Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 118 119 120 121 122 123 124 125 126 127 128 .. 5022
— 502,152 events total
2020-09-21 15:33:11 <siraben> phadej: Bracket pattern is like RAII in C++ and Rust?
2020-09-21 15:33:14 <merijn> siraben: I like the idea of that, but I'm not sure about the reality :p
2020-09-21 15:33:19 <phadej> (in fact explicitly unsolved, IIRC)
2020-09-21 15:33:29 <merijn> siraben: IndexedMonads might be a better solution for a "better bracket", tbh
2020-09-21 15:33:38 <siraben> phadej: How do they interact? I know that Rust does not have exceptions but I wonder if it's do to the theoretical limitations.
2020-09-21 15:33:46 <merijn> But I don't see Haskell en masse adopting indexed monads :p
2020-09-21 15:33:46 <siraben> IndexedMonads is a language extension?
2020-09-21 15:33:56 <merijn> siraben: that's a missing space ;)
2020-09-21 15:34:06 <phadej> I think someone have to do a bit of homework at this point ;)
2020-09-21 15:34:11 → eric___ joins (~eric@2804:431:c7d4:a280:a420:f119:9a1a:c48)
2020-09-21 15:34:41 × eric_ quits (~eric@2804:431:c7d4:402a:842:24e2:eb65:718) (Ping timeout: 246 seconds)
2020-09-21 15:34:58 <phadej> I don't know what mechanism Rust uses to report division by zero errors
2020-09-21 15:34:59 <phadej> e.g.
2020-09-21 15:35:05 <siraben> Oh, it's like the LockM monad in https://www.cs.tufts.edu/comp/150FP/archive/simon-peyton-jones/typefun.pdf
2020-09-21 15:35:06 <siraben> indexed monads, that is
2020-09-21 15:35:07 <phadej> so I cannot comment further
2020-09-21 15:35:31 <siraben> phadej: it panics
2020-09-21 15:35:34 <siraben> so, stack unwind
2020-09-21 15:35:39 <merijn> siraben: indexed monads would be "(>>>=) :: IxMonad m => m i j a -> (a -> m j k b) -> m i k b"
2020-09-21 15:35:49 → coffeeturtle joins (~coffeetur@cpc103420-donc13-2-0-cust24.17-1.cable.virginm.net)
2020-09-21 15:35:55 → mastarija joins (~mastarija@93-142-74-181.adsl.net.t-com.hr)
2020-09-21 15:36:00 × mastarija quits (~mastarija@93-142-74-181.adsl.net.t-com.hr) (Client Quit)
2020-09-21 15:36:02 <merijn> siraben: And then you can use i/j/k etc. to encode "file descript is open" file descriptor is closed, etc.
2020-09-21 15:36:07 <phadej> siraben: can you recover from panics?
2020-09-21 15:36:17 <siraben> phadej: no
2020-09-21 15:36:21 <siraben> merijn: right
2020-09-21 15:36:29 <siraben> panics are like error in Haskell, I suppose
2020-09-21 15:36:36 <phadej> you can catch `error` in Haskell
2020-09-21 15:36:55 <phadej> or just let the thread die (but not the whole program)
2020-09-21 15:37:13 × wroathe quits (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net) (Ping timeout: 246 seconds)
2020-09-21 15:37:37 × acidjnk_new2 quits (~acidjnk@p200300d0c73658819d57354728142a5b.dip0.t-ipconnect.de) (Ping timeout: 260 seconds)
2020-09-21 15:37:41 × coffeeturtle quits (~coffeetur@cpc103420-donc13-2-0-cust24.17-1.cable.virginm.net) (Remote host closed the connection)
2020-09-21 15:37:55 <phadej> :t error
2020-09-21 15:37:57 <lambdabot> [Char] -> a
2020-09-21 15:37:58 <siraben> heh, well technically you can catch panics in Rust, but it's definitely not recommended https://doc.rust-lang.org/beta/std/panic/fn.catch_unwind.html
2020-09-21 15:38:02 → coffeeturtle joins (~coffeetur@cpc103420-donc13-2-0-cust24.17-1.cable.virginm.net)
2020-09-21 15:38:09 <phadej> :t throwIO . userError
2020-09-21 15:38:11 <lambdabot> String -> IO a
2020-09-21 15:38:18 <phadej> :t throw . userError
2020-09-21 15:38:20 <lambdabot> String -> c
2020-09-21 15:38:27 × suppi quits (~suppi@2605:f700:40:c00::e6fc:6842) (Ping timeout: 260 seconds)
2020-09-21 15:38:30 <phadej> error = throw . userError
2020-09-21 15:38:37 <phadej> you can try to catch them in IO then
2020-09-21 15:39:10 <phadej> which is e.g. useful when you write a web-server framework, so you can log how users' buggy handlers failed
2020-09-21 15:39:31 <siraben> https://users.cs.northwestern.edu/~jesse/pubs/substructural-control/CtlURAL.pdf comes up when I search for linear types and exceptions
2020-09-21 15:39:40 <siraben> "In short, exceptions and linear types refuse to get along, because linear types make promises that exceptions do not let them keep."
2020-09-21 15:39:49 <siraben> But in an affine system apparently this is alright, just drop the variables.
2020-09-21 15:40:20 → heatsink joins (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net)
2020-09-21 15:40:28 <phadej> yes.
2020-09-21 15:40:44 <phadej> this is why I said that LinearHaskell + exceptions is unsolved problem
2020-09-21 15:41:25 <phadej> one can be optimistic and think "we need an implementation to see how things fit (or not fit) together, so we can develop LinearHaskell further"
2020-09-21 15:41:39 <phadej> or you can be pessimistic and just declare upfront "it won't work"
2020-09-21 15:41:40 <phadej> :)
2020-09-21 15:41:50 × chele quits (~chele@ip5b416ea2.dynamic.kabel-deutschland.de) (Ping timeout: 256 seconds)
2020-09-21 15:41:55 → pingiun joins (~pingiun@ip-213-124-184-182.ip.prioritytelecom.net)
2020-09-21 15:42:22 <siraben> Hehe
2020-09-21 15:42:25 <phadej> IMHO anyone who says *now* that LinearHaskell will solve any problem is just selling you a hype.
2020-09-21 15:42:38 <phadej> yeah, it might solve a problem, but give you dozen new ones.
2020-09-21 15:42:40 <siraben> Why emphasis on now?
2020-09-21 15:42:44 <siraben> Ah, right.
2020-09-21 15:42:45 <Taneb> It might solve the problem of "I would like another thing to mess around with"
2020-09-21 15:43:11 <siraben> I'm not expecting it to solve problems immediately, but linear types are interesting and I'd like to mess around with them, heh.
2020-09-21 15:43:18 × kenran quits (~maier@b2b-37-24-119-190.unitymedia.biz) (Ping timeout: 256 seconds)
2020-09-21 15:43:18 → revprez_anzio joins (~revprez_a@pool-108-49-213-40.bstnma.fios.verizon.net)
2020-09-21 15:43:22 <phadej> Taneb: that can be fun :)
2020-09-21 15:43:24 <siraben> AffineHaskell, anyone?
2020-09-21 15:43:56 × bahamas quits (~lucian@unaffiliated/bahamas) (Ping timeout: 272 seconds)
2020-09-21 15:45:25 → vdiazroa joins (578036c2@p578036c2.dip0.t-ipconnect.de)
2020-09-21 15:45:28 → mmohammadi98129 joins (~mmohammad@2.178.188.114)
2020-09-21 15:45:38 × Chi1thangoo quits (~Chi1thang@87.112.60.168) (Remote host closed the connection)
2020-09-21 15:45:49 × mmohammadi9812 quits (~mmohammad@2.178.188.114) (Ping timeout: 265 seconds)
2020-09-21 15:45:49 mmohammadi98129 is now known as mmohammadi9812
2020-09-21 15:47:16 → cole-h joins (~cole-h@c-73-48-197-220.hsd1.ca.comcast.net)
2020-09-21 15:49:41 × nbloomf quits (~nbloomf@2600:1700:83e0:1f40:c87d:6f52:4c16:4130) (Quit: My MacBook has gone to sleep. ZZZzzz…)
2020-09-21 15:49:43 → wroathe joins (~wroathe@c-73-24-27-54.hsd1.mn.comcast.net)
2020-09-21 15:50:16 × cr3 quits (~cr3@192-222-143-195.qc.cable.ebox.net) (Ping timeout: 272 seconds)
2020-09-21 15:50:56 <phadej> GHC-9.2 will change the syntax for linear arrows, so at least that should make it clear that LinearHaskell is very experimental feature ;)
2020-09-21 15:51:54 <dolio> Presumably the emphasis is due to, right now, no one having any significant experience actually designing things with LinearHaskell, so any claims about what it will solve are just speculation.
2020-09-21 15:52:07 <phadej> yes
2020-09-21 15:52:18 → qwerty joins (59b76b94@a89-183-107-148.net-htp.de)
2020-09-21 15:52:25 → ph88_ joins (~ph88@2a02:8109:9e40:2704:a1b2:872a:487a:52ae)
2020-09-21 15:52:31 × eric___ quits (~eric@2804:431:c7d4:a280:a420:f119:9a1a:c48) (Read error: Connection reset by peer)
2020-09-21 15:52:35 <phadej> obviously authors claim it will solve all problems ;)
2020-09-21 15:52:43 qwerty is now known as qwerty123
2020-09-21 15:52:57 qwerty123 is now known as qwerty93
2020-09-21 15:53:03 <dolio> Even the authors mostly have toy examples from research papers and blog posts.
2020-09-21 15:53:09 → eric_ joins (~eric@2804:431:c7d4:a280:a420:f119:9a1a:c48)
2020-09-21 15:54:00 × coffeeturtle quits (~coffeetur@cpc103420-donc13-2-0-cust24.17-1.cable.virginm.net) (Quit: leaving)
2020-09-21 15:54:09 <phadej> the implicit claim? is that you examples generalise
2020-09-21 15:54:15 <phadej> s/you/toy/
2020-09-21 15:54:18 × utopic_int0x80 quits (~lucid_0x8@188.253.235.98) (Quit: Leaving)
2020-09-21 15:54:30 <phadej> but in this case it is not obvious
2020-09-21 15:55:23 → plutoniix joins (~q@ppp-27-55-67-24.revip3.asianet.co.th)
2020-09-21 15:56:21 → nikosky joins (ba0e194f@186.14.25.79)
2020-09-21 15:56:29 × ph88^ quits (~ph88@2a02:8109:9e40:2704:3dae:c0b9:eddf:4cf0) (Ping timeout: 272 seconds)
2020-09-21 15:56:56 <qwerty93> I want to write a function `mycatch :: a -> Either String a` which returns its argument if no error was thrown and otherwise the error message. For example, `mycatch ((error "asd")::Int)` should return `Left "asd"`. How can this be done with `unsafePerformIO`?
2020-09-21 15:57:28 <dminuoso> qwerty93: That is not possible in standard haskell/
2020-09-21 15:57:36 × cpressey quits (~cpressey@79-65-248-204.host.pobb.as13285.net) (Quit: WeeChat 1.9.1)
2020-09-21 15:57:36 <dminuoso> You have to do this in IO
2020-09-21 15:57:42 → nbloomf joins (~nbloomf@2600:1700:83e0:1f40:f82a:5e5e:a65a:42ad)
←Prev  Next→
Page 1 .. 118 119 120 121 122 123 124 125 126 127 128 .. 5022

All times are in UTC.