 15:39:19 × fendor quits (~fendor@078132040188.public.t-mobile.at) (Ping timeout: 246 seconds)
2020-10-26 15:39:53 <thblt> Game is type Game = ReaderT GConf (StateT GState IO)
2020-10-26 15:40:05 × elliott_ quits (~elliott_@pool-108-51-141-12.washdc.fios.verizon.net) (Read error: Connection reset by peer)
2020-10-26 15:40:52 <tomsmeding> what about https://hackage.haskell.org/package/mtl-2.2.2/docs/Control-Monad-Reader.html#v:runReader and https://hackage.haskell.org/package/mtl-2.2.2/docs/Control-Monad-State-Lazy.html#v:runState ?
2020-10-26 15:41:04 <thblt> tomsmeding: reading this!
2020-10-26 15:41:13 <tomsmeding> oh runReaderT and runStateT of course, because transformers
2020-10-26 15:41:14 → Tuplanolla joins (~Tuplanoll@91-159-68-239.elisa-laajakaista.fi)
2020-10-26 15:42:08 × firstlove quits (~firstlove@58.246.122.242) (Ping timeout: 256 seconds)
2020-10-26 15:43:16 → Ivan__1 joins (~yudin@gw.mat.uc.pt)
2020-10-26 15:46:16 → elliott_ joins (~elliott_@pool-108-51-141-12.washdc.fios.verizon.net)
2020-10-26 15:46:23 × alp quits (~alp@2a01:e0a:58b:4920:f036:1a06:bf83:85b) (Ping timeout: 272 seconds)
2020-10-26 15:47:06 → Deide joins (~Deide@217.155.19.23)
2020-10-26 15:48:47 <thblt> I feel stupid, but in ghci, `runReaderT play def` just prints `runReaderT play def :: StateT GState IO ()`
2020-10-26 15:49:40 <monsterchrom> add runStateT on top of that
2020-10-26 15:49:40 <__monty__> You'll probably want to runStateT on that?
2020-10-26 15:50:04 → alp joins (~alp@2a01:e0a:58b:4920:c83a:4ff:4d6e:6312)
2020-10-26 15:50:07 <monsterchrom> Yes this is what happens when you go crazy with monad transformers.
2020-10-26 15:50:24 <monsterchrom> piled higher and deeper
2020-10-26 15:50:25 <[exa]> thblt: runReaderT translates ReaderT r x a to `x a`, in your case this continues because x=StateT...
2020-10-26 15:50:54 <thblt> monsterchrom: is that so crazy? a config ReaderT, a runtime state and IO?
2020-10-26 15:51:10 <thblt> (Thanks all)
2020-10-26 15:51:21 × geowiesnot quits (~user@i15-les02-ix2-87-89-181-157.sfr.lns.abo.bbox.fr) (Ping timeout: 256 seconds)
2020-10-26 15:51:35 <monsterchrom> You look at "runStateT (runReaderT play def) initial", you tell me.
2020-10-26 15:51:43 × heatsink quits (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net) (Remote host closed the connection)
2020-10-26 15:51:54 <thblt> monsterchrom: i honestly don't know.
2020-10-26 15:51:59 <monsterchrom> for starters, the sheer obligation to say that in the correct order
2020-10-26 15:52:34 → idhugo_ joins (~idhugo@563472ae.rev.stofanet.dk)
2020-10-26 15:52:37 <T0pH4t> is there a way to put a type constraint on a type family decl eg "class Foo a where type Bar a" where i limit the type set for "Bar a" to be an instance of some class Baz
2020-10-26 15:52:44 × elliott__ quits (~elliott@pool-108-51-141-12.washdc.fios.verizon.net) (Ping timeout: 258 seconds)
2020-10-26 15:53:30 → unihernandez22 joins (~Unai@168.197.200.20)
2020-10-26 15:54:20 × unihernandez22 quits (~Unai@168.197.200.20) (Client Quit)
2020-10-26 15:55:10 → Guest23433 joins (~Waithamai@185.244.214.217)
2020-10-26 15:55:48 <thblt> monsterchrom: I think I see what you mean. I thought it was a common pattern.
2020-10-26 15:55:50 <__monty__> monsterchrom: Would you recommend MTL instead (or another effect system)? Or rather just define a monad per-project?
2020-10-26 15:56:01 <tomsmeding> T0pH4t: I don't think so, but maybe you can make this work: "class Baz a => Foo a where type Bar a"
2020-10-26 15:56:28 <dolio> thblt: I don't see a problem.
2020-10-26 15:56:41 <monsterchrom> Define a monad per project. And, ironically, its definition is a newtype wrapper around using MTL stuff.
2020-10-26 15:56:50 × alx741 quits (~alx741@186.178.110.169) (Ping timeout: 260 seconds)
2020-10-26 15:57:00 <monsterchrom> err I guess transformers stuff.
2020-10-26 15:57:09 <T0pH4t> ]tomsmeding: that woudl only constrain 'a' and not the type of "Bar a"
2020-10-26 15:57:26 <T0pH4t> tomsmeding: ^
2020-10-26 15:57:26 <tomsmeding> monsterchrom: wanting the newtype does not mean that you can't have more than 1 monad stack per project
2020-10-26 15:58:03 <tomsmeding> and you just move the problem: you still have to say runStateT (runReaderT blabla), but now just in runMyMonad instead of in main
2020-10-26 15:58:09 <monsterchrom> So define two newtype wrappers of two stacks?
2020-10-26 15:58:13 <tomsmeding> yes!
2020-10-26 15:58:27 <thblt> GHC has been screaming at me a few times talking about "functional dependencies". What can I read to understand what it's screaming about?
2020-10-26 15:58:35 → heatsink joins (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net)
2020-10-26 15:58:40 <tomsmeding> T0pH4t: ah that's true; then I don't think that's possible
2020-10-26 15:58:44 → fendor__ joins (~fendor@078132040188.public.t-mobile.at)
2020-10-26 15:58:45 × fendor_ quits (~fendor@91.141.2.240.wireless.dyn.drei.com) (Ping timeout: 240 seconds)
2020-10-26 15:58:54 <monsterchrom> OK, inside the newtype wrapper it doesn't have to be transformers stuff. You can handcode it yourself. That's a choice.
2020-10-26 15:59:38 <texasmynsted> The ~/.cabal/config has only seven options listed in the init section. How do people normally make a template for cabal init? I have done this with "template" programs, but it seems like there must be a better way.
2020-10-26 15:59:43 <[exa]> thblt: functional dependencies arise to disambiguate various stuff with multiparameter type classes, I guess you just hit some minor problematic spot
2020-10-26 15:59:56 <monsterchrom> I have done "newtype X a = MkX (String -> Maybe (String, a))" by hand.
2020-10-26 16:00:24 × GyroW_ quits (~GyroW@unaffiliated/gyrow) (Quit: Someone ate my pie)
2020-10-26 16:00:34 → GyroW joins (~GyroW@d54C03E98.access.telenet.be)
2020-10-26 16:00:34 × GyroW quits (~GyroW@d54C03E98.access.telenet.be) (Changing host)
2020-10-26 16:00:34 → GyroW joins (~GyroW@unaffiliated/gyrow)
2020-10-26 16:00:34 → Rudd0 joins (~Rudd0@185.189.115.103)
2020-10-26 16:01:05 <monsterchrom> If you're too lazy to write its >>=, you could choose "newtype X a = MkX (StateT String Maybe a)" and write a few "deriving" clauses.
2020-10-26 16:01:36 <tomsmeding> I'm not sure whether coding something like that by hand would be easier than saying runStateT . runMaybeT for some definition of (.)
2020-10-26 16:02:05 <tomsmeding> not saying it's particularly difficult, but I guess I don't really see your argument against transformer stacks anymore then :p
2020-10-26 16:02:06 <dolio> Yeah, I don't get what's being optimized here.
2020-10-26 16:03:12 <monsterchrom> Bah, that was a tangent.
2020-10-26 16:03:13 × chkno quits (~chkno@75-7-2-127.lightspeed.sntcca.sbcglobal.net) (Read error: Connection reset by peer)
2020-10-26 16:03:31 <monsterchrom> For this particular game, one doesn't need the StateT GState stage.
2020-10-26 16:03:55 → chkno joins (~chkno@75-7-2-127.lightspeed.sntcca.sbcglobal.net)
2020-10-26 16:04:02 <tomsmeding> ... or, apparently, I misunderstood you. Sorry for the confusion :p
2020-10-26 16:04:04 <thblt> monsterchrom: of course! I wanted to learn about transformers, which was easier with a simple problem.
2020-10-26 16:04:12 <T0pH4t> so this doesn't seem to work either :/ "class Baz a => Foo a where type Bar a, toBaz :: Baz (Bar a) => a -> Bar a"
2020-10-26 16:04:31 <monsterchrom> And the ReaderT stage could be replaced by the equivalent but lighter (GConf ->)
2020-10-26 16:04:58 <tomsmeding> T0pH4t: remove the 'Baz a =>' constraint on Foo?
2020-10-26 16:05:11 → johnw joins (~johnw@haskell/developer/johnw)
2020-10-26 16:05:15 <tomsmeding> oh oh oh maaaybe you can "class Baz (Bar a) => Foo a where type Bar a"
2020-10-26 16:05:16 <thblt> monsterchrom: thanks for the tip!
2020-10-26 16:05:16 <T0pH4t> woops, thats not there, wrong copy
←Prev  Next→
Page 1 .. 872 873 874 875 876 877 878 879 880 881 882 .. 5022

All times are in UTC.