Logs: liberachat/#haskell
| 2021-05-24 16:23:21 | → | juhp joins (~juhp@128.106.188.199) |
| 2021-05-24 16:23:34 | × | fabfianda quits (~fabfianda@net-93-148-125-174.cust.vodafonedsl.it) (Ping timeout: 264 seconds) |
| 2021-05-24 16:23:41 | <maerwald> | discord feels like a place for esport kiddies |
| 2021-05-24 16:24:36 | → | fabfianda joins (~fabfianda@mob-5-90-252-55.net.vodafone.it) |
| 2021-05-24 16:25:35 | <kuribas> | how is haskell superior when rust has a different use case? |
| 2021-05-24 16:27:04 | → | favonia joins (~favonia@99-36-111-163.lightspeed.dllstx.sbcglobal.net) |
| 2021-05-24 16:29:22 | × | kuribas quits (~user@ptr-25vy0i9i2khzv5pzmly.18120a2.ip6.access.telenet.be) (Quit: ERC (IRC client for Emacs 26.3)) |
| 2021-05-24 16:29:46 | × | amk quits (~amk@176.61.106.150) (Read error: Connection reset by peer) |
| 2021-05-24 16:30:13 | → | amk joins (~amk@176.61.106.150) |
| 2021-05-24 16:31:44 | × | xkapastel quits (uid17782@id-17782.tinside.irccloud.com) (Quit: Connection closed for inactivity) |
| 2021-05-24 16:32:12 | × | favonia quits (~favonia@99-36-111-163.lightspeed.dllstx.sbcglobal.net) (Changing host) |
| 2021-05-24 16:32:12 | → | favonia joins (~favonia@user/favonia) |
| 2021-05-24 16:33:17 | × | kluk quits (~kluk@cpe-69-203-82-73.nyc.res.rr.com) (Ping timeout: 272 seconds) |
| 2021-05-24 16:37:13 | × | ddellacosta quits (~ddellacos@86.106.121.82) (Remote host closed the connection) |
| 2021-05-24 16:37:40 | → | ddellacosta joins (~ddellacos@86.106.121.82) |
| 2021-05-24 16:38:03 | × | sondre quits (~sondrelun@cm-84.212.100.140.getinternet.no) (Ping timeout: 264 seconds) |
| 2021-05-24 16:39:29 | → | ddellaco_ joins (~ddellacos@86.106.121.50) |
| 2021-05-24 16:39:49 | <zzz> | haskell is so superior that it doesnt have a use case |
| 2021-05-24 16:40:32 | <zzz> | (mozilla moved to matrix btw. discord maybe a bridge?) |
| 2021-05-24 16:41:14 | <maerwald> | I think rust is probably superior on the paper in most regards, but it's WAY more verbose |
| 2021-05-24 16:42:10 | × | ddellacosta quits (~ddellacos@86.106.121.82) (Ping timeout: 264 seconds) |
| 2021-05-24 16:42:15 | × | oxide quits (~lambda@user/oxide) (Ping timeout: 264 seconds) |
| 2021-05-24 16:42:19 | <Vq> | I like to call rust the ugly ML dialect. |
| 2021-05-24 16:42:44 | <zzz> | depending on what you mean by "on the paper". im inclined to bet on haskell if we're talking "on paper" |
| 2021-05-24 16:42:52 | <Vq> | I quite like rust, but it doesn't look as pretty as other ML languages. |
| 2021-05-24 16:42:53 | <maerwald> | platform support, cross compiling, verified TLS implementation, less ecosystem fragmentation, predictable performance, low memory footprint, good debuggers, ... |
| 2021-05-24 16:42:53 | → | oxide joins (~lambda@user/oxide) |
| 2021-05-24 16:43:36 | × | ddellaco_ quits (~ddellacos@86.106.121.50) (Remote host closed the connection) |
| 2021-05-24 16:43:39 | × | merijn quits (~merijn@83-160-49-249.ip.xs4all.nl) (Ping timeout: 265 seconds) |
| 2021-05-24 16:43:58 | <maerwald> | I would do it for work, but it's more typing :p |
| 2021-05-24 16:44:02 | → | ddellacosta joins (~ddellacos@86.106.121.50) |
| 2021-05-24 16:44:14 | <zzz> | who likes predictable performance? i want to feel the rush of being alive |
| 2021-05-24 16:44:35 | × | ddellacosta quits (~ddellacos@86.106.121.50) (Remote host closed the connection) |
| 2021-05-24 16:44:41 | → | ddellacosta joins (~ddellacos@86.106.121.50) |
| 2021-05-24 16:45:36 | <maerwald> | zzz: after having worked on a haskell platform that took 64GB of memory, multiple failed consulting attempts and a couple of years to fix memory issues... I do care about it |
| 2021-05-24 16:46:02 | <maerwald> | *some* of it was due to architectural issues, other things was haskell being haskell |
| 2021-05-24 16:46:17 | <zzz> | bah... premature optimization |
| 2021-05-24 16:46:30 | → | safinaskar joins (~user@109-252-90-89.nat.spd-mgts.ru) |
| 2021-05-24 16:46:58 | <Cale> | zzz: What, are you going to go up to 128GB first? :) |
| 2021-05-24 16:47:07 | ← | favonia parts (~favonia@user/favonia) (Leaving) |
| 2021-05-24 16:47:20 | → | kluk joins (~kluk@2603-7000-9b3f-6934-6417-29db-6208-fab1.res6.spectrum.com) |
| 2021-05-24 16:47:26 | <maerwald> | sure, pretty easy if you find someone like Cale, who probably knows GHC internals and can fix it in one weekend ;) |
| 2021-05-24 16:48:10 | <zzz> | whatever it takes unitl the compiler starts showing signs of sentience. then it can take care of optimization for you |
| 2021-05-24 16:48:12 | <maerwald> | having to know GHC internals IS the problem |
| 2021-05-24 16:48:46 | → | dwt_ joins (~dwt_@2601:2c6:8080:618:48b9:7938:1ce:33ec) |
| 2021-05-24 16:48:55 | <maerwald> | it's already hard enough to keep up with the language itself, let alone esoteric performance things |
| 2021-05-24 16:48:55 | × | kluk quits (~kluk@2603-7000-9b3f-6934-6417-29db-6208-fab1.res6.spectrum.com) (Read error: Connection reset by peer) |
| 2021-05-24 16:49:21 | <maerwald> | but... tbf... most programs will never get to the point that this becomes a problem |
| 2021-05-24 16:49:25 | <maerwald> | but some do |
| 2021-05-24 16:49:26 | <Cale> | I really wish we could cool it on the questionable language extensions |
| 2021-05-24 16:50:03 | × | bgamari quits (~int0x0c@72.65.101.148) (Ping timeout: 264 seconds) |
| 2021-05-24 16:50:13 | → | kluk joins (~kluk@2603-7000-9b3f-6934-6417-29db-6208-fab1.res6.spectrum.com) |
| 2021-05-24 16:50:32 | × | horex539 quits (~horex539@2a02:a03f:6aa5:a00:c0c2:de35:17c6:545e) (Remote host closed the connection) |
| 2021-05-24 16:51:01 | <Cale> | I've been working on a change to let GHC do unpacking of sum-typed fields, and running into some frustrating LinearHaskell nonsense. |
| 2021-05-24 16:52:19 | <maerwald> | that would disrupt a lot of PhD thesis, I guess :P |
| 2021-05-24 16:53:33 | → | jao joins (~jao@cpc103048-sgyl39-2-0-cust502.18-2.cable.virginm.net) |
| 2021-05-24 16:53:35 | → | horex539 joins (~horex539@2a02:a03f:6aa5:a00:c0c2:de35:17c6:545e) |
| 2021-05-24 16:55:43 | × | kluk quits (~kluk@2603-7000-9b3f-6934-6417-29db-6208-fab1.res6.spectrum.com) (Read error: Connection reset by peer) |
| 2021-05-24 16:56:26 | → | kluk joins (~kluk@2603-7000-9b3f-6934-6417-29db-6208-fab1.res6.spectrum.com) |
| 2021-05-24 16:56:56 | <zzz> | i see harm in success at cost of PhD thesis |
| 2021-05-24 16:57:54 | <Rembane> | Isn't that one of the tensions of Haskell? Research language vs industrial language? |
| 2021-05-24 16:57:59 | <Cale> | Not all Ph.D. theses are created equal when it comes to practicality. |
| 2021-05-24 16:58:36 | <maerwald> | I believe the solution is interoperable languages, instead of one language that does it all. MS is closest to it with C# -> F# -> F* |
| 2021-05-24 16:59:18 | <maerwald> | but that requires more pragmatism than we have in the haskell community (and more resources) |
| 2021-05-24 16:59:57 | <Rembane> | I think resources are a bigger problem than pragmatism in this case, and also what the community optimizes for. |
| 2021-05-24 16:59:59 | <Cale> | One thing which would be really good is if we can sort out GHC so that it's more modular. |
| 2021-05-24 17:00:35 | → | slep joins (~slep@cpc150002-brnt4-2-0-cust437.4-2.cable.virginm.net) |
| 2021-05-24 17:00:59 | <boxscape> | any idea why I might be getting "Text.ParserCombinators.Parsec.Prim.many: combinator 'many' is applied to a parser that accepts an empty string" when the stack trace shows I'm in this function? |
| 2021-05-24 17:00:59 | <boxscape> | nonNewLineSpaces = skipMany $ satisfy \c -> isSpace c && c /= '\n' |
| 2021-05-24 17:01:26 | → | tose joins (~tose@ip-85-160-8-188.eurotel.cz) |
| 2021-05-24 17:01:52 | <Cale> | Don't you need another $ there? But yeah, that's not a parser that can accept an empty string, so I'm not sure. |
| 2021-05-24 17:02:10 | <boxscape> | not with one of those questionable extensions, BlockArguments |
| 2021-05-24 17:02:18 | <Cale> | haha, okay |
| 2021-05-24 17:02:23 | <Cale> | Where'd you get the stack trace from? |
| 2021-05-24 17:02:23 | × | kluk quits (~kluk@2603-7000-9b3f-6934-6417-29db-6208-fab1.res6.spectrum.com) (Read error: Connection reset by peer) |
| 2021-05-24 17:02:46 | <boxscape> | ghc check-notes.hs -prof -fprof-auto -fprof-cafs, followed by cat tmp | ./check-notes +RTS -xc |
| 2021-05-24 17:02:58 | <Cale> | Ah, okay, a cost-centre trace |
| 2021-05-24 17:03:15 | <boxscape> | ah, yes I suppose that's a more accurate name |
| 2021-05-24 17:03:46 | <boxscape> | output looks like this which seems like it would mean that it has to be from that skipMany https://paste.tomsmeding.com/3p9tmXnx |
| 2021-05-24 17:04:24 | → | kluk joins (~kluk@2603-7000-9b3f-6934-6417-29db-6208-fab1.res6.spectrum.com) |
| 2021-05-24 17:05:33 | → | Tomurb joins (~tom@92-17.net.optinet.cz) |
| 2021-05-24 17:05:34 | <Cale> | What does your parser look like? Maybe the problem is just somewhere nearby? |
| 2021-05-24 17:06:02 | → | shapr joins (~user@pool-100-36-247-68.washdc.fios.verizon.net) |
| 2021-05-24 17:06:10 | × | berberman quits (~berberman@2408:8207:2560:3e20::a44) (Ping timeout: 264 seconds) |
| 2021-05-24 17:06:20 | × | coot quits (~coot@37.30.58.122.nat.umts.dynamic.t-mobile.pl) (Quit: coot) |
| 2021-05-24 17:06:26 | × | imdoor quits (~imdoor@balticom-142-78-50.balticom.lv) (Quit: imdoor) |
| 2021-05-24 17:07:06 | → | Deide joins (~Deide@217.155.19.23) |
| 2021-05-24 17:07:06 | × | Deide quits (~Deide@217.155.19.23) (Changing host) |
| 2021-05-24 17:07:06 | → | Deide joins (~Deide@user/deide) |
| 2021-05-24 17:07:12 | <boxscape> | yeah it contains a bunch of manys and some parsers that can produce empty strings. I guess I'll go through them and see disabling which one fixes it, maybe I'm misunderstanding how the trace works. |
| 2021-05-24 17:07:22 | → | sondre joins (~sondrelun@cm-84.212.100.140.getinternet.no) |
| 2021-05-24 17:07:23 | <boxscape> | s/produce/accept |
| 2021-05-24 17:07:37 | <Cale> | boxscape: Perhaps nonNewLineSpaces is being used in a many? |
| 2021-05-24 17:07:51 | <Cale> | What does singleLineNotes look like? |
| 2021-05-24 17:08:52 | <Cale> | I wouldn't trust the trace too much, they're often kinda... approximate. They mean a whole lot less than you'd want for debugging, but can be better than nothing. |
| 2021-05-24 17:09:06 | × | kluk quits (~kluk@2603-7000-9b3f-6934-6417-29db-6208-fab1.res6.spectrum.com) (Read error: Connection reset by peer) |
| 2021-05-24 17:09:57 | → | kluk joins (~kluk@2603-7000-9b3f-6934-6417-29db-6208-fab1.res6.spectrum.com) |
| 2021-05-24 17:10:24 | <boxscape> | this is singleLineNotes, there's a bunch of manys in here, let me check if I'm inadverntantly nesting them somewhere https://paste.tomsmeding.com/W3XdvVtm |
| 2021-05-24 17:10:25 | <Cale> | It might also help to test your parsers individually, perhaps just do some quick examples in GHCi using testParser -- usually a problem like this one will show up immediately if it's going to show up, unless it's being buried by a conditional somehow |
| 2021-05-24 17:10:32 | <boxscape> | yeah that's a good idea |
All times are in UTC.