Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 783 784 785 786 787 788 789 790 791 792 793 .. 5022
— 502,152 events total
2020-10-21 19:33:34 <hyperisco> doesn't have to go through a shell
2020-10-21 19:37:35 <hololeap> % import Control.Monad.Logic
2020-10-21 19:37:36 <yahb> hololeap: ; <no location info>: error:; Could not find module `Control.Monad.Logic'; Perhaps you meant; Control.Monad.Co (from kan-extensions-5.2); Control.Monad.Cont (from mtl-2.2.2); Control.Monad.Fail (from base-4.14.0.0)
2020-10-21 19:38:03 <[exa]> no prolog today
2020-10-21 19:38:21 <monochrom> yahb probably doesn't have logict installed
2020-10-21 19:38:36 <Chousuke> monochrom: I guess I'm more a fan of the simple and stupid approach of just treating everything as bytes. for HTTP, the header is a type hint, but you still can't trust that the data is anything other than a stream of bytes.
2020-10-21 19:38:52 <monochrom> "become a galaxy-level patron so you can instruct yahb to install one more package!"
2020-10-21 19:39:05 <geekosaur> but you have to have some idea of what parser to use to validate it, at minimum
2020-10-21 19:39:17 <Chousuke> yeah
2020-10-21 19:39:29 <Chousuke> I'm not sure that that's useful on the filesystem level though
2020-10-21 19:39:46 <geekosaur> it's useful at the level of apps getting data from the filesystem
2020-10-21 19:39:50 <monochrom> You are right that a receiver still has to parse, and to expect parse error. But the receiver also expects to not bother trying other parsers.
2020-10-21 19:40:24 <monochrom> If you say "text/html" I'm not going to try gunzip it.
2020-10-21 19:40:47 <geekosaur> unless the compression header says to
2020-10-21 19:40:50 <monochrom> even if the bytes scream "I'm gzip"
2020-10-21 19:40:59 <monochrom> Yeah, that.
2020-10-21 19:41:38 <monochrom> If you say "text/html" I'm not going to run objdump on it, even if the bytes scream "ELF"
2020-10-21 19:41:50 <hololeap> i'm confused as to how you use ListT
2020-10-21 19:41:57 <hololeap> *LogicT
2020-10-21 19:42:17 <[exa]> hololeap: any short example / confusing snippet?
2020-10-21 19:42:18 <monochrom> For the most part you can use it like you use a list monad.
2020-10-21 19:42:25 <hololeap> i know it has some kind of similarities to ListT, but i don't see how
2020-10-21 19:42:49 <monochrom> There is just one addition command to learn if you want the search to be BFS instead of DFS.
2020-10-21 19:43:12 <monochrom> Perhaps start with Logic instead of LogicT?
2020-10-21 19:43:58 <hololeap> i've played around with Logic a bit, but simply constructing a `Logic a` value is weird
2020-10-21 19:44:35 × Franciman quits (~francesco@host-82-54-10-114.retail.telecomitalia.it) (Quit: Leaving)
2020-10-21 19:44:46 <[exa]> hololeap: how weird? (can't you construct `Logic a` using just `pure` ?
2020-10-21 19:44:58 hackage ukrainian-phonetics-basic 0.1.0.0 - A library to work with the basic Ukrainian phonetics and syllable segmentation. https://hackage.haskell.org/package/ukrainian-phonetics-basic-0.1.0.0 (OleksandrZhabenko)
2020-10-21 19:44:58 <Chousuke> monochrom: the other side of the coin is that if the metadata claims the file to be something and it looks like something else, it might get opened by an application you didn't expect which has a long history of causing problems. So I guess file extensions aren't completely awful for deciding the application to open, as long as they're visible to the user.
2020-10-21 19:45:45 <hololeap> odds = pure 1 `mplus` (odds >>= \i -> if i >= 100 then mzero else pure (i+2))
2020-10-21 19:46:01 <hololeap> is this roughly equivalent to [1,3..100]?
2020-10-21 19:46:07 <Chousuke> come to think of it, hashbangs are basically file metadata to decide the opening program. :P
2020-10-21 19:46:50 × knupfer quits (~Thunderbi@i59F7FF96.versanet.de) (Quit: knupfer)
2020-10-21 19:46:54 → knupfer1 joins (~Thunderbi@200116b82c30ba0001bfef634072ae05.dip.versatel-1u1.de)
2020-10-21 19:47:25 <[exa]> monochrom, Chousuke: btw sorry to jump into discussion, but the golden rule says that if the user/operator/etc is unable to instruct the programs to use correct data access methods, so he will fail at maintaining the metadata...unless the metadata becomes inseparable part of data, as it once had already, with file magic numbers
2020-10-21 19:47:27 <geekosaur> to decide the interpreter, to be specific (and the dynamic loader is treated as an interpreter)
2020-10-21 19:47:43 <monochrom> You can instantiate that to [] and get: odds = [1] ++ (odds >>= \i -> if i>=100 then [] else [i+2])
2020-10-21 19:49:18 knupfer1 is now known as knupfer
2020-10-21 19:49:43 <hololeap> ok, but i'm trying to construct something similar to [1,3..100], except i want it to be `Logic Int`
2020-10-21 19:49:54 <[exa]> (that said, I'd welcome if all file formats would have extensible magic, just like #!)
2020-10-21 19:50:00 <Chousuke> geekosaur: yeah. and if you're putting +x on any random file and running it, I don't think you can expect to stay secure. :P
2020-10-21 19:50:04 <hololeap> `observeAll odds` doesn't terminate
2020-10-21 19:50:08 <monochrom> Just add "odds :: Logic Int"
2020-10-21 19:51:18 <geekosaur> because if you try to observe all of it, it just keeps going through values after it hits the limit (empty list doesn't mean "stop", any more than it does for the list comprehension example)
2020-10-21 19:51:24 <monochrom> Probably the [] instantiation won't either.
2020-10-21 19:51:28 → chaosmasttter joins (~chaosmast@p200300c4a70942015d50de0935368812.dip0.t-ipconnect.de)
2020-10-21 19:51:40 <geekosaur> Chousuke, yet it's common for scripts to work that way
2020-10-21 19:51:47 <monochrom> > let odds = pure 1 `mplus` (odds >>= \i -> if i >= 100 then mzero else pure (i+2)) in odds :: [Int]
2020-10-21 19:51:50 <lambdabot> [1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31,33,35,37,39,41,43,45,47,49,51,53...
2020-10-21 19:51:54 <hololeap> your right, it doesn't
2020-10-21 19:51:58 <monochrom> actually need a smaller test
2020-10-21 19:52:02 <monochrom> > let odds = pure 1 `mplus` (odds >>= \i -> if i >= 10 then mzero else pure (i+2)) in odds :: [Int]
2020-10-21 19:52:05 <lambdabot> [1,3,5,7,9,11*Exception: <<loop>>
2020-10-21 19:52:23 <monochrom> Not a good program in the first place IMO
2020-10-21 19:52:30 <geekosaur> huh, I expected a timeout, not <<loop>>
2020-10-21 19:52:53 <monochrom> You don't do left recursion like that (odds = odds >>= ...)
2020-10-21 19:53:14 <geekosaur> oh, hm, I guess it does do that, yeh
2020-10-21 19:53:56 <hololeap> so, does a Logic computation by default not terminate, but requires observeMany to terminate it?
2020-10-21 19:54:14 <geekosaur> it's just the way this one was constructed
2020-10-21 19:54:16 <Chousuke> geekosaur: it is. I read most scripts before I run them, though of course I have to trust some things. Generally whatever is installed from OS package management is auto-trusted.
2020-10-21 19:54:34 <monochrom> I would rather code up a forward constraint propagation algorithm for learning logict.
2020-10-21 19:54:58 <Chousuke> geekosaur: I am very uncomfortable every time I have to use pip or gem or whatnot for whatever purpose.
2020-10-21 19:55:11 <hololeap> well, how would i write a `Logic Int` computation that would evaluate to [1,3,5,7,9] when used with observeAll?
2020-10-21 19:55:12 × shafox quits (~shafox@106.51.234.111) (Remote host closed the connection)
2020-10-21 19:55:45 × errst quits (~tirej@unaffiliated/tirej) (Quit: Lost terminal)
2020-10-21 19:56:36 <monochrom> odds i | i < 100 = pure i <|> odds (i+2)
2020-10-21 19:56:50 <monochrom> | otherwise = empty
2020-10-21 19:56:54 → Athas joins (athas@2a01:7c8:aaac:1cf:9a0:fad3:fdf2:2ed)
2020-10-21 19:57:03 × ech quits (~user@gateway/tor-sasl/ech) (Ping timeout: 240 seconds)
2020-10-21 19:57:11 <monochrom> change to mplus/mzero if Logic doesn't support Applicative
2020-10-21 19:58:26 <monochrom> May the odds be ever in your favour.
2020-10-21 19:58:30 <hololeap> hold on, i thought mzero/empty wouldn't stop the computation
2020-10-21 19:58:50 → Sheilong joins (uid293653@gateway/web/irccloud.com/x-zkmqxnjkmjnikczm)
2020-10-21 19:59:30 <hololeap> oh, i see. it never calls `odds` on the second branch. in my original example it calls it no matter what.
2020-10-21 19:59:54 → ZarTek joins (znc@static.138.236.76.144.clients.your-server.de)
2020-10-21 19:59:57 hackage ordinal 0.3.0.0 - Convert numbers to words in different languages. https://hackage.haskell.org/package/ordinal-0.3.0.0 (wvanonsem90)
2020-10-21 20:00:00 → acarrico joins (~acarrico@dhcp-68-142-39-249.greenmountainaccess.net)
2020-10-21 20:00:17 × justsomeguy quits (~justsomeg@unaffiliated/--/x-3805311) (Ping timeout: 260 seconds)
2020-10-21 20:00:37 → rprije joins (~rprije@110-175-117-18.tpgi.com.au)
2020-10-21 20:01:23 <monochrom> Oddly enough, "odds = pure 1 <|> fmap (+ 2) odds" is much more well defined that adding "if i>=100 then empty"
2020-10-21 20:01:25 × supercoven_ quits (~Supercove@dsl-hkibng32-54fb54-166.dhcp.inet.fi) (Ping timeout: 240 seconds)
2020-10-21 20:02:08 <hololeap> ok, i see now. it looks like i was over-complicating things, although i still don't have an intuition for this. the type signature `forall r. (a -> r -> r) -> r -> r` is not intuitive to me
2020-10-21 20:02:09 × conal quits (~conal@ip-66-115-176-174.creativelink.net) (Quit: Computer has gone to sleep.)
2020-10-21 20:02:18 → tabaqui joins (~tabaqui@2604:a880:800:c1::21b:3001)
2020-10-21 20:02:19 <hololeap> monochrom: how is it more well defined?
2020-10-21 20:03:07 <tabaqui> Allright, I'm using simple recusion to implement my dsl
2020-10-21 20:03:15 <tabaqui> Now I have another issue
2020-10-21 20:03:24 → justsomeguy joins (~justsomeg@unaffiliated/--/x-3805311)
2020-10-21 20:03:46 <tabaqui> type family VarType (a :: Symbol) :: *
2020-10-21 20:03:50 <monochrom> The different between "an infinite 'loop' that actually is productive in giving you an infinite list" and "an infinite loop that doesn't accomplish anything"
2020-10-21 20:04:11 <tabaqui> f :: Sing a -> VarType a
2020-10-21 20:04:29 <tabaqui> How can I write any function with such signature?
2020-10-21 20:04:48 <hololeap> monochrom: oh, i see what you mean
2020-10-21 20:05:04 → ech joins (~user@gateway/tor-sasl/ech)
2020-10-21 20:05:07 → z0 joins (~z0@bl15-167-204.dsl.telepac.pt)
2020-10-21 20:05:13 <monochrom> If you're going to do an infinite "x0 ++ x1 ++ x2 ++ x3 ++ ...", you would rather each xi be a non-empty list, so that the whole thing is an infinite list, rather than...
2020-10-21 20:05:16 <tabaqui> I've tried to compare singletons with %==, but Haskell doesn't accept it as a proof
2020-10-21 20:05:30 z0 is now known as Guest46583
2020-10-21 20:05:53 <monochrom> rather than from x50 onwards it's all xi=[] so the list should end but the computer doesn't know and chases gooses.
2020-10-21 20:06:10 → Franciman joins (~francesco@host-82-54-10-114.retail.telecomitalia.it)
←Prev  Next→
Page 1 .. 783 784 785 786 787 788 789 790 791 792 793 .. 5022

All times are in UTC.