Home liberachat/#haskell: Logs Calendar

Logs: liberachat/#haskell

←Prev  Next→
Page 1 .. 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 .. 18040
— 1,803,949 events total
2021-08-02 06:18:51 <euouae> you can throw away the result of an IO action right?
2021-08-02 06:18:51 <yahb> dminuoso: String -> IO ()
2021-08-02 06:18:57 <dminuoso> Yes.
2021-08-02 06:19:23 <euouae> So there's always a projection `M a -> M t`
2021-08-02 06:19:30 <euouae> Sorry, `M a -> M ()`
2021-08-02 06:19:36 <dminuoso> Correct
2021-08-02 06:19:40 <Maxdamantus> You can't have an `IO` action that emits no value, so the value `()` is used as a placeholder.
2021-08-02 06:19:45 <dminuoso> % :t () <$ putStrLn "foo" -- euouae
2021-08-02 06:19:45 <yahb> dminuoso: IO ()
2021-08-02 06:19:55 <dminuoso> This combinator is also, very confusingly, named `void` in base
2021-08-02 06:19:58 <dminuoso> % :t void
2021-08-02 06:19:58 <yahb> dminuoso: Functor f => f a -> f ()
2021-08-02 06:20:02 <euouae> Which means that function taking `M ()` may as well take `M a`, it's just a hint that they do nothing with `a`
2021-08-02 06:20:08 <dminuoso> euouae: Precisely.
2021-08-02 06:20:14 <dminuoso> You're spot on.
2021-08-02 06:20:21 <euouae> However, they can't take `M a`, right?
2021-08-02 06:20:30 <dminuoso> Also correct.
2021-08-02 06:20:36 <dminuoso> Which is why `void` exists.
2021-08-02 06:20:37 <euouae> Alright, thank you for clarifying everything :)
2021-08-02 06:20:57 <dminuoso> It's for massaging values like IO actions into the right type, for when some other function demands say `IO ()`
2021-08-02 06:21:10 <dminuoso> But you can also just write `() <$ foo` if you like
2021-08-02 06:21:51 <euouae> But what does `() <$ putStrLn "Hello"` print
2021-08-02 06:22:02 <dminuoso> Oh, well to be fair that's a silly thing to do
2021-08-02 06:22:10 <dminuoso> % :T () <$ getLine
2021-08-02 06:22:10 <yahb> dminuoso: unknown command ':T'; use :? for help.
2021-08-02 06:22:12 <dminuoso> % :t () <$ getLine
2021-08-02 06:22:12 <yahb> dminuoso: IO ()
2021-08-02 06:22:14 <dminuoso> This is a etter example
2021-08-02 06:22:19 → Lord_of_Life_ joins (~Lord@user/lord-of-life/x-2819915)
2021-08-02 06:22:45 <euouae> I see
2021-08-02 06:22:47 <dminuoso> euouae: It sitll prints "Hello". The example is poor because putStrLn at the end already produces IO ().
2021-08-02 06:23:04 <dminuoso> But with getLine, you still get the effect of consuing a string on stdin, but it throws away the result
2021-08-02 06:23:20 <euouae> Right the type information is different than the side ffects
2021-08-02 06:23:31 <dminuoso> Right.
2021-08-02 06:23:32 × Lord_of_Life quits (~Lord@user/lord-of-life/x-2819915) (Ping timeout: 256 seconds)
2021-08-02 06:23:33 <euouae> in some sense at least. of course IO is always there
2021-08-02 06:23:34 Lord_of_Life_ is now known as Lord_of_Life
2021-08-02 06:25:43 <dminuoso> euouae: Right, this is the beauty of this parametrized IO type. We separate the side-effect (IO) from the result of carrying out that side-effect.
2021-08-02 06:26:04 <dminuoso> Now, of course we cant know what exact side-effects actually happen inside IO. And this is where the idea of effect systems begins.
2021-08-02 06:26:13 <dibblego> IO is unrelated to side-effects
2021-08-02 06:27:06 × gambpang quits (~ian@207.181.230.156) (Ping timeout: 272 seconds)
2021-08-02 06:27:22 → gambpang joins (~ian@207.181.230.156)
2021-08-02 06:28:31 → Obo joins (~roberto@70.pool90-171-81.dynamic.orange.es)
2021-08-02 06:34:13 × guest86 quits (~guest86@2001:8004:1420:14ac:b1cc:c092:84b0:86f6) (Ping timeout: 246 seconds)
2021-08-02 06:36:15 <aegon> Axman6: if your still around i found it out. This is the wierdest syntax yet fromJust (fromList x) is the way to get it from a list that has been built up by a function
2021-08-02 06:36:33 <aegon> is the fromJust part of IsList stuff or is that something specific that the Hasktorch implementation did :?
2021-08-02 06:38:10 <Axman6> :t fromJust
2021-08-02 06:38:11 <lambdabot> Maybe a -> a
2021-08-02 06:38:38 <euouae> dminuoso: Is this related? http://okmij.org/ftp/Haskell/extensible/exteff.pdf
2021-08-02 06:38:41 <Axman6> I guess there's a IsList instnce for Maybe? that code looks quite confusing though
2021-08-02 06:39:41 <dminuoso> euouae: Yes, this is one example.
2021-08-02 06:42:05 <euouae> dminuoso: Ok I don't see it myself, but I trust that it's being researched :P I'll probably be learning the basics of cabal and the rest of the tools for now :)
2021-08-02 06:42:12 × phma_ quits (phma@2001:5b0:211b:e2d8:c072:cd43:475:d917) (Read error: Connection reset by peer)
2021-08-02 06:42:54 → gehmehgeh joins (~user@user/gehmehgeh)
2021-08-02 06:43:08 <dminuoso> euouae: Yeah. Much in the effect systems is ongoing research and experimentation, and in Haskell in particular it's mostly limited to custom effects (in the sense of how do we compose various monadic effects together in an extensible and declarative way).
2021-08-02 06:43:08 → phma_ joins (phma@2001:5b0:210d:4948:1e1a:34b5:d23e:1981)
2021-08-02 06:47:47 → merijn joins (~merijn@83-160-49-249.ip.xs4all.nl)
2021-08-02 06:55:08 → _ht joins (~quassel@82-169-194-8.biz.kpn.net)
2021-08-02 06:57:07 → acidjnk_new joins (~acidjnk@p200300d0c72b9504b8b6f59e7f78d4aa.dip0.t-ipconnect.de)
2021-08-02 07:01:27 <sergio812> I’m having trouble with QuickCheck’s “collect” when testing IO code...
2021-08-02 07:01:37 <sergio812> I can easily modify a basic property like “prop_pure list = myImpl list === refImpl list” to add case distribution like so “prop_pure list = collect (length list) $ myImpl list === refImpl list”, and all works fine.
2021-08-02 07:02:03 <sergio812> Now, how should I do to add case distribution to IO code like “prop_io list = monadicIO $ do { res <- run $ myImplIO list; assert $ res == refImpl list }”?
2021-08-02 07:02:19 <sergio812> I tried tucking in “collect (length l) $” before “run” or “assert” and GHC complained about expecting type “PropertyM IO Int” or “PropertyM IO a0”, while actual type is “Property”.
2021-08-02 07:02:20 → lortabac joins (~lortabac@2a01:e0a:541:b8f0:4a3f:fe34:92c5:875c)
2021-08-02 07:02:33 <sergio812> I tried tucking in “collect (length l) $” after “run” or “assert” and GHC complained about expecting type “IO Int” or “Bool”, while actual type is “Property”.
2021-08-02 07:02:42 <sergio812> Can I use “collect” for IO code? If so, how? If not, how can I add case distribution?
2021-08-02 07:04:56 → dhouthoo joins (~dhouthoo@178-117-36-167.access.telenet.be)
2021-08-02 07:10:29 → fendor joins (~fendor@77.119.222.253.wireless.dyn.drei.com)
2021-08-02 07:11:12 <c_wraith> sergio812: Why are you trying to do it inside the do block?
2021-08-02 07:11:23 <c_wraith> sergio812: it would work just fine if you did it outside
2021-08-02 07:12:15 <nshepperd> sergio812: i think you're meant to use 'monitor' in io tests
2021-08-02 07:12:38 <nshepperd> sergio812: https://hackage.haskell.org/package/QuickCheck-2.14.2/docs/Test-QuickCheck-Monadic.html#v:monitor
2021-08-02 07:13:40 <sergio812> c_wraith: you mean, tuck in "collect (length list) $" before "monadicIO"?
2021-08-02 07:14:09 <c_wraith> oh, if it is a scoping error, then yes. monitor is what you want
2021-08-02 07:14:36 × takuan quits (~takuan@178-116-218-225.access.telenet.be) (Ping timeout: 272 seconds)
2021-08-02 07:14:52 <sergio812> c_wraith: you're probably right that I can make do with putting it outside on this example
2021-08-02 07:16:40 × tzh quits (~tzh@c-24-21-73-154.hsd1.or.comcast.net) (Quit: zzz)
2021-08-02 07:16:41 <sergio812> nshepperd: thank you for pointing me to "monitor"!
2021-08-02 07:17:13 × azeem quits (~azeem@dynamic-adsl-94-34-48-122.clienti.tiscali.it) (Remote host closed the connection)
2021-08-02 07:17:26 → azeem joins (~azeem@dynamic-adsl-94-34-48-122.clienti.tiscali.it)
2021-08-02 07:20:10 × geekosaur quits (~geekosaur@xmonad/geekosaur) (Ping timeout: 240 seconds)
2021-08-02 07:21:52 × merijn quits (~merijn@83-160-49-249.ip.xs4all.nl) (Ping timeout: 258 seconds)
2021-08-02 07:21:52 × jonathanx quits (~jonathan@h-178-174-176-109.A357.priv.bahnhof.se) (Read error: Connection reset by peer)
2021-08-02 07:22:03 → jonathanx joins (~jonathan@h-178-174-176-109.A357.priv.bahnhof.se)
2021-08-02 07:22:58 → mastarija joins (~mastarija@31.217.8.174)
2021-08-02 07:24:06 × shriekingnoise quits (~shrieking@186.137.144.80) (Quit: Quit)
2021-08-02 07:24:06 × favonia quits (~favonia@user/favonia) (Ping timeout: 272 seconds)
2021-08-02 07:24:21 <mastarija> Do we have something like the "Alternative" type class, but that is related to "Semigroup" instead of "Monoid". I'd really like an instance of the Alternative, but without that annoying "empty"
2021-08-02 07:24:48 <dminuoso> mastarija: Check out semigroupoids
2021-08-02 07:24:51 → cfricke joins (~cfricke@user/cfricke)
2021-08-02 07:25:05 <mastarija> dminuoso, nothing in base, huh?
2021-08-02 07:25:08 <dminuoso> Right
2021-08-02 07:25:17 <dminuoso> https://hackage.haskell.org/package/semigroupoids-5.3.5/docs/Data-Functor-Alt.html
2021-08-02 07:25:29 <dminuoso> mastarija: ^- this is one of the reasons quite a few packages depend on semigroupoids. :)
2021-08-02 07:25:36 <mastarija> :D
2021-08-02 07:25:48 <mastarija> I guess mine will too
2021-08-02 07:29:55 → jay-invariant joins (~jay@c-24-4-6-169.hsd1.ca.comcast.net)
2021-08-02 07:30:06 <dibblego> Either is a good example instance
2021-08-02 07:31:03 <mastarija> Would it be fine, if I write in my library, you are free to use the "empty", but if you do, you might loose the data you've been accumulating?
2021-08-02 07:31:22 <dibblego> as long as it is lawful, sure
←Prev  Next→
Page 1 .. 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 .. 18040

All times are in UTC.