Logs on 2025-03-28 (liberachat/#haskell)
| 00:01:32 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 00:05:23 | → | anselmschueler joins (~quassel@user/schuelermine) |
| 00:05:24 | <anselmschueler> | Hi |
| 00:05:24 | × | sprotte24 quits (~sprotte24@p200300d16f0a80004582713a1a0df846.dip0.t-ipconnect.de) (Read error: Connection reset by peer) |
| 00:05:28 | <anselmschueler> | ’s been a while |
| 00:05:55 | → | xff0x joins (~xff0x@2405:6580:b080:900:879:4ff2:52d5:5929) |
| 00:06:05 | geekosaur | waves |
| 00:07:51 | <anselmschueler> | does anyone know when stimes was changed to document that it’ll work with 0 only for some implementors, and that stimesMonoid was added to enforce stimesMonoid 0 = mempty? |
| 00:07:59 | <anselmschueler> | I seem to remember it wasn’t so some time ago |
| 00:08:19 | <anselmschueler> | specifically when I wrote this rejected proposal: https://github.com/haskell/core-libraries-committee/issues/72 |
| 00:12:23 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 00:12:40 | → | peterbecich joins (~Thunderbi@syn-047-229-123-186.res.spectrum.com) |
| 00:14:56 | × | tdammers quits (~tdammers@110-136-178-143.ftth.glasoperator.nl) (Ping timeout: 265 seconds) |
| 00:17:25 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 00:27:51 | → | tdammers joins (~tdammers@110-136-178-143.ftth.glasoperator.nl) |
| 00:28:09 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 00:30:30 | × | toby-bro quits (~toby-bro@user/toby-bro) (Ping timeout: 276 seconds) |
| 00:32:03 | → | malte joins (~malte@mal.tc) |
| 00:34:41 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 00:35:21 | × | st_aldini quits (~Thunderbi@136.48.22.91) (Remote host closed the connection) |
| 00:36:12 | × | jacopovalanzano quits (~jacopoval@cpc151911-cove17-2-0-cust105.3-1.cable.virginm.net) (Quit: Client closed) |
| 00:39:30 | → | JuanDaugherty joins (~juan@user/JuanDaugherty) |
| 00:41:52 | × | aetepe quits (~aetepe@188.119.58.34) (Ping timeout: 252 seconds) |
| 00:44:34 | × | olivial quits (~benjaminl@user/benjaminl) (Ping timeout: 260 seconds) |
| 00:46:13 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 00:51:47 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 268 seconds) |
| 00:54:15 | × | acidjnk quits (~acidjnk@p200300d6e71c4f64e0b826361c3e438a.dip0.t-ipconnect.de) (Ping timeout: 268 seconds) |
| 00:55:34 | × | anselmschueler quits (~quassel@user/schuelermine) (Remote host closed the connection) |
| 00:55:57 | × | Tuplanolla quits (~Tuplanoll@91-159-69-59.elisa-laajakaista.fi) (Quit: Leaving.) |
| 00:58:11 | → | anselmschueler joins (~quassel@user/schuelermine) |
| 00:58:14 | <EvanR> | :t \n -> (!! n) . iterate |
| 00:58:15 | <lambdabot> | error: |
| 00:58:15 | <lambdabot> | • Couldn't match type ‘a -> [a]’ with ‘[c]’ |
| 00:58:15 | <lambdabot> | Expected type: (a -> a) -> [c] |
| 00:58:35 | <EvanR> | :t \n f -> (!! n) . iterate f |
| 00:58:36 | <lambdabot> | Int -> (c -> c) -> c -> c |
| 01:01:59 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 01:04:21 | <anselmschueler> | :h |
| 01:04:24 | <anselmschueler> | :? |
| 01:04:30 | <anselmschueler> | hm |
| 01:05:31 | <geekosaur> | it only supports :t and :k (from ghci) plus "> " to evaluate expressions, otherwise commands start with @ |
| 01:05:36 | <anselmschueler> | oh |
| 01:05:38 | <anselmschueler> | ok |
| 01:06:25 | <geekosaur> | (the in-channel help isn't so helpful, you might prefer https://github.com/lambdabot/lambdabot/pull/205/files) |
| 01:06:48 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 01:08:02 | <anselmschueler> | @djinn [a] -> (a -> [b]) -> [b] |
| 01:08:02 | <lambdabot> | Error: Undefined type [] |
| 01:09:18 | <geekosaur> | djinn doesn't understand recursively defined types. there was a bot that did (exference) but it's bitrotted, and even when working it could fail to find solutions |
| 01:09:42 | <anselmschueler> | geekosaur: the source code is just lost? |
| 01:09:44 | <anselmschueler> | sad |
| 01:09:50 | <anselmschueler> | I assume it’s also a hard problem |
| 01:09:54 | <anselmschueler> | to generate those |
| 01:10:18 | <geekosaur> | no, it's there, it was broken by compiler changes and I couldn't figure out how to update it when I tried forking it |
| 01:10:23 | <geekosaur> | ghc internals change a lot |
| 01:10:52 | <geekosaur> | https://github.com/lspitzner/exference |
| 01:11:02 | <anselmschueler> | ok |
| 01:11:13 | × | xff0x quits (~xff0x@2405:6580:b080:900:879:4ff2:52d5:5929) (Ping timeout: 248 seconds) |
| 01:12:16 | <anselmschueler> | hm @free isn’t working |
| 01:12:19 | <anselmschueler> | @free fmap |
| 01:12:20 | <lambdabot> | Extra stuff at end of line in retrieved type "Functor f => (a -> b) -> f a -> f b" |
| 01:12:48 | <Leary> | @djinn (r -> (a -> r -> r) -> r) -> (a -> (s -> (b -> s -> s) -> s)) -> (t -> (b -> t -> t) -> t) |
| 01:12:48 | <lambdabot> | f _ _ a _ = a |
| 01:12:51 | <geekosaur> | @free map |
| 01:12:52 | <lambdabot> | g . h = k . f => $map g . map h = map k . $map f |
| 01:13:16 | <geekosaur> | which is actually fmap; this is something of a hack because @free doesn't understand typeclasses |
| 01:13:25 | × | otto_s quits (~user@p5de2f7cc.dip0.t-ipconnect.de) (Ping timeout: 265 seconds) |
| 01:13:25 | <Leary> | (i.e. `f _ _ = []`) |
| 01:13:29 | <anselmschueler> | oh |
| 01:13:52 | <anselmschueler> | @free fmap @[] |
| 01:13:52 | <lambdabot> | Extra stuff at end of line |
| 01:14:01 | <anselmschueler> | :( |
| 01:14:09 | <Axman6> | you can play with lambdabot in private messages too |
| 01:14:15 | <anselmschueler> | ok |
| 01:14:19 | <Leary> | @free fmap :: (a -> b) -> F a -> F b |
| 01:14:19 | <lambdabot> | g . h = k . f => $map_F g . fmap h = fmap k . $map_F f |
| 01:14:28 | <anselmschueler> | it’s been too long since I’ve done Haskell |
| 01:14:33 | × | vanishingideal quits (~vanishing@user/vanishingideal) (Ping timeout: 245 seconds) |
| 01:14:34 | <anselmschueler> | like proper |
| 01:14:54 | → | otto_s joins (~user@p5de2fe2d.dip0.t-ipconnect.de) |
| 01:14:59 | <geekosaur> | well, most of these failures are because most of the bot plugins weren't updated |
| 01:15:10 | <geekosaur> | like, no TypeApplications support |
| 01:15:15 | <anselmschueler> | yeah |
| 01:16:11 | <anselmschueler> | two years ago I wrote a small (230 lines) code (still my best Haskell code to date) for directory traversal |
| 01:16:26 | <anselmschueler> | do you think using existentials for that is overkill? |
| 01:16:40 | → | vanishingideal joins (~vanishing@user/vanishingideal) |
| 01:17:03 | <anselmschueler> | I had data Landmark = forall t. Landmark { prepare :: FilePath -> IO t, check :: t -> Check } |
| 01:17:09 | <anselmschueler> | so when the program starts it collects data |
| 01:17:14 | × | JuanDaugherty quits (~juan@user/JuanDaugherty) (Quit: praxis.meansofproduction.biz (juan@acm.org)) |
| 01:17:26 | <anselmschueler> | and then it starts traversing and each time it checks for each file |
| 01:17:31 | <anselmschueler> | something |
| 01:17:46 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 01:18:00 | <anselmschueler> | and existentials seemed like a natural solution |
| 01:18:04 | <anselmschueler> | but idk if it is anymore |
| 01:19:13 | × | peterbecich quits (~Thunderbi@syn-047-229-123-186.res.spectrum.com) (Ping timeout: 248 seconds) |
| 01:20:08 | <Leary> | It's not really any different from `newtype Landmark = Landmark{ prepareAndCheck :: FilePath -> IO Check }`. |
| 01:20:47 | → | olivial joins (~benjaminl@user/benjaminl) |
| 01:22:05 | × | notdabs quits (~Owner@2600:1700:69cf:9000:9c7c:26a2:e9ed:cf3a) (Read error: Connection reset by peer) |
| 01:22:31 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 01:33:32 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 01:34:08 | <EvanR> | there's usually a way to reword your datatype so that existentials, and sometimes higher rank types aren't required |
| 01:34:42 | <EvanR> | alternatively with higher rank polymorphism, you can make data types not required |
| 01:34:56 | <EvanR> | whatever is more convenient |
| 01:35:14 | → | ljdarj1 joins (~Thunderbi@user/ljdarj) |
| 01:35:43 | × | Googulator quits (~Googulato@2a01-036d-0106-01d5-c415-995d-99e3-7810.pool6.digikabel.hu) (Quit: Client closed) |
| 01:36:00 | → | Googulator joins (~Googulato@2a01-036d-0106-01d5-c415-995d-99e3-7810.pool6.digikabel.hu) |
| 01:37:22 | × | ljdarj quits (~Thunderbi@user/ljdarj) (Ping timeout: 248 seconds) |
| 01:37:22 | ljdarj1 | is now known as ljdarj |
| 01:38:39 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 268 seconds) |
| 01:41:02 | → | aetepe joins (~aetepe@188.119.58.34) |
| 01:44:12 | × | anselmschueler quits (~quassel@user/schuelermine) (Ping timeout: 268 seconds) |
| 01:45:59 | × | aetepe quits (~aetepe@188.119.58.34) (Ping timeout: 260 seconds) |
| 01:49:19 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 01:53:57 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 248 seconds) |
| 01:56:09 | → | weary-traveler joins (~user@user/user363627) |
| 01:59:24 | × | ljdarj quits (~Thunderbi@user/ljdarj) (Quit: ljdarj) |
| 02:03:48 | × | gmg quits (~user@user/gehmehgeh) (Remote host closed the connection) |
| 02:03:54 | × | jmcantrell quits (~weechat@user/jmcantrell) (Quit: WeeChat 4.6.0) |
| 02:04:32 | → | gmg joins (~user@user/gehmehgeh) |
| 02:04:39 | → | jmcantrell joins (~weechat@user/jmcantrell) |
| 02:04:48 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 02:05:20 | → | xff0x joins (~xff0x@fsb6a9491c.tkyc517.ap.nuro.jp) |
| 02:11:38 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 245 seconds) |
| 02:17:39 | → | werneta joins (~werneta@syn-071-083-160-242.res.spectrum.com) |
| 02:22:51 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 02:27:50 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 02:38:36 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 02:39:54 | → | aetepe joins (~aetepe@188.119.58.34) |
| 02:43:30 | → | ensyde joins (~ensyde@2601:5c6:c200:6dc0::8955) |
| 02:43:38 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 02:44:44 | × | aetepe quits (~aetepe@188.119.58.34) (Ping timeout: 260 seconds) |
| 02:51:13 | × | tomku quits (~tomku@user/tomku) (Ping timeout: 245 seconds) |
| 02:51:24 | <haskellbridge> | <Bowuigi> EvanR programming with rank N stuff only can be annoying sometimes, specially on a total system |
| 02:51:28 | → | tomku joins (~tomku@user/tomku) |
| 02:52:16 | <haskellbridge> | <Bowuigi> Mendler-style recursion schemes are fantastic though, if only there was a way to get histomorphisms working on a total setting... |
| 02:54:22 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 02:55:33 | → | aetepe joins (~aetepe@188.119.58.34) |
| 02:59:06 | → | LainExperiments joins (~LainExper@user/LainExperiments) |
| 02:59:22 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 03:00:02 | × | aetepe quits (~aetepe@188.119.58.34) (Ping timeout: 248 seconds) |
| 03:02:48 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 03:07:24 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 03:11:17 | → | aetepe joins (~aetepe@188.119.58.34) |
| 03:12:19 | × | tromp quits (~textual@2001:1c00:3487:1b00:f8c0:4558:f838:6f52) (Ping timeout: 260 seconds) |
| 03:13:23 | × | LainExperiments quits (~LainExper@user/LainExperiments) (Quit: Client closed) |
| 03:14:36 | → | emmanuelux_ joins (~emmanuelu@user/emmanuelux) |
| 03:16:05 | × | emmanuelux quits (~emmanuelu@user/emmanuelux) (Ping timeout: 268 seconds) |
| 03:16:15 | × | aetepe quits (~aetepe@188.119.58.34) (Ping timeout: 276 seconds) |
| 03:17:35 | <haskellbridge> | <thirdofmay18081814goya> what exactly is a higher-order effect that algebraic effects can't handle? |
| 03:17:51 | <haskellbridge> | <thirdofmay18081814goya> is it an effect that takes another effect? something like "Effect a -> Effect b"? |
| 03:18:33 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 03:21:33 | → | LainExperiments joins (~LainExper@user/LainExperiments) |
| 03:23:25 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 03:34:20 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 03:38:59 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 260 seconds) |
| 03:39:10 | <Leary> | thirdofmay: `data Catch e m a where Catch :: m a -> (e -> m a) -> Catch e m a`. The effect `runCatchEM :: forall a. Catch E M (M a) -> M a` is a higher order effect because `Catch` needs to know about `M`, and it isn't algebraic because `runCatch c >>= k /= runCatch (c <&> (>>= k))`. |
| 03:40:00 | → | LainExperiments3 joins (LainExperi@user/LainExperiments) |
| 03:41:15 | <jackdk> | Is LainExperiments serially-numbered? |
| 03:42:06 | × | LainExperiments quits (~LainExper@user/LainExperiments) (Ping timeout: 240 seconds) |
| 03:45:08 | <haskellbridge> | <thirdofmay18081814goya> Leary: I see! thanks a lot for the comment! |
| 03:46:52 | <EvanR> | I had to look at join parts to get that one |
| 03:49:31 | <haskellbridge> | <Liamzee> are people doing much with civilian linear types these days? |
| 03:49:52 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 03:56:39 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 260 seconds) |
| 03:57:12 | <haskellbridge> | <Liamzee> i still feel bad about not writing a linearly typed growable vector lib, because it wouldn't have been any good |
| 03:58:15 | <jackdk> | Axman6 and I have a half-done port of streaming-bytestring to linear types but never finished it |
| 03:58:44 | <Axman6> | We should finish that |
| 03:59:17 | <Axman6> | I feel like there wasn't all that much left |
| 03:59:44 | <EvanR> | https://hackage.haskell.org/package/linear-base-0.4.0/docs/Data-Vector-Mutable-Linear.html |
| 04:03:00 | <Axman6> | Is there a linear state monad or something that can make passing around linear values a little nicer? |
| 04:06:28 | <jackdk> | https://hackage.haskell.org/package/linear-base-0.4.0/docs/Control-Functor-Linear.html#g:4 Not the module I would've expected |
| 04:07:56 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 04:09:10 | <haskellbridge> | <Liamzee> oh, i never realized that linear base had dynamic vectors, only mutable ones |
| 04:10:00 | <EvanR> | there is also a destination array thing in there |
| 04:10:43 | <Leary> | thirdofmay: Err, I guess that should have been `data Catch e m a where Catch :: m x -> (Either e x -> a) -> Catch e m a`, which probably clarifies the rest of what I said. |
| 04:12:21 | → | aetepe joins (~aetepe@188.119.58.34) |
| 04:13:01 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 04:14:05 | <haskellbridge> | <Liamzee> but no one has built a linearly typed webserver and put it on hackage? |
| 04:16:49 | <haskellbridge> | <Liamzee> fun recent linear types: |
| 04:16:50 | <haskellbridge> | <Liamzee> https://hackage.haskell.org/package/text-builder-linear |
| 04:16:54 | <haskellbridge> | <Liamzee> by bodigrim |
| 04:17:03 | × | aetepe quits (~aetepe@188.119.58.34) (Ping timeout: 245 seconds) |
| 04:18:36 | × | LainExperiments3 quits (LainExperi@user/LainExperiments) (Quit: Client closed) |
| 04:19:40 | <EvanR> | it's not linearly typed but it does do the thing everyone thinks linear types are for http://www.impredicative.com/ur/ |
| 04:19:49 | <EvanR> | Ur/Web |
| 04:23:23 | → | michalz joins (~michalz@185.246.207.205) |
| 04:23:43 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 04:24:38 | <haskellbridge> | <thirdofmay18081814goya> does anyone know if there's a way to embed a max number of recursive calls in an arbitrary recursive function? |
| 04:24:57 | <haskellbridge> | <thirdofmay18081814goya> i guess necessarily some method with recursion-schemes right |
| 04:26:19 | <haskellbridge> | <thirdofmay18081814goya> the idea would be to have a recursive function whose specification is separate from the requirement that it terminates after n calls |
| 04:26:36 | <haskellbridge> | <thirdofmay18081814goya> so that you can also use it without enforcing the max recursive call |
| 04:27:26 | <haskellbridge> | <thirdofmay18081814goya> trivially a catamorphism whose algebra returns a pair can do this, any other ideas? |
| 04:28:06 | → | aetepe joins (~aetepe@188.119.58.34) |
| 04:28:21 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 246 seconds) |
| 04:28:46 | <EvanR> | this isn't doesn't "monkey patch" the way you want, but the partiality monad will let you decide elsewhere how long a recursive algorithm gets to terminate |
| 04:29:04 | <Leary> | I would write `fixMaxDepth :: Int -> (a -> a) -> Maybe a`. |
| 04:29:24 | <Leary> | Err, maybe not. |
| 04:29:53 | <Leary> | `fixMaxDepth :: Int -> (a -> Either a b) -> Maybe b`? |
| 04:29:55 | → | LainExperiments joins (~LainExper@user/LainExperiments) |
| 04:30:10 | <EvanR> | well, now it's nearly the partiality monad |
| 04:30:48 | × | aforemny_ quits (~aforemny@2001:9e8:6cdb:6d00:dc2e:87c9:d5a9:a75) (Ping timeout: 245 seconds) |
| 04:31:23 | → | aforemny joins (~aforemny@i577B139C.versanet.de) |
| 04:31:30 | <haskellbridge> | <thirdofmay18081814goya> i'll check out the partiality monad, thanks! |
| 04:33:14 | × | aetepe quits (~aetepe@188.119.58.34) (Ping timeout: 260 seconds) |
| 04:34:47 | → | LainExperiments4 joins (~LainExper@user/LainExperiments) |
| 04:36:06 | × | LainExperiments quits (~LainExper@user/LainExperiments) (Ping timeout: 240 seconds) |
| 04:36:07 | → | bitdex joins (~bitdex@gateway/tor-sasl/bitdex) |
| 04:38:54 | → | peterbecich joins (~Thunderbi@syn-047-229-123-186.res.spectrum.com) |
| 04:39:30 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 04:40:52 | Digitteknohippie | is now known as Digit |
| 04:44:37 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 248 seconds) |
| 04:51:33 | × | haskellbridge quits (~hackager@syn-024-093-192-219.res.spectrum.com) (Ping timeout: 252 seconds) |
| 04:53:05 | → | Digitteknohippie joins (~user@user/digit) |
| 04:54:07 | × | Digit quits (~user@user/digit) (Ping timeout: 252 seconds) |
| 04:54:24 | → | haskellbridge joins (~hackager@syn-024-093-192-219.res.spectrum.com) |
| 04:54:24 | ChanServ | sets mode +v haskellbridge |
| 04:55:18 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 04:55:23 | × | xff0x quits (~xff0x@fsb6a9491c.tkyc517.ap.nuro.jp) (Ping timeout: 245 seconds) |
| 04:55:35 | <jackdk> | thirdofmay18081814goya: Hughes does this for a game tree by generating the infinite tree and then pruning it to N levels, which works because laziness |
| 05:00:02 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 05:03:49 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 05:06:30 | × | LainExperiments4 quits (~LainExper@user/LainExperiments) (Ping timeout: 240 seconds) |
| 05:08:08 | <haskellbridge> | <thirdofmay18081814goya> jackdk: ah very neat, will check this out too, have not considered any notion of laziness so far |
| 05:08:10 | <haskellbridge> | <thirdofmay18081814goya> thanks! |
| 05:08:49 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 05:10:44 | × | peterbecich quits (~Thunderbi@syn-047-229-123-186.res.spectrum.com) (Ping timeout: 265 seconds) |
| 05:10:54 | → | xff0x joins (~xff0x@fsb6a9491c.tkyc517.ap.nuro.jp) |
| 05:12:03 | × | gmg quits (~user@user/gehmehgeh) (Remote host closed the connection) |
| 05:12:45 | → | gmg joins (~user@user/gehmehgeh) |
| 05:13:33 | → | LainExperiments joins (~LainExper@user/LainExperiments) |
| 05:17:39 | Digitteknohippie | is now known as Digit |
| 05:19:33 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 05:21:56 | <ski> | Leary : `Either e x -> a' a la "Exceptional Syntax" ? |
| 05:22:06 | <ski> | EvanR : Ur has uniqueness ? |
| 05:22:53 | <EvanR> | does it? |
| 05:23:01 | <ski> | i dunno |
| 05:23:10 | <ski> | you said "it does do the thing everyone thinks linear types are for" |
| 05:24:07 | <EvanR> | memory management, or lack there of |
| 05:24:34 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 248 seconds) |
| 05:27:19 | <ski> | mhm |
| 05:29:09 | → | aetepe joins (~aetepe@188.119.58.34) |
| 05:32:48 | → | Digitteknohippie joins (~user@user/digit) |
| 05:33:06 | <Leary> | ski: I'm not familiar with the paper. I pulled that `data Catch` out of my arse, though it may have been inspired by 'A Framework for Higher-Order Effects & Handlers'. |
| 05:33:37 | × | aetepe quits (~aetepe@188.119.58.34) (Ping timeout: 244 seconds) |
| 05:34:13 | × | Digit quits (~user@user/digit) (Ping timeout: 268 seconds) |
| 05:35:20 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 05:36:40 | Digitteknohippie | is now known as Digit |
| 05:41:03 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 05:44:35 | × | ChaiTRex quits (~ChaiTRex@user/chaitrex) (Remote host closed the connection) |
| 05:46:08 | → | ChaiTRex joins (~ChaiTRex@user/chaitrex) |
| 05:46:09 | → | takuan joins (~takuan@d8D86B601.access.telenet.be) |
| 05:51:54 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 05:53:14 | → | Square joins (~Square@user/square) |
| 05:57:08 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 265 seconds) |
| 05:57:19 | × | weary-traveler quits (~user@user/user363627) (Remote host closed the connection) |
| 05:58:01 | <ski> | Leary : "Exceptional Syntax" by Nick Benton,Andrew Kennedy in 2001 at <https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/exceptionalsyntax.pdf> (<http://lambda-the-ultimate.org/node/1193) |
| 05:58:19 | → | nvoid joins (~nvoid@2601:140:8700:25fe:5820:136c:b0ce:f5b1) |
| 05:59:07 | <ski> | basically argues for (in Haskell terms) `catchBind :: Exception e => IO a -> (e -> IO b) -> (a -> IO b) -> IO b' being more appropriate as a primitive, than `catch :: Exception e => IO a -> (e -> IO a) -> IO a' |
| 05:59:40 | <ski> | one reason being tail calls |
| 06:04:46 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 06:05:47 | × | nvoid quits (~nvoid@2601:140:8700:25fe:5820:136c:b0ce:f5b1) (Quit: nvoid) |
| 06:09:40 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 06:10:06 | × | LainExperiments quits (~LainExper@user/LainExperiments) (Ping timeout: 240 seconds) |
| 06:10:32 | <sim590> | Consider the following problem: https://projecteuler.net/problem=32. Which I have solved. My algorithm uses some math analysis to determine an upper bound for the search space. Anyway, I use the following function to recognize a pandigital triplet: https://paste.debian.net/1365804/. Using this function, my algorithms gave me the right answer. Then, I decided to rewrite the function a bit to make |
| 06:10:34 | <sim590> | it simpler. I rewrote it to https://paste.debian.net/1365805/. I don't understand why, but now my whole algorithme doesn't give me the right answer anymore. |
| 06:12:49 | <sim590> | To me, I thought that the second version made more sense... I'm now doubting if I accidently got the right answer because of a bad upper bound and a bad algorithm... lol |
| 06:16:51 | × | rvalue quits (~rvalue@user/rvalue) (Read error: Connection reset by peer) |
| 06:17:23 | → | rvalue joins (~rvalue@user/rvalue) |
| 06:19:09 | → | LainExperiments joins (~LainExper@user/LainExperiments) |
| 06:20:34 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 06:20:57 | <sim590> | The funny thing is with the first function, the algorithm executes in 280 seconds while with the second function, it executes in 420 seconds... (it's by now means optimal, I tried a brute force approach with the help of an upper bound) |
| 06:22:51 | × | emmanuelux_ quits (~emmanuelu@user/emmanuelux) (Read error: Connection reset by peer) |
| 06:23:49 | <sim590> | Ah. I get it. The second version doesn't take into account repeating numbers while the Set approach took care of that. |
| 06:24:46 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 06:24:49 | × | hattckory quits (~hattckory@bras-base-toroon4524w-grc-30-70-27-118-207.dsl.bell.ca) (Ping timeout: 248 seconds) |
| 06:30:14 | → | aetepe joins (~aetepe@188.119.58.34) |
| 06:30:52 | <Leary> | thirdofmay: I might be beating a dead horse at this point, but there are some details I kinda bungled due to algebraicity depending on precisely how on operation is expressed (and I prefer to fix these things). `catch :: M a -> (E -> M a) -> M a` naively becomes `catch :: Sig_Catch (M a) -> M a` with `Sig_Catch r = (r, E -> r)` which /isn't/ algebraic, but the last `data Catch` I wrote has a `runCatch` that actually /is/. That's because I cheated and us |
| 06:30:53 | <Leary> | ed `try` disguised as `catch`: `M a -> M (Either E a)` ~ `M a -> Codensity M (Either E a)` ~ `M a -> (Either E a -> M b) -> M b` ~ `Sig_Try (M b) -> M b` where `Sig_Try r = (M a, Either E a -> r)`. |
| 06:31:40 | → | hattckory joins (~hattckory@70.27.118.207) |
| 06:33:50 | <Leary> | ski: I see. Those tail calls and algebraicity are probably related. |
| 06:34:57 | × | aetepe quits (~aetepe@188.119.58.34) (Ping timeout: 248 seconds) |
| 06:35:57 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 06:41:00 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 06:43:22 | × | echoreply quits (~echoreply@2001:19f0:9002:1f3b:5400:ff:fe6f:8b8d) (Quit: WeeChat 2.8) |
| 06:43:55 | → | echoreply joins (~echoreply@45.32.163.16) |
| 06:43:59 | → | Digitteknohippie joins (~user@user/digit) |
| 06:45:13 | × | Digit quits (~user@user/digit) (Ping timeout: 252 seconds) |
| 06:48:50 | <haskellbridge> | <thirdofmay18081814goya> Leary: am parsing these comments, thanks for the answer! |
| 06:51:45 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 06:56:09 | × | hattckory quits (~hattckory@70.27.118.207) (Ping timeout: 260 seconds) |
| 06:58:18 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 245 seconds) |
| 07:00:00 | × | caconym quits (~caconym@user/caconym) (Quit: bye) |
| 07:01:02 | → | caconym joins (~caconym@user/caconym) |
| 07:03:08 | Digitteknohippie | is now known as Digit |
| 07:05:36 | → | aetepe joins (~aetepe@188.119.58.34) |
| 07:05:46 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 07:10:15 | × | aetepe quits (~aetepe@188.119.58.34) (Ping timeout: 276 seconds) |
| 07:10:32 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 07:15:55 | × | Lord_of_Life quits (~Lord@user/lord-of-life/x-2819915) (Ping timeout: 244 seconds) |
| 07:15:59 | × | ensyde quits (~ensyde@2601:5c6:c200:6dc0::8955) (Ping timeout: 260 seconds) |
| 07:17:48 | → | ensyde joins (~ensyde@2601:5c6:c200:6dc0::6f7f) |
| 07:18:43 | × | jmcantrell quits (~weechat@user/jmcantrell) (Ping timeout: 244 seconds) |
| 07:19:11 | × | ft quits (~ft@p508db463.dip0.t-ipconnect.de) (Quit: leaving) |
| 07:19:25 | → | Lord_of_Life joins (~Lord@user/lord-of-life/x-2819915) |
| 07:21:32 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 07:26:39 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 260 seconds) |
| 07:30:06 | × | LainExperiments quits (~LainExper@user/LainExperiments) (Quit: Client closed) |
| 07:31:37 | → | ash3en joins (~Thunderbi@ip1f10cbd6.dynamic.kabel-deutschland.de) |
| 07:33:39 | → | wootehfoot joins (~wootehfoo@user/wootehfoot) |
| 07:37:21 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 07:42:14 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 260 seconds) |
| 07:44:16 | × | connrs quits (~connrs@user/connrs) (Quit: ZNC 1.9.1 - https://znc.in) |
| 07:52:39 | × | ash3en quits (~Thunderbi@ip1f10cbd6.dynamic.kabel-deutschland.de) (Ping timeout: 265 seconds) |
| 07:53:18 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 07:54:35 | × | L29Ah quits (~L29Ah@wikipedia/L29Ah) (Read error: Connection timed out) |
| 07:56:56 | → | acidjnk joins (~acidjnk@p200300d6e71c4f71c835c1b6e3010b6c.dip0.t-ipconnect.de) |
| 07:58:11 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 08:00:15 | → | lortabac joins (~lortabac@2a01:e0a:541:b8f0:55ab:e185:7f81:54a4) |
| 08:00:19 | → | connrs joins (~connrs@user/connrs) |
| 08:02:36 | → | CiaoSen joins (~Jura@2a02:8071:64e1:da0:5a47:caff:fe78:33db) |
| 08:04:51 | → | developer_ joins (~developer@85.50.149.196) |
| 08:06:46 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 08:07:14 | → | vpan joins (~vpan@212.117.1.172) |
| 08:11:39 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 245 seconds) |
| 08:12:06 | × | Googulator quits (~Googulato@2a01-036d-0106-01d5-c415-995d-99e3-7810.pool6.digikabel.hu) (Ping timeout: 240 seconds) |
| 08:14:37 | → | auri joins (~auri@fsf/member/auri) |
| 08:21:30 | → | __monty__ joins (~toonn@user/toonn) |
| 08:22:34 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 08:29:29 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 260 seconds) |
| 08:33:44 | × | econo_ quits (uid147250@id-147250.tinside.irccloud.com) (Quit: Connection closed for inactivity) |
| 08:40:37 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 08:43:44 | → | sord937 joins (~sord937@gateway/tor-sasl/sord937) |
| 08:44:54 | × | sprout quits (~sprout@84-80-106-227.fixed.kpn.net) (Ping timeout: 246 seconds) |
| 08:45:30 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 08:52:36 | × | Sgeo quits (~Sgeo@user/sgeo) (Read error: Connection reset by peer) |
| 08:53:26 | → | hattckory joins (~hattckory@bras-base-toroon4524w-grc-30-70-27-118-207.dsl.bell.ca) |
| 08:56:23 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 08:58:09 | × | hattckory quits (~hattckory@bras-base-toroon4524w-grc-30-70-27-118-207.dsl.bell.ca) (Ping timeout: 276 seconds) |
| 08:58:57 | <haskellbridge> | <Liamzee> also, Axman6 |
| 08:58:58 | <haskellbridge> | <Liamzee> https://media.discordapp.net/attachments/968989726633779215/1355089415214465144/image.png?ex=67e7a8dc&is=67e6575c&hm=c6a608fb0a4529004be2bd4ee61b350e88d825a2d5cbbfada84b9841929023db&=&format=webp&quality=lossless&width=2548&height=1212 |
| 08:59:25 | → | lxsameer joins (~lxsameer@Serene/lxsameer) |
| 09:01:44 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 260 seconds) |
| 09:02:23 | → | ash3en joins (~Thunderbi@185.209.196.192) |
| 09:06:17 | → | sprout joins (~sprout@84-80-106-227.fixed.kpn.net) |
| 09:07:46 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 09:10:15 | × | sord937 quits (~sord937@gateway/tor-sasl/sord937) (Remote host closed the connection) |
| 09:10:35 | → | sord937 joins (~sord937@gateway/tor-sasl/sord937) |
| 09:11:08 | → | aetepe joins (~aetepe@188.119.58.34) |
| 09:12:33 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 246 seconds) |
| 09:15:20 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 09:17:20 | × | sord937 quits (~sord937@gateway/tor-sasl/sord937) (Quit: sord937) |
| 09:18:14 | → | Googulator joins (~Googulato@81.183.235.203) |
| 09:18:43 | × | tzh quits (~tzh@c-76-115-131-146.hsd1.or.comcast.net) (Quit: zzz) |
| 09:19:36 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 09:25:05 | → | fp joins (~Thunderbi@130.233.70.95) |
| 09:25:16 | × | Googulator quits (~Googulato@81.183.235.203) (Quit: Client closed) |
| 09:25:29 | → | Googulator joins (~Googulato@81.183.235.203) |
| 09:25:46 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 09:27:00 | → | tromp joins (~textual@2001:1c00:3487:1b00:6095:11f3:6fa5:fb3d) |
| 09:27:01 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Read error: Connection reset by peer) |
| 09:28:34 | → | merijn joins (~merijn@77.242.116.146) |
| 09:28:51 | × | Square quits (~Square@user/square) (Remote host closed the connection) |
| 09:31:59 | → | sord937 joins (~sord937@gateway/tor-sasl/sord937) |
| 09:38:27 | → | chele joins (~chele@user/chele) |
| 09:43:41 | × | j1n37 quits (~j1n37@user/j1n37) (Ping timeout: 244 seconds) |
| 09:44:25 | → | j1n37 joins (~j1n37@user/j1n37) |
| 09:44:54 | → | alp joins (~alp@2001:861:8ca0:4940:97d8:ee2c:4105:6ce6) |
| 09:48:20 | × | sord937 quits (~sord937@gateway/tor-sasl/sord937) (Remote host closed the connection) |
| 09:48:39 | → | sord937 joins (~sord937@gateway/tor-sasl/sord937) |
| 09:53:25 | × | arahael quits (~arahael@user/arahael) (Ping timeout: 248 seconds) |
| 09:53:28 | × | sord937 quits (~sord937@gateway/tor-sasl/sord937) (Remote host closed the connection) |
| 09:53:48 | → | sord937 joins (~sord937@gateway/tor-sasl/sord937) |
| 09:54:41 | → | tabaqui joins (~tabaqui@167.71.80.236) |
| 09:55:48 | × | CiaoSen quits (~Jura@2a02:8071:64e1:da0:5a47:caff:fe78:33db) (Ping timeout: 245 seconds) |
| 09:56:46 | × | j1n37 quits (~j1n37@user/j1n37) (Read error: Connection reset by peer) |
| 09:56:54 | × | Googulator quits (~Googulato@81.183.235.203) (Quit: Client closed) |
| 09:57:12 | → | Googulator joins (~Googulato@81.183.235.203) |
| 09:58:17 | → | wildsalander joins (~wildsalan@47.red-80-29-238.dynamicip.rima-tde.net) |
| 09:58:42 | × | wildsalander quits (~wildsalan@47.red-80-29-238.dynamicip.rima-tde.net) (Client Quit) |
| 09:59:58 | → | j1n37 joins (~j1n37@user/j1n37) |
| 10:01:08 | → | Square joins (~Square@user/square) |
| 10:04:02 | × | xff0x quits (~xff0x@fsb6a9491c.tkyc517.ap.nuro.jp) (Ping timeout: 248 seconds) |
| 10:06:29 | → | ljdarj joins (~Thunderbi@user/ljdarj) |
| 10:17:15 | → | dhil joins (~dhil@2a0c:b381:52e:3600:cbb4:807:319f:c7af) |
| 10:21:52 | → | CiaoSen joins (~Jura@2a02:8071:64e1:da0:5a47:caff:fe78:33db) |
| 10:25:42 | × | pavonia quits (~user@user/siracusa) (Quit: Bye!) |
| 10:26:18 | × | tt12310978324354 quits (~tt1231@2603:6010:8700:4a81:219f:50d3:618a:a6ee) (Quit: The Lounge - https://thelounge.chat) |
| 10:27:20 | → | tt12310978324354 joins (~tt1231@2603:6010:8700:4a81:219f:50d3:618a:a6ee) |
| 10:28:12 | <haskellbridge> | <Morj> I've been playing around with using texinfo as a target for structured docs of haddock and other tools. Here's the unstyled result: https://morj.men/ron_html/Quick-start.html and the hackage page for comparison: https://hackage.haskell.org/package/ron-hs-0.4.0/docs/Data-Ron.html |
| 10:29:33 | <haskellbridge> | <Morj> It's fun to see how much importance the creators of texinfo placed on the printed material, compared to now where I want options for easier hierarchy (like a module has classes, classes have functions and implementors, all hyperlinked). This forced me to use '@node' everywhere which inserts a page break |
| 10:31:07 | <haskellbridge> | <Morj> I also wouldn't call it good for hyperlinked manuals: the dichotomy how there is a node hierarchy, and a chapter-subchapter hierarchy is weird, plus there isn't even an option to make the table of contents automatically |
| 10:31:25 | <haskellbridge> | <Morj> But this was fun, thanks to someone from here for telling me of this format |
| 10:42:18 | × | ash3en quits (~Thunderbi@185.209.196.192) (Ping timeout: 265 seconds) |
| 10:49:49 | × | Square quits (~Square@user/square) (Ping timeout: 244 seconds) |
| 10:53:49 | <haskellbridge> | <Liamzee> wait, what? |
| 10:53:51 | <haskellbridge> | <Liamzee> https://hackage.haskell.org/package/linear-base-0.4.0/docs/src/Data.List.Linear.html#take |
| 10:53:55 | <haskellbridge> | <Liamzee> look at the take function, why is it like this? |
| 10:56:45 | → | ash3en joins (~Thunderbi@185.209.196.192) |
| 11:00:06 | → | L29Ah joins (~L29Ah@wikipedia/L29Ah) |
| 11:15:01 | × | ensyde quits (~ensyde@2601:5c6:c200:6dc0::6f7f) (Ping timeout: 248 seconds) |
| 11:18:08 | × | lortabac quits (~lortabac@2a01:e0a:541:b8f0:55ab:e185:7f81:54a4) (Ping timeout: 272 seconds) |
| 11:20:48 | × | ljdarj quits (~Thunderbi@user/ljdarj) (Ping timeout: 245 seconds) |
| 11:20:50 | → | ubert1 joins (~Thunderbi@2a02:8109:ab8a:5a00:7d86:d0e8:c2d1:2ee3) |
| 11:26:49 | × | acidjnk quits (~acidjnk@p200300d6e71c4f71c835c1b6e3010b6c.dip0.t-ipconnect.de) (Ping timeout: 252 seconds) |
| 11:27:44 | × | ash3en quits (~Thunderbi@185.209.196.192) (Ping timeout: 265 seconds) |
| 11:36:37 | <haskellbridge> | <jv> Liamzee: i remember having it done beautifully on this channel (or maybe #haskell-beginners). should be in the logs from about 6 years ago |
| 11:36:54 | <haskellbridge> | <jv> or so |
| 11:36:55 | <jackdk> | Because if you have [a] %1 -> b, then you need to do away with the entire list, not just the prefix that you take |
| 11:37:36 | <haskellbridge> | <jv> under a different nick though |
| 11:38:07 | → | JuanDaugherty joins (~juan@user/JuanDaugherty) |
| 11:39:18 | × | Googulator quits (~Googulato@81.183.235.203) (Ping timeout: 240 seconds) |
| 11:57:13 | × | aetepe quits (~aetepe@188.119.58.34) (Ping timeout: 265 seconds) |
| 11:57:24 | → | lortabac joins (~lortabac@2a01:e0a:541:b8f0:55ab:e185:7f81:54a4) |
| 11:57:35 | → | tremon joins (~tremon@83.80.159.219) |
| 11:58:00 | × | fp quits (~Thunderbi@130.233.70.95) (Ping timeout: 252 seconds) |
| 11:59:04 | → | aetepe joins (~aetepe@188.119.58.34) |
| 11:59:56 | × | CiaoSen quits (~Jura@2a02:8071:64e1:da0:5a47:caff:fe78:33db) (Ping timeout: 272 seconds) |
| 12:00:52 | <tomsmeding> | Liamzee: in other words, that `lseq` is actually doing something, see its definition and the Consumable class |
| 12:03:30 | → | ljdarj joins (~Thunderbi@user/ljdarj) |
| 12:03:51 | × | merijn quits (~merijn@77.242.116.146) (Ping timeout: 252 seconds) |
| 12:04:31 | → | fp joins (~Thunderbi@130.233.70.95) |
| 12:10:58 | × | bitdex quits (~bitdex@gateway/tor-sasl/bitdex) (Quit: = "") |
| 12:16:53 | → | acidjnk joins (~acidjnk@p200300d6e71c4f71c835c1b6e3010b6c.dip0.t-ipconnect.de) |
| 12:17:01 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 12:22:53 | → | Googulator joins (~Googulato@81.183.235.203) |
| 12:23:48 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 265 seconds) |
| 12:28:36 | <__monty__> | I thought `%1 ->` meant at most once, not exactly once? |
| 12:29:23 | × | tromp quits (~textual@2001:1c00:3487:1b00:6095:11f3:6fa5:fb3d) (Quit: My iMac has gone to sleep. ZZZzzz…) |
| 12:33:59 | <yushyin> | i think it is exactly once |
| 12:35:18 | × | Googulator quits (~Googulato@81.183.235.203) (Ping timeout: 240 seconds) |
| 12:35:22 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 12:35:56 | → | ft joins (~ft@p508db463.dip0.t-ipconnect.de) |
| 12:37:58 | <tomsmeding> | __monty__: it means exactly once. At most once would be an "affine" arrow, not a "linear" one, I think, in the prevailing jargon |
| 12:40:36 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 12:43:24 | × | zungi quits (~tory@user/andrewchawk) (Ping timeout: 264 seconds) |
| 12:43:42 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 12:44:27 | <__monty__> | In what context is it important to consume the argument? Limiting to once is easier to reason about, side-effects are a good reason for example. |
| 12:47:14 | → | Googulator joins (~Googulato@81.183.235.203) |
| 12:48:06 | → | zungi joins (~tory@user/andrewchawk) |
| 12:48:21 | <vpan> | hi, `foldMapM` from GHC.Utils.Monad seems useful (found it on hoogle), but calling something from the compiler's namespace feels wrong. There must be a reason those functions aren't in base. Is my intuition correct that typical applications (that don't process Haskell code) should stay away from GHC.*? |
| 12:51:03 | <__monty__> | I think the reason not to use the GHC modules is the size of the package. |
| 12:52:50 | × | Unicorn_Princess quits (~Unicorn_P@user/Unicorn-Princess/x-3540542) (Remote host closed the connection) |
| 12:54:36 | <__monty__> | @hoogle foldrM |
| 12:54:37 | <lambdabot> | Data.Foldable foldrM :: (Foldable t, Monad m) => (a -> b -> m b) -> b -> t a -> m b |
| 12:54:37 | <lambdabot> | Data.Vector.Fusion.Bundle.Monadic foldrM :: Monad m => (a -> b -> m b) -> b -> Bundle m v a -> m b |
| 12:54:37 | <lambdabot> | Data.Vector.Fusion.Stream.Monadic foldrM :: Monad m => (a -> b -> m b) -> b -> Stream m a -> m b |
| 12:54:45 | <__monty__> | That does seem similar. |
| 12:54:54 | <Leary> | :t (getAp .) . foldMap . (Ap .) |
| 12:54:55 | <lambdabot> | forall k (t :: * -> *) (f :: k -> *) (a1 :: k) a2. (Foldable t, Monoid (Ap f a1)) => (a2 -> f a1) -> t a2 -> f a1 |
| 12:57:30 | → | hattckory joins (~hattckory@bras-base-toroon4524w-grc-30-70-27-118-207.dsl.bell.ca) |
| 12:57:37 | → | weary-traveler joins (~user@user/user363627) |
| 12:58:44 | × | chele quits (~chele@user/chele) (Remote host closed the connection) |
| 13:01:18 | <vpan> | thanks for the hints. The trick using `Ap` looks cool. :) |
| 13:03:41 | <tomsmeding> | __monty__: if you require linearity, you can use linear types as a memory management system. Opened files need to be closed, allocated arrays need to be freed; if you have affine arrows only, you still need a GC to clean them up |
| 13:04:49 | × | hattckory quits (~hattckory@bras-base-toroon4524w-grc-30-70-27-118-207.dsl.bell.ca) (Ping timeout: 260 seconds) |
| 13:04:52 | <tomsmeding> | in that sense, linear arrows are "more general": they allow expressing safe APIs for more kinds of resources (i.e. ones that need to be closed/freed/... -- i.e. "consumed" -- at a predictable time) |
| 13:05:30 | <__monty__> | And `free` is the only possible operation that consumes consumes an array without producing one? |
| 13:05:32 | <tomsmeding> | I guess with affine arrows you can still opt-in to explicit consumption as the user, but with linear arrows, the API designer can _require_ explicit consumption |
| 13:05:53 | <tomsmeding> | well, that depends on the API |
| 13:05:57 | <tomsmeding> | for arrays, kind of? |
| 13:06:16 | <tomsmeding> | though sometimes it comes in a different guise, like `toList :: Array a -> [a]` |
| 13:06:25 | <tomsmeding> | er, `toList :: Array a %1-> [a]`, of course |
| 13:06:35 | <tomsmeding> | (I hate the visuals of that %1-> syntax.) |
| 13:07:04 | <__monty__> | Yeah, bring back the lollipop. |
| 13:07:20 | <tomsmeding> | or hell, if you don't want letters in a symbol, do -%> or something |
| 13:08:01 | <tomsmeding> | originally the syntax allowed multiplicity polymorphism, where you could have `forall p. a %p-> b`, but IIRC ghc doesn't currently implement that or something |
| 13:08:22 | <__monty__> | How about array indexing? It'd be pretty inconvenient if that had to free the array. |
| 13:08:53 | <tomsmeding> | oh, apparently ghc does support multiplicity polymorphism |
| 13:09:12 | <tomsmeding> | __monty__: I suggest you peruse this page :) https://hackage.haskell.org/package/linear-base-0.4.0/docs/Data-Array-Mutable-Linear.html#v:get |
| 13:09:31 | <tomsmeding> | % :seti -XLinearTypes |
| 13:09:31 | <yahb2> | <no output> |
| 13:09:39 | <tomsmeding> | % foo :: forall a b p. (a %p-> b) -> a %p-> b ; foo g x = g x |
| 13:09:39 | <yahb2> | <no output> |
| 13:09:55 | × | JuanDaugherty quits (~juan@user/JuanDaugherty) (Quit: praxis.meansofproduction.biz (juan@acm.org)) |
| 13:09:56 | <__monty__> | Is it still all smoke and mirrors or does linear data actually get allocated/freed outside of the GCed memory? |
| 13:09:58 | <tomsmeding> | % :seti -fprint-explicit-foralls |
| 13:09:58 | <yahb2> | <no output> |
| 13:10:01 | <tomsmeding> | % :t foo |
| 13:10:01 | <yahb2> | foo ; :: forall a b (p :: GHC.Types.Multiplicity). ; (a %p -> b) -> a %p -> b |
| 13:10:54 | <tomsmeding> | __monty__: I think for this module, it's indeed smoke and mirrors, as you say |
| 13:11:06 | <tomsmeding> | as in: you do get the advantage that the array can be used outside of IO or ST |
| 13:11:15 | <tomsmeding> | but indeed the strict linearity is inessential |
| 13:12:41 | <Leary> | __monty__: GHC doesn't do any magic for you, it gives you the tools to create safe interfaces over your own manually allocated/freed data. |
| 13:12:52 | <tomsmeding> | ^ |
| 13:12:56 | × | malte quits (~malte@mal.tc) (Ping timeout: 244 seconds) |
| 13:13:53 | <tomsmeding> | not unlikely that for arrays, allocating the array itself outside of the GC'd heap would be counterproductive, because if it's a boxed array then you'd have to register all the contained pointers as additional GC roots anyway, and it's not boxed then the GC spends only O(1) time on it |
| 13:17:18 | → | jacopovalanzano joins (~jacopoval@cpc151911-cove17-2-0-cust105.3-1.cable.virginm.net) |
| 13:17:21 | <jacopovalanzano> | o/ |
| 13:21:00 | × | paotsaq quits (~paotsaq@127.209.37.188.rev.vodafone.pt) (Ping timeout: 272 seconds) |
| 13:21:56 | → | malte joins (~malte@mal.tc) |
| 13:26:29 | → | agentultra joins (~user@104-195-156-122.cpe.teksavvy.com) |
| 13:29:12 | <agentultra> | i'm trying to compile a static executable with `cabal -v build --enable-static --enable-executable-static` but I am finding that theres a build step that fails as cabal seems to be passing `-dynamic` and `-shared` regardless. anyone have links to a guide on how to get cabal to build statically-linked executables? |
| 13:30:40 | <tomsmeding> | agentultra: https://cs-syd.eu/posts/2024-04-20-static-linking-haskell-nix or https://hasufell.github.io/posts/2024-04-21-static-linking.html perhaps |
| 13:31:11 | <agentultra> | thanks tomsmeding |
| 13:31:21 | <tomsmeding> | agentultra: I'm assuming here that you're talking about statically-linking _system_ dependencies, not haskell dependencies |
| 13:31:39 | <tomsmeding> | haskell dependencies are already linked statically by default, except (essentially) on Arch Linux system GHCs |
| 13:33:17 | <tomsmeding> | (PSA: if you're on arch linux, as always but especially for you, please use ghcup and not the GHC from the system package manager) |
| 13:34:43 | <agentultra> | yes. i'm linking in a shared object file that seems to be compiled in a funny way... when i link dynamically, I get a relink error from `sin` from the shared object. it seems to conflict with GHC's dynamically link with libm. I'm trying to statically link to see if that resolves the issue. |
| 13:35:50 | → | hattckory joins (~hattckory@bras-base-toroon4524w-grc-30-70-27-118-207.dsl.bell.ca) |
| 13:36:08 | <tomsmeding> | agentultra: what exactly is that relink error? |
| 13:36:21 | <tomsmeding> | compiling haskell programs statically is a bit of a rabbit hole, as you can see from those posts |
| 13:37:06 | <agentultra> | It's a bit long: /home/jking/Projects/tigerbeetle-hs/dist-newstyle/build/x86_64-linux/ghc-9.6.6/tigerbeetle-hs-0.1.0.0/x/tigerbeetle-hs/build/tigerbeetle-hs/tigerbeetle-hs: Relink `/nix/store/x3v27x8fz9di2nv2j91xgsypvxgy34g9-libtb_client/lib/libtb_client.so' with `/nix/store/h7zcxabfxa7v5xdna45y2hplj31ncf8a-glibc-2.40-36/lib/libm.so.6' for IFUNC symbol `sin' |
| 13:37:13 | → | ash3en joins (~Thunderbi@185.209.196.192) |
| 13:37:43 | tomsmeding | has never seen that particular error phrasing before |
| 13:37:58 | <tomsmeding> | if you're using nix already anyway, though, getting a static link _may_ be easy enough |
| 13:38:04 | tomsmeding | doesn't use nix |
| 13:38:10 | → | ystael joins (~ystael@user/ystael) |
| 13:38:31 | <agentultra> | i'm learning nix as I go here myself :D |
| 13:38:57 | <agentultra> | but I'm mostly just using it to create the developer shell and compiling like normal with cabal |
| 13:39:01 | <tomsmeding> | my first step in debugging this would be to check how that libtb_client.so refers to 'sin', first by running `nm` on it and if necessary perhaps with objdump |
| 13:39:07 | <tomsmeding> | ah |
| 13:40:18 | × | developer_ quits (~developer@85.50.149.196) (Ping timeout: 252 seconds) |
| 13:40:21 | × | robobub quits (uid248673@id-248673.uxbridge.irccloud.com) (Quit: Connection closed for inactivity) |
| 13:40:54 | × | hattckory quits (~hattckory@bras-base-toroon4524w-grc-30-70-27-118-207.dsl.bell.ca) (Ping timeout: 276 seconds) |
| 13:42:25 | <tomsmeding> | agentultra: bit of a shot in the dark: what happens if you don't put only `libtb_client` in your extra-libraries, but also `m`? |
| 13:43:32 | <tomsmeding> | but also output of `ldd /nix/store/x3v27x8fz9di2nv2j91xgsypvxgy34g9-libtb_client/lib/libtb_client.so` might be helpful |
| 13:43:36 | → | random-jellyfish joins (~developer@user/random-jellyfish) |
| 13:44:45 | <agentultra> | yeah, so the symbol for `sin` from that library is listed as a W,F (Warning/Normal, Function) |
| 13:45:16 | <yushyin> | agentultra: can't hurt to also ask in #haskell:nixos.org (matrix room) in case you suspect a nix-related issue |
| 13:45:21 | <agentultra> | and ldd says it's linked with linux-vdso, libpthread, libc.so.6... but not libm |
| 13:45:49 | <tomsmeding> | agentultra: "W" is weak, not warning |
| 13:45:59 | <agentultra> | ah right, my b |
| 13:46:14 | <tomsmeding> | do try also telling ghc to link with libm |
| 13:46:24 | <agentultra> | a dynamically linked GHC executable does dynamically link with libm |
| 13:46:26 | <tomsmeding> | (as I suggested a few lines above) |
| 13:46:32 | <tomsmeding> | oh |
| 13:46:42 | <agentultra> | ah k, i'll double check that |
| 13:47:15 | → | chele joins (~chele@user/chele) |
| 13:51:02 | → | tromp joins (~textual@2001:1c00:3487:1b00:6095:11f3:6fa5:fb3d) |
| 13:55:38 | → | jespada joins (~jespada@2800:a4:2225:2c00:21b7:d8ab:6242:25ba) |
| 14:01:46 | <agentultra> | yeah, shot in the dark doesn't change anything... the executable is still dynamically linked with libm.so.6, the libtb_client.so file is not. So the "relink" issue, as I understand it, is because that library's symbol table is looking for the implementation of that function which is conflicting with one already defined globally by the executable? |
| 14:02:25 | × | malte quits (~malte@mal.tc) (Ping timeout: 248 seconds) |
| 14:02:31 | <tomsmeding> | I _think_ the error is saying that you need to re-link libtb_client.so with libm.so |
| 14:02:48 | <tomsmeding> | that it to say, it's not an observation that a re-link is happening, it's an instruction to change libtb_client.so |
| 14:03:18 | <tomsmeding> | your "it's a bit long" is also funny because in fact the error was really short: _despite_ very long paths, it still fit in a single IRC message |
| 14:03:38 | <tomsmeding> | which is typical for linker errors, unfortunately: concise to the point of incomprehensibility |
| 14:03:54 | <agentultra> | true! :D |
| 14:04:23 | <tomsmeding> | so my feeling (!) is that there is no conflict at all, and that instead, the symbol 'sin' is _missing_ somehow |
| 14:04:29 | <tomsmeding> | but I don't know enough here |
| 14:05:12 | <agentultra> | it gives me some ideas though, so thanks for sounding off with me |
| 14:05:34 | <tomsmeding> | also try linking that libtb_client.so with a simple C file |
| 14:05:39 | <tomsmeding> | just with plain gcc |
| 14:08:13 | × | ash3en quits (~Thunderbi@185.209.196.192) (Ping timeout: 244 seconds) |
| 14:08:18 | → | malte joins (~malte@mal.tc) |
| 14:09:24 | → | toby-bro joins (~toby-bro@user/toby-bro) |
| 14:14:35 | <hellwolf> | what's the alternative to "replicateM constant" where I can pattern matching. Pattern matching the list is not as convenient, as far as I can tell. |
| 14:15:02 | <hellwolf> | say `replicateN 2 mb >>= \(x1, x2) -> _` |
| 14:17:22 | → | notdabs joins (~Owner@2600:1700:69cf:9000:a99a:9123:20e6:65c6) |
| 14:22:38 | <mauke> | :t replicateM 2 ?mb >>= \[x1, x2] -> ?result |
| 14:22:38 | <lambdabot> | (Monad m, ?mb::m a, ?result::m b) => m b |
| 14:27:29 | <hellwolf> | maybe my skill issue, but I don't like pattermatching these list output when I gave a Int constant as the "size" input. |
| 14:27:42 | <hellwolf> | what's the idiomatic way of doing it? |
| 14:28:09 | <__monty__> | Isn't static the default with GHC, I thought Arch's packaging was always so problematic because they want dynamic everything? |
| 14:29:03 | → | Square joins (~Square@user/square) |
| 14:29:10 | <agentultra> | statically linking the Haskell object files, yes... but we still dynamically link libc, libgmp, etc. |
| 14:29:34 | × | malte quits (~malte@mal.tc) (Ping timeout: 260 seconds) |
| 14:30:24 | × | jrm quits (~jrm@user/jrm) (Quit: ciao) |
| 14:30:26 | <Leary> | hellwolf: `mb >>= \x1 -> mb >>= \x2 -> ...` |
| 14:30:27 | <haskellbridge> | <Morj> I think the tuple replicate would look something like "replicateN @2 mb >>= \(x1 :/\ x2) -> _" |
| 14:30:35 | → | malte joins (~malte@mal.tc) |
| 14:30:47 | <haskellbridge> | <Morj> I vaguely remember seing those tuples somewhere, but that might have been purescript |
| 14:31:57 | → | jrm joins (~jrm@user/jrm) |
| 14:33:18 | <Leary> | You can of course write something like `replicateA :: Applicative f => SNat n -> f a -> f (Vec n a)`, but it's not more idiomatic than the dead-simple approach. |
| 14:34:13 | <haskellbridge> | <Morj> But can ghc already prove that pattern-matching is exhastive here? |
| 14:34:22 | <haskellbridge> | <Morj> With Vec runtimeValue I mean |
| 14:34:29 | × | vpan quits (~vpan@212.117.1.172) (Quit: Leaving.) |
| 14:35:00 | <haskellbridge> | <Morj> Or, ugh, is it even runtime here, I got confused |
| 14:35:07 | <hellwolf> | we have lift2M, we just need liftNM |
| 14:35:55 | <hellwolf> | with "Vec n", we probably can pattern matching, perhaps pattern synonyms would be required. |
| 14:36:16 | × | alp quits (~alp@2001:861:8ca0:4940:97d8:ee2c:4105:6ce6) (Ping timeout: 268 seconds) |
| 14:36:21 | → | paotsaq joins (~paotsaq@127.209.37.188.rev.vodafone.pt) |
| 14:36:29 | <hellwolf> | excuse me, from lift2, to liftN, I meant to say. |
| 14:37:11 | <hellwolf> | but unless it's in the base/core library, it's not idiomatic... |
| 14:37:17 | <hellwolf> | I know how to write my own, but that's not the point. |
| 14:38:38 | <agentultra> | hm, it does dynamically link with gcc. |
| 14:39:16 | <haskellbridge> | <Morj> I sometimes think it would be better if `,` were a right-associative operator instead of built-in construct. This way all (a, b, c) = (a, (b, c)) and we can create our own functions that work with tuples of arbitrary arity |
| 14:39:51 | <haskellbridge> | <Morj> I think ocaml and sml already do that with `*`? |
| 14:39:56 | <agentultra> | gcc doesn't explicitly link `libm` though and we don't get a runtime relink error if we try to execute it. |
| 14:42:39 | <__monty__> | Morj: Are you secretly a LISPer? -_- |
| 14:42:53 | <haskellbridge> | <Morj> I dream of learning lisp one day! |
| 14:43:04 | <hellwolf> | use emacs :) |
| 14:43:15 | <hellwolf> | and have elisp nightmare, instead |
| 14:43:32 | <haskellbridge> | <Morj> I read a blog of a big janet fan, and for some reason it's very hard for me to understand their quoting and unquoting, even when I have no problems using them in template haskell |
| 14:44:01 | <haskellbridge> | <Morj> > emacs |
| 14:44:01 | <haskellbridge> | I've already written 100k lines of vimscript >_< |
| 14:44:30 | × | Googulator quits (~Googulato@81.183.235.203) (Ping timeout: 240 seconds) |
| 14:45:41 | <tomsmeding> | agentultra: so it works with gcc? What if you pass -lm to gcc as well? |
| 14:50:17 | → | prasad joins (~Thunderbi@c-73-246-138-70.hsd1.in.comcast.net) |
| 14:59:28 | → | jmcantrell joins (~weechat@user/jmcantrell) |
| 14:59:44 | × | jespada quits (~jespada@2800:a4:2225:2c00:21b7:d8ab:6242:25ba) (Ping timeout: 260 seconds) |
| 15:03:00 | → | jespada joins (~jespada@2800:a4:231e:8900:903f:fbe:20bf:5608) |
| 15:04:11 | × | jmcantrell quits (~weechat@user/jmcantrell) (Client Quit) |
| 15:04:27 | → | jmcantrell joins (~weechat@user/jmcantrell) |
| 15:05:44 | × | malte quits (~malte@mal.tc) (Ping timeout: 252 seconds) |
| 15:08:47 | → | hattckory joins (~hattckory@bras-base-toroon4524w-grc-30-70-27-118-207.dsl.bell.ca) |
| 15:12:12 | <agentultra> | tomsmeding: oof, it links fine with `-lm` and no runtime relink error :( |
| 15:12:35 | <tomsmeding> | agentultra: you're saying "runtime" again. Does that "Relink" error occur at _runtime_ of your haskell application? |
| 15:12:56 | → | malte joins (~malte@mal.tc) |
| 15:13:24 | <agentultra> | yes |
| 15:13:59 | <tomsmeding> | O.o |
| 15:14:16 | <merijn> | I mean, depending on your definition of runtime that's logical |
| 15:14:17 | <tomsmeding> | what does `ldd` say on your compiled haskell application and on that simple C executable? |
| 15:14:42 | <merijn> | everything be linked statically by default only applies to the Haskell bits |
| 15:14:58 | <tomsmeding> | merijn: I know what a runtime linker is and roughly what it does, but I was running on the wrong assumption here that it was `ld` that was complaining, not `ld-linux.so` |
| 15:15:22 | <merijn> | But yeah, pastebin ldd output |
| 15:15:39 | <tomsmeding> | compile-time linking succeeding and then runtime linking _failing_ sounds like a LD_RUNTIME_PATH mistake somewhere |
| 15:15:45 | <tomsmeding> | and given that this is nix, it smells a lot like nix issues |
| 15:15:51 | <merijn> | oh, oof |
| 15:15:54 | merijn | taps out |
| 15:16:02 | <tomsmeding> | because of the mention of nix? :p |
| 15:16:06 | <merijn> | Yes :p |
| 15:16:31 | <tomsmeding> | *LD_LIBRARY_PATH is what I meant |
| 15:16:51 | <agentultra> | is there a pastebin for this channel? |
| 15:16:54 | × | fp quits (~Thunderbi@130.233.70.95) (Ping timeout: 272 seconds) |
| 15:16:58 | <tomsmeding> | @where paste |
| 15:16:58 | <lambdabot> | Help us help you: please paste full code, input and/or output at e.g. https://paste.tomsmeding.com |
| 15:17:00 | <tomsmeding> | (also see topic) |
| 15:19:07 | <agentultra> | tomsmeding: https://paste.tomsmeding.com/FmcDXN4e |
| 15:19:31 | × | hattckory quits (~hattckory@bras-base-toroon4524w-grc-30-70-27-118-207.dsl.bell.ca) (Read error: Connection reset by peer) |
| 15:20:39 | <tomsmeding> | agentultra: and the former gives the Relink error at runtime and the latter doesn't? |
| 15:20:48 | <agentultra> | tomsmeding: correct |
| 15:20:50 | <tomsmeding> | Do you actually use functions from libtb_client.so in the C program? |
| 15:21:01 | <tomsmeding> | perhaps the runtime linker is lazy and only loads it in once it's actually used |
| 15:21:37 | <agentultra> | no, just linking. I was in the middle of adding a usage, I can get that going and try running it. |
| 15:21:51 | <tomsmeding> | lazy runtime linking is a thing |
| 15:22:37 | <tomsmeding> | (if you manually invoke the runtime linker with dlopen(2), it's the RTLD_LAZY flag) |
| 15:22:53 | <tomsmeding> | er, dlopen(3) |
| 15:24:33 | <agentultra> | odd that the Haskell executable doesn't call anything from the library either, it just links... wonder if something in the generated FFI bindings invokes something in the library? |
| 15:26:33 | <tomsmeding> | agentultra: another thing you can try: set LD_DEBUG=files,libs,bindings (or perhaps LD_DEBUG=all) when starting the two executables and check if there is a difference in the resulting outputs |
| 15:26:42 | <tomsmeding> | *difference between |
| 15:26:54 | <tomsmeding> | (LD_DEBUG is from ld-linux(8)) |
| 15:26:58 | × | ystael quits (~ystael@user/ystael) (Ping timeout: 244 seconds) |
| 15:27:31 | <agentultra> | roger, roger |
| 15:29:04 | → | ystael joins (~ystael@user/ystael) |
| 15:32:52 | × | malte quits (~malte@mal.tc) (Ping timeout: 252 seconds) |
| 15:34:38 | <agentultra> | So calling the library from C, no run-time relink error. |
| 15:34:52 | <agentultra> | Checking the differences |
| 15:35:01 | <merijn> | Is the link invocation for the C code and Haskell the same, though? |
| 15:36:22 | tomsmeding | . o O ( `cabal build --ghc-options=-v` ) |
| 15:36:25 | → | xff0x joins (~xff0x@2405:6580:b080:900:73c3:617e:93dd:971) |
| 15:36:36 | <tomsmeding> | still doesn't print the ld invocation, though |
| 15:37:10 | <tomsmeding> | merijn: do you know of a way to make ghc print the actual ld invocation? `bear intercept` doesn't work because ghc passes options through `@` files so the actual flags are not in the cmdline. |
| 15:37:32 | <merijn> | If you set -v high enough it just prints out everything it does |
| 15:37:42 | <merijn> | Last I tried anyway |
| 15:37:42 | <tomsmeding> | (I still have to hack bear at some point to understand this and open those options files) |
| 15:38:57 | <tomsmeding> | oh wait -v does already show the gcc call, I somehow missed that |
| 15:39:31 | <merijn> | :) |
| 15:39:45 | <tomsmeding> | (gcc and ghc look too much alike) |
| 15:41:04 | → | malte joins (~malte@mal.tc) |
| 15:42:05 | <agentultra> | https://paste.tomsmeding.com/rde825xQ |
| 15:42:44 | <tomsmeding> | agentultra: you need `--ghc-options=-v`, this makes the cabal build verbose but not what ghc itself does |
| 15:42:59 | <agentultra> | tomsmeding: ah, sorry will do. |
| 15:43:14 | <tomsmeding> | we're interested in that "Linking..." line at :589 ;) |
| 15:44:50 | <agentultra> | tomsmeding: https://paste.tomsmeding.com/9QHmZK9i |
| 15:46:13 | × | werneta quits (~werneta@syn-071-083-160-242.res.spectrum.com) (Ping timeout: 245 seconds) |
| 15:46:16 | <tomsmeding> | always nice when ghc links in your library three times |
| 15:46:29 | <tomsmeding> | agentultra: the line in question is :952 there |
| 15:47:03 | <tomsmeding> | no less than 5 occurrences of -lm ! |
| 15:47:34 | <agentultra> | oh weird |
| 15:47:40 | <tomsmeding> | that shouldn't matter though |
| 15:49:08 | <tomsmeding> | what happens if you compile your C file with `gcc thing.c -o a.out -fuse-ld=gold -Wl,--no-as-needed -lm -no-pie -Wl,--gc-sections -ltb_client`? |
| 15:49:28 | <tomsmeding> | (an attempt to manually remove all the superfluous stuff from that long command) |
| 15:50:12 | <tomsmeding> | oh perhaps with `-L/home/jking/Projects/tigerbeetle-hs/lib/x86_64-linux-gnu.2.27` |
| 15:50:39 | <tomsmeding> | somehow the Relink error should be reproducible without ghc :p |
| 15:51:12 | → | Square2 joins (~Square@user/square) |
| 15:52:01 | <agentultra> | tomsmeding: `cc test.c -Wall -Iinclude -Llib/x86_64-linux-gnu.2.27 -fuse-ld=gold -Wl,--no-as-needed -lm -no-pie -Wl,--gc-sections -ltb_client` |
| 15:52:08 | <agentultra> | Creates the relink error when you run it! |
| 15:52:22 | <agentultra> | and a segfault |
| 15:52:27 | <tomsmeding> | lol |
| 15:52:27 | × | tromp quits (~textual@2001:1c00:3487:1b00:6095:11f3:6fa5:fb3d) (Quit: My iMac has gone to sleep. ZZZzzz…) |
| 15:52:29 | × | jacopovalanzano quits (~jacopoval@cpc151911-cove17-2-0-cust105.3-1.cable.virginm.net) (Quit: Client closed) |
| 15:52:52 | <tomsmeding> | now reduce it to figure out what the problem is :) |
| 15:52:53 | → | hattckory joins (~hattckory@70.27.118.207) |
| 15:52:59 | × | malte quits (~malte@mal.tc) (Ping timeout: 260 seconds) |
| 15:53:04 | <agentultra> | yup, excellent. |
| 15:53:05 | <tomsmeding> | also, yay progress |
| 15:58:00 | <agentultra> | yes, thanks a bunch for narrowing that down... I didn't know about the --ghc-options=-v flag |
| 15:58:28 | <merijn> | agentultra: --ghc-options just pass flags as-is on to ghc :) |
| 16:06:08 | <agentultra> | tomsmeding: turns out, it's -lm |
| 16:06:34 | <tomsmeding> | one cannot use libtb_client.so in a program if you also link in libm? |
| 16:06:42 | <tomsmeding> | that... rather sounds like a problem with tb_client. |
| 16:07:04 | <tomsmeding> | or at least how it's built |
| 16:07:21 | <agentultra> | i suspect they statically link libm or include their own implementation |
| 16:07:49 | <tomsmeding> | if they have their own implementation, then internal references to `sin` in tb_client should just refer to the internal symbol, regardless of how many libm's are around |
| 16:07:50 | <agentultra> | the ldd output suggests so... is there a way to tell cabal to tell ghc to not link `-lm` when linking this library? |
| 16:08:09 | <tomsmeding> | how would you then use `sin` from haskell? |
| 16:08:29 | <tomsmeding> | (I don't think so) |
| 16:09:22 | <agentultra> | Hm. So creating a static executable wouldn't even work around it? |
| 16:09:41 | <merijn> | You can just override the linker flags GHC passes |
| 16:10:00 | <tomsmeding> | merijn: GHC already passes -lm 4 times, are you sure you can make it not pass all 4 of thouse? |
| 16:10:02 | <tomsmeding> | *those |
| 16:10:10 | <merijn> | No :p |
| 16:10:18 | <merijn> | Well yes |
| 16:10:32 | <merijn> | But the question is "would I be willing to invest that effort?" :p |
| 16:10:41 | <tomsmeding> | also, I don't think it's a good idea |
| 16:10:55 | <tomsmeding> | e.g. floor(3) is in libm too |
| 16:11:06 | <tomsmeding> | so no more floor() for you, I guess, then? |
| 16:11:36 | <tomsmeding> | (though on a non-ancient CPU, it's probably translated to a CPU instruction) |
| 16:12:04 | × | hattckory quits (~hattckory@70.27.118.207) (Ping timeout: 260 seconds) |
| 16:13:17 | → | tromp joins (~textual@2001:1c00:3487:1b00:6095:11f3:6fa5:fb3d) |
| 16:13:23 | <agentultra> | oh weird... it might be the link order |
| 16:13:34 | <tomsmeding> | link order is significant in general |
| 16:13:54 | <merijn> | tomsmeding: not in general |
| 16:13:58 | <merijn> | But on linux, certainly |
| 16:14:04 | × | weary-traveler quits (~user@user/user363627) (Remote host closed the connection) |
| 16:14:16 | <agentultra> | `$ cc test.c -Wall -Iinclude -Llib/x86_64-linux-gnu.2.27 -ltb_client -lm` compiles and runs with no error |
| 16:14:19 | <tomsmeding> | I meant that as in: there are situations where link order is significant, I'm not sure if this is one of those |
| 16:14:41 | <tomsmeding> | and swapping the -ltb_client and -lm makes it fail? |
| 16:14:42 | <agentultra> | but if you put `-lm` before `-ltb_client`, you get the relink error. |
| 16:14:48 | <tomsmeding> | merijn: any ideas? |
| 16:14:58 | <merijn> | Whisky and crying? |
| 16:15:04 | <tomsmeding> | that's what I suspected |
| 16:15:16 | <merijn> | All I know about linking is against my will :p |
| 16:15:41 | <merijn> | The scary moment you realise that you've become "the person everyone comes to when they hit linker errors" :p |
| 16:15:53 | <tomsmeding> | heh |
| 16:16:15 | <tomsmeding> | that happened to me with "linux stuff" |
| 16:16:16 | <agentultra> | yeah, linkers are pain. |
| 16:16:58 | <merijn> | And I'm too distracted trying to learn tree-sitter to think about linkers |
| 16:17:08 | <tomsmeding> | agentultra: any chance you can do, well, precisely what the error is saying -- change the tb_client build process to include -lm when building the shared library? |
| 16:17:08 | <agentultra> | wonder if this would be ameliorated if ghc had it's own |
| 16:17:44 | <agentultra> | tomsmeding: i might try that unless there's an easy way to pass the link flags in the desired order to ghc from cabal |
| 16:18:23 | <tomsmeding> | I strongly suspect ghc just unconditionally adds -lm early on the command line, from looking at the structure of that command line in the log you sent |
| 16:18:49 | <tomsmeding> | which is overall not a very strange thing to do, tb_client should just declare that it depends on -lm, and it doesn't |
| 16:21:22 | × | lortabac quits (~lortabac@2a01:e0a:541:b8f0:55ab:e185:7f81:54a4) (Quit: WeeChat 4.5.2) |
| 16:22:14 | <EvanR> | tomsmeding, I can hit you up for any help with "linux stuff" ? cool |
| 16:22:50 | tomsmeding | hides |
| 16:23:19 | yushyin | makes a note |
| 16:23:48 | tomsmeding | despairs |
| 16:25:18 | <agentultra> | tomsmeding: agree, I think it's because the link order.. the executable has `sin` first in the elf of the executable and `dlopen` loads this library with it's own `sin` and tries to relink it? So if we change the link order when we call the library it can resolve it's own `sin`. |
| 16:25:34 | × | jespada quits (~jespada@2800:a4:231e:8900:903f:fbe:20bf:5608) (Quit: My Mac has gone to sleep. ZZZzzz…) |
| 16:26:36 | <agentultra> | I'm going to see if they'll accept a patch from me to to change their linker options, if that's even possible. The library is actually written in Zig :S |
| 16:27:59 | <tomsmeding> | agentultra: I think this is what's happening (but take this with a grain of salt): the `sin` symbol is Weak in libtb_client.so (as we saw), which means (man 1 nm) that it's allowed to be undefined at link time. When compiling with -ltb_client -lm, `sin` is first undefined, and then libm provides a definition. When compiling with -lm -tb_client -lm, the first -lm resolves any missing references to |
| 16:28:01 | <tomsmeding> | `sin`, but then afterwards, tb_client introduces new ones. The second occurrence of -lm (this is where I'm a bit unsure) is ignored because libm was already specified (?) |
| 16:28:20 | <tomsmeding> | hence in the second case, `sin` remains unresolved, which results in an error at runtime |
| 16:28:34 | <tomsmeding> | dlopen is not involved here, that is for manually calling into the runtime linker at runtime |
| 16:28:55 | <EvanR> | it is indeed difficult to resolve one's own sin |
| 16:28:56 | <agentultra> | gotcha... that seems reasonable |
| 16:29:08 | <tomsmeding> | it's ld-linux.so, the runtime linker itself, that is relevant here (which is, despite its name, actually an executable) |
| 16:29:28 | <tomsmeding> | EvanR: it's not at all, just declare that you depend on libm |
| 16:29:40 | <EvanR> | the good news about libm |
| 16:29:52 | <agentultra> | well, they'll have a bug report coming soon to get the process started. |
| 16:29:57 | <tomsmeding> | :) |
| 16:30:14 | <agentultra> | thanks again for all your help :D |
| 16:30:20 | <EvanR> | (actually, sin has unlimited radius of convergence so there's that) |
| 16:30:23 | <tomsmeding> | cheers :) |
| 16:30:37 | <tomsmeding> | EvanR: floating-point numbers only go so far though |
| 16:30:37 | <jle`> | that's deep |
| 16:31:19 | <EvanR> | it's true "correct rounded result" guarantee doesn't cover the entire domain of sin in ieee754 |
| 16:31:27 | <EvanR> | watch yourself |
| 16:31:31 | <EvanR> | in the far lands |
| 16:34:05 | × | prasad quits (~Thunderbi@c-73-246-138-70.hsd1.in.comcast.net) (Read error: Connection reset by peer) |
| 16:38:11 | → | malte joins (~malte@mal.tc) |
| 16:41:48 | × | chele quits (~chele@user/chele) (Remote host closed the connection) |
| 16:44:23 | → | prasad joins (~Thunderbi@c-73-246-138-70.hsd1.in.comcast.net) |
| 16:46:54 | <monochrom> | "You must not sin." >:) |
| 16:51:17 | → | pyooque joins (~puke@user/puke) |
| 16:51:18 | × | puke quits (~puke@user/puke) (Killed (erbium.libera.chat (Nickname regained by services))) |
| 16:51:18 | pyooque | is now known as puke |
| 16:51:19 | × | nitrix quits (~nitrix@user/meow/nitrix) (Ping timeout: 260 seconds) |
| 16:51:40 | <monochrom> | But I thought you would want to absolve, not resolve, your sin. >:) |
| 16:55:04 | × | acidjnk quits (~acidjnk@p200300d6e71c4f71c835c1b6e3010b6c.dip0.t-ipconnect.de) (Ping timeout: 272 seconds) |
| 16:58:09 | → | nitrix joins (~nitrix@user/meow/nitrix) |
| 16:58:20 | × | L29Ah quits (~L29Ah@wikipedia/L29Ah) (Ping timeout: 265 seconds) |
| 17:06:00 | → | jespada joins (~jespada@2800:a4:231e:8900:903f:fbe:20bf:5608) |
| 17:06:07 | <int-e> | :t abs |
| 17:06:09 | <lambdabot> | Num a => a -> a |
| 17:06:17 | <int-e> | :t abs . sin |
| 17:06:18 | <lambdabot> | Floating c => c -> c |
| 17:12:53 | → | acidjnk joins (~acidjnk@p200300d6e71c4f71c835c1b6e3010b6c.dip0.t-ipconnect.de) |
| 17:17:35 | → | Sgeo joins (~Sgeo@user/sgeo) |
| 17:22:00 | → | weary-traveler joins (~user@user/user363627) |
| 17:22:39 | <EvanR> | :t fix . error |
| 17:22:40 | <lambdabot> | [Char] -> c |
| 17:23:02 | <absence> | Is there a way to do something like "(\f g h a -> f (g (h a))) <$> f' <*> g' <*> h' <*> a'" without having to manually specify the combining function to the left? |
| 17:23:47 | <merijn> | I can't tell if that's trying to be idiom brackets or something else |
| 17:23:54 | <EvanR> | are you sure you don't want to reverse the applicative or something |
| 17:24:14 | <merijn> | absence: What's it trying to accomplish? |
| 17:24:24 | → | OlzhasYergali joins (~OlzhasYer@188.130.156.6) |
| 17:29:41 | → | L29Ah joins (~L29Ah@wikipedia/L29Ah) |
| 17:29:52 | × | tromp quits (~textual@2001:1c00:3487:1b00:6095:11f3:6fa5:fb3d) (Quit: My iMac has gone to sleep. ZZZzzz…) |
| 17:30:39 | × | euphores quits (~SASL_euph@user/euphores) (Quit: Leaving.) |
| 17:30:59 | <absence> | merijn: f' g' and h' are applicative functors that return some operation (e.g. "pure (+1)"), so I just want to apply them. |
| 17:31:28 | <merijn> | absence: okay, so how about we simplify this |
| 17:31:38 | <merijn> | ah, wait |
| 17:31:50 | <merijn> | absence: Are they all Endo functors? |
| 17:31:56 | <merijn> | i.e. "a -> a" functions? |
| 17:32:17 | <absence> | No, otherwise I guess it would be easier... |
| 17:32:44 | <merijn> | yeah, I was thinking just traverse + some fold |
| 17:35:41 | → | euphores joins (~SASL_euph@user/euphores) |
| 17:38:51 | <absence> | Hm, I guess it's kind of an applicative version of (.)? |
| 17:39:25 | → | tromp joins (~textual@2001:1c00:3487:1b00:6095:11f3:6fa5:fb3d) |
| 17:39:39 | <merijn> | absence: I mean, you could just define that? :) |
| 17:39:56 | × | ubert1 quits (~Thunderbi@2a02:8109:ab8a:5a00:7d86:d0e8:c2d1:2ee3) (Remote host closed the connection) |
| 17:40:03 | <haskellbridge> | <Bowuigi> Would some fold/traversal with liftA2 work? |
| 17:40:19 | <merijn> | Bowuigi: No, because that only works for Endo |
| 17:41:02 | <merijn> | absence: You an just define <.> as custom operator? :) |
| 17:41:24 | → | alp joins (~alp@2001:861:8ca0:4940:10d7:35fd:1caf:385) |
| 17:41:48 | <merijn> | :t let f <.> g = (.) <$> f <*> g in (<.>) |
| 17:41:49 | <lambdabot> | Applicative f => f (b -> c) -> f (a -> b) -> f (a -> c) |
| 17:42:13 | <merijn> | Maybe reverse the arguments depending on how you want it to line up |
| 17:42:27 | → | target_i joins (~target_i@user/target-i/x-6023099) |
| 17:45:01 | × | OlzhasYergali quits (~OlzhasYer@188.130.156.6) (Quit: Client closed) |
| 17:45:15 | <absence> | Ugh, "of course" I'm using impredicative types, so that fails in the same way as the regular (.) does, but other than that I guess it would've worked. |
| 17:46:02 | <merijn> | absence: Can't you fix it with explicit RankN type annotations? |
| 17:48:07 | × | lxsameer quits (~lxsameer@Serene/lxsameer) (Ping timeout: 265 seconds) |
| 17:50:16 | <absence> | Is it possible to define a generic (.) that works with RankN, or would it have to be a more specific one for the structures I have? |
| 17:52:08 | <monochrom> | I think (.) already works with RankN. It is why the lens library has no type error. |
| 17:53:42 | <c_wraith> | Eh. The lens library uses ALens (or whatever) in a lot of contexts where it'd otherwise need RankN |
| 17:55:16 | <monochrom> | Does "(\f g h a -> f (g (h a))) <$> f' <*> g' <*> h' <*> a'" simply mean f' <*> (g' <*> (h' <*> a')) ? |
| 17:56:36 | × | aetepe quits (~aetepe@188.119.58.34) (Ping timeout: 252 seconds) |
| 17:56:48 | <monochrom> | Come to think of it, I probably once used Applicative laws to prove something related. |
| 18:00:15 | × | malte quits (~malte@mal.tc) (Remote host closed the connection) |
| 18:02:35 | × | tromp quits (~textual@2001:1c00:3487:1b00:6095:11f3:6fa5:fb3d) (Quit: My iMac has gone to sleep. ZZZzzz…) |
| 18:05:18 | <absence> | You may be right, but I run into the same RankN problem: Couldn't match type: forall a. m a -> n a with: b -> c |
| 18:09:35 | → | malte joins (~malte@mal.tc) |
| 18:18:30 | → | tromp joins (~textual@2001:1c00:3487:1b00:6095:11f3:6fa5:fb3d) |
| 18:23:05 | → | aetepe joins (~aetepe@188.119.58.34) |
| 18:23:16 | → | Googulator joins (~Googulato@2a01-036d-0106-01d5-ac5d-24c1-ad5e-7f2b.pool6.digikabel.hu) |
| 18:31:12 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 18:31:13 | × | acidjnk quits (~acidjnk@p200300d6e71c4f71c835c1b6e3010b6c.dip0.t-ipconnect.de) (Ping timeout: 245 seconds) |
| 18:39:16 | × | euphores quits (~SASL_euph@user/euphores) (Remote host closed the connection) |
| 18:40:09 | → | euphores joins (~SASL_euph@user/euphores) |
| 18:44:10 | → | slack1256 joins (~slack1256@2803:c600:5111:9ab8:84f4:ae24:2907:edab) |
| 18:44:48 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 18:57:34 | <tomsmeding> | absence: what about simply "f' <$> g' <$> h' <$> a'"? |
| 18:57:40 | <tomsmeding> | % :i <$> |
| 18:57:40 | <yahb2> | (<$>) :: Functor f => (a -> b) -> f a -> f b ; -- Defined in ‘GHC.Internal.Data.Functor’ ; infixl 4 <$> |
| 19:00:01 | × | caconym quits (~caconym@user/caconym) (Quit: bye) |
| 19:00:03 | <tomsmeding> | fun exercise: prove that "(x <$> y) <$> z" and "x <$> (y <$> z)" do exactly the same thin |
| 19:00:05 | <tomsmeding> | +g |
| 19:00:15 | → | wildsalander joins (~wildsalan@37-33-20-77.bb.dnainternet.fi) |
| 19:00:44 | → | caconym joins (~caconym@user/caconym) |
| 19:00:48 | × | zungi quits (~tory@user/andrewchawk) (Ping timeout: 264 seconds) |
| 19:00:54 | × | wildsalander quits (~wildsalan@37-33-20-77.bb.dnainternet.fi) (Client Quit) |
| 19:01:23 | → | acidjnk joins (~acidjnk@p200300d6e71c4f71c835c1b6e3010b6c.dip0.t-ipconnect.de) |
| 19:01:32 | <tomsmeding> | monochrom: no, (<$>) not (<*>) |
| 19:02:37 | <tomsmeding> | (hint for my "fun exercise": first infer the types of x, y and z in both cases) |
| 19:03:54 | × | malte quits (~malte@mal.tc) (Ping timeout: 268 seconds) |
| 19:04:11 | → | malte joins (~malte@mal.tc) |
| 19:04:18 | → | rvalue- joins (~rvalue@user/rvalue) |
| 19:04:33 | <haskellbridge> | <Liamzee> tomsmeding: no, the problem is that it's keyed so that Prelude.take 3 and Prelude.Linear.take3 do separate things |
| 19:04:45 | × | rvalue quits (~rvalue@user/rvalue) (Ping timeout: 246 seconds) |
| 19:05:24 | <tomsmeding> | Liamzee: "keyed"? |
| 19:05:35 | <haskellbridge> | <Liamzee> namely, Linear.take 3 takes 4 elements of the list |
| 19:05:47 | <tomsmeding> | _what_ |
| 19:06:06 | <tomsmeding> | surely that's a bug |
| 19:06:28 | <haskellbridge> | <Liamzee> <0, as opposed to <= 0 |
| 19:06:35 | <tomsmeding> | that makes 0 sense |
| 19:06:48 | × | sord937 quits (~sord937@gateway/tor-sasl/sord937) (Quit: sord937) |
| 19:08:08 | <tomsmeding> | Liamzee: any more of such functions? |
| 19:08:24 | × | malte quits (~malte@mal.tc) (Ping timeout: 244 seconds) |
| 19:09:20 | <haskellbridge> | <Liamzee> only noticed that one |
| 19:09:23 | <haskellbridge> | <Liamzee> you want to push the patch? |
| 19:09:27 | <tomsmeding> | drop also has it |
| 19:09:30 | <[exa]> | O_o |
| 19:09:38 | <haskellbridge> | <Liamzee> yeah, i saw |
| 19:09:46 | <haskellbridge> | <Liamzee> that's a breaking change to API, though |
| 19:09:59 | <tomsmeding> | that's why it'll just be an issue, not a PR yet |
| 19:10:03 | <haskellbridge> | <Liamzee> I wonder who's actually using it if drop / take being bugged like this didn't get caught |
| 19:10:08 | <tomsmeding> | exactly |
| 19:10:17 | <tomsmeding> | I think Data.List.Linear is not what people are using prelude-linear for |
| 19:10:41 | rvalue- | is now known as rvalue |
| 19:10:41 | <tomsmeding> | Data.Array.Mutable.Linear is much more useful |
| 19:12:21 | <[exa]> | anyway yeah there's a literal off-by-one error here right? https://hackage.haskell.org/package/linear-base-0.4.0/docs/src/Data.List.Linear.html#take |
| 19:12:26 | <absence> | tomsmeding: That doesn't compile, since f' isn't a function... |
| 19:13:13 | <tomsmeding> | absence: well, you used it as a function in your original snippet. |
| 19:13:39 | <tomsmeding> | Liamzee: https://github.com/tweag/linear-base/issues/484 |
| 19:14:01 | <[exa]> | wow, I wanted to check the issue tracker if the bug is there already |
| 19:14:06 | <[exa]> | 2 issues "opened just now" |
| 19:14:08 | <[exa]> | good job guys. :D |
| 19:14:24 | <tomsmeding> | absence: the only difference between the type of what I posted, and the type of what you posted, is that mine has a Functor constraint only |
| 19:14:30 | <tomsmeding> | oh oops |
| 19:14:52 | <tomsmeding> | Liamzee: shall I close mine? |
| 19:15:02 | <absence> | tomsmeding: The type of f' is something like f (a -> b). |
| 19:15:18 | <tomsmeding> | [exa]: what's cooler is that my issue has a lower number, but it's higher up in the list |
| 19:15:41 | <haskellbridge> | <Liamzee> closed mine |
| 19:15:44 | <tomsmeding> | ah I see |
| 19:15:44 | <[exa]> | yeah I'm just looking at that |
| 19:15:48 | × | tomku quits (~tomku@user/tomku) (Ping timeout: 252 seconds) |
| 19:16:07 | <tomsmeding> | absence: _oops_, sorry |
| 19:16:50 | × | hgolden quits (~hgolden@2603:8000:9d00:3ed1:6ff3:8389:b901:6363) (Remote host closed the connection) |
| 19:17:49 | → | tomku joins (~tomku@user/tomku) |
| 19:20:33 | → | hgolden joins (~hgolden@2603:8000:9d00:3ed1:6ff3:8389:b901:6363) |
| 19:25:45 | <[exa]> | tomsmeding Liamzee: you guys clicked it literally at the very same second, and although Liamzee has a little tiny bit higher global ID it still sorts by something different (I guess they sort by opening time in seconds (equal) and then probably by title or something) |
| 19:26:40 | <tomsmeding> | cool |
| 19:27:51 | → | sprotte24 joins (~sprotte24@p200300d16f1244002c79175af1485053.dip0.t-ipconnect.de) |
| 19:28:43 | × | AlexZenon quits (~alzenon@178.34.150.194) (Ping timeout: 245 seconds) |
| 19:28:59 | → | zungi joins (~tory@user/andrewchawk) |
| 19:29:31 | → | polyphem joins (~rod@p3ee3f49b.dip0.t-ipconnect.de) |
| 19:32:56 | → | pavonia joins (~user@user/siracusa) |
| 19:33:41 | × | alp quits (~alp@2001:861:8ca0:4940:10d7:35fd:1caf:385) (Ping timeout: 248 seconds) |
| 19:35:54 | → | AlexZenon joins (~alzenon@178.34.150.194) |
| 19:41:09 | × | remedan quits (~remedan@ip-62-245-108-153.bb.vodafone.cz) (Quit: Bye!) |
| 19:41:39 | → | ash3en joins (~Thunderbi@185.209.196.192) |
| 19:42:10 | → | remedan joins (~remedan@ip-62-245-108-153.bb.vodafone.cz) |
| 19:44:07 | <EvanR> | 1 second is a very long time in computers |
| 19:44:23 | <EvanR> | two events within the same 1 second interval might not even be able to see each other xD |
| 19:45:26 | → | Tuplanolla joins (~Tuplanoll@91-159-69-59.elisa-laajakaista.fi) |
| 19:45:30 | × | AlexZenon quits (~alzenon@178.34.150.194) (Ping timeout: 252 seconds) |
| 19:45:52 | × | ash3en quits (~Thunderbi@185.209.196.192) (Ping timeout: 252 seconds) |
| 19:48:56 | <[exa]> | is there any good/recommended way to have a feedable attoparsec parser which can output partial data between the feeds, and has to keep some state information for the whole stream? |
| 19:49:30 | <merijn> | [exa]: Yes, it's called "attoparsec"? |
| 19:50:34 | <merijn> | [exa]: It has built in methods for incremental feeding. So you just wrap it in a tuple with some extra state? |
| 19:51:14 | <[exa]> | merijn: like, I'm doing the lifting manually now (I've got a `parseSomething :: st -> Parser (st, [something])`, feeding it and passing the `st` around manually when one piece finishes |
| 19:52:06 | <[exa]> | I hoped for something where I could have a `tell`, `put` and `get` and no other thinking to do :D |
| 19:52:24 | tomsmeding | . o O ( `tell`? ) |
| 19:52:29 | <merijn> | [exa]: I mean, you can just use `StateT` on top of attoparsec? |
| 19:52:56 | <merijn> | [exa]: If you wanna get fancier than that, you'll have to consider swapping to megaparsec |
| 19:53:00 | <[exa]> | yes, but then I have to pass the state around that myself anyway |
| 19:53:06 | merijn | .o O ( is trifecta still a thing? ) |
| 19:53:15 | → | AlexZenon joins (~alzenon@178.34.150.194) |
| 19:53:25 | <[exa]> | tomsmeding: the tell from writer, "we parsed out this information, emit it" |
| 19:53:27 | → | ljdarj1 joins (~Thunderbi@user/ljdarj) |
| 19:53:28 | <merijn> | [exa]: megaparsec has an embedded state parser |
| 19:53:42 | <merijn> | s/state parser/state monad |
| 19:53:52 | <[exa]> | hm let me check |
| 19:54:04 | <tomsmeding> | [exa]: that sounds not like how parser combinator parsers work :p |
| 19:54:11 | <tomsmeding> | you return the parsed value, you don't `tell` it |
| 19:54:38 | <tomsmeding> | wanting to output the parsed values via `tell` sounds like you want a non-Parser that optionally uses Parser inside to parse little chunks |
| 19:55:02 | <[exa]> | yeah this is for something like RDF turtle, I was thinking about emitting the triples directly without structure. |
| 19:55:20 | <[exa]> | sounds kinda like I want the actual streamy libraries, right. |
| 19:55:41 | <tomsmeding> | [exa]: what if the Parser backtracks? Do you keep your writer state? (Flashback to a few days ago) |
| 19:55:54 | <merijn> | [exa]: You probably want streaming libraries then, yeah |
| 19:56:11 | <[exa]> | tomsmeding: good point, it should hold until it's clear that it won't backtrack |
| 19:56:21 | <merijn> | [exa]: conduit works really well with the various parsecs where you just write a parser for one record and emit as you finish them |
| 19:56:27 | <[exa]> | ok this actually at least proves that my idea with writer is off |
| 19:56:39 | <absence> | What's the best way to deal with forall mismatches after simplified subsumption? I just made some code compile by writing "fmap (\f x -> f x)", which looks a bit silly, and isn't very intuitive... |
| 19:56:41 | × | ljdarj quits (~Thunderbi@user/ljdarj) (Ping timeout: 265 seconds) |
| 19:56:41 | ljdarj1 | is now known as ljdarj |
| 19:56:50 | <[exa]> | merijn: yeah I think that will be it |
| 19:56:52 | × | Square2 quits (~Square@user/square) (Ping timeout: 252 seconds) |
| 19:56:52 | × | Square quits (~Square@user/square) (Ping timeout: 252 seconds) |
| 19:57:01 | <tomsmeding> | the standard fix for simplified subsumption breaking your code is indeed eta-expansion |
| 19:57:17 | <[exa]> | ok thanks guys for consultation again! :) |
| 19:57:57 | × | AlexZenon quits (~alzenon@178.34.150.194) (Ping timeout: 252 seconds) |
| 19:58:10 | <absence> | tomsmeding: So I've heard, it just feels extra weird when it's inside a functor. But I guess there's no other way? |
| 19:59:21 | → | CiaoSen joins (~Jura@2a02:8071:64e1:da0:5a47:caff:fe78:33db) |
| 20:00:36 | → | AlexZenon joins (~alzenon@178.34.150.194) |
| 20:07:38 | × | slack1256 quits (~slack1256@2803:c600:5111:9ab8:84f4:ae24:2907:edab) (Remote host closed the connection) |
| 20:08:55 | → | machinedgod joins (~machinedg@d108-173-18-100.abhsia.telus.net) |
| 20:09:29 | → | j1n37- joins (~j1n37@user/j1n37) |
| 20:10:14 | × | j1n37 quits (~j1n37@user/j1n37) (Ping timeout: 260 seconds) |
| 20:10:48 | × | aetepe quits (~aetepe@188.119.58.34) (Remote host closed the connection) |
| 20:11:29 | × | dhil quits (~dhil@2a0c:b381:52e:3600:cbb4:807:319f:c7af) (Ping timeout: 248 seconds) |
| 20:12:48 | × | zungi quits (~tory@user/andrewchawk) (Ping timeout: 264 seconds) |
| 20:13:18 | × | CiaoSen quits (~Jura@2a02:8071:64e1:da0:5a47:caff:fe78:33db) (Ping timeout: 272 seconds) |
| 20:18:12 | <haskellbridge> | <Liamzee> hmmm, that's interesting |
| 20:18:13 | <haskellbridge> | <Liamzee> Ur isn't a monad? |
| 20:21:16 | <EvanR> | instance Prelude.Monad Ur where |
| 20:21:26 | <EvanR> | Ur a >>= f = f a |
| 20:25:55 | × | polyphem quits (~rod@p3ee3f49b.dip0.t-ipconnect.de) (Ping timeout: 268 seconds) |
| 20:27:30 | <haskellbridge> | <Liamzee> i'm not sure what the technical term is, i spoke to a friend who did stuff with Linear Haskell |
| 20:27:49 | <haskellbridge> | <Liamzee> but we talked about the possibility of Linear Haskell simply being a concern for library writers, but you called LH for better memory usage or a disabled GC |
| 20:28:14 | <haskellbridge> | <Liamzee> having Ur being a monad would be useful for LInear Haskell, so it's interesting to ask why it's not |
| 20:28:24 | <haskellbridge> | <Liamzee> "Rust as a monadic eDSL" ;) |
| 20:30:34 | <EvanR> | I'm concerned about your premise that Ur is not a moand |
| 20:30:45 | <EvanR> | which seems to be plainly false |
| 20:31:16 | <EvanR> | disabling of the GC when using linear haskell doesn't sound like a thing |
| 20:31:42 | <haskellbridge> | <Liamzee> https://hackage.haskell.org/package/linear-base-0.1.0/docs/Control-Functor-Linear.html |
| 20:31:51 | × | weary-traveler quits (~user@user/user363627) (Quit: Konversation terminated!) |
| 20:31:56 | <haskellbridge> | <Liamzee> oh, i realize the problem now |
| 20:32:04 | <haskellbridge> | <Liamzee> it's up to 0.4.0, and maybe 0.5.0 :) |
| 20:32:13 | → | weary-traveler joins (~user@user/user363627) |
| 20:32:56 | <haskellbridge> | <Liamzee> https://hackage.haskell.org/package/linear-base-0.4.0/docs/src/Data.Unrestricted.Linear.Internal.Ur.html#line-75 |
| 20:35:58 | <haskellbridge> | <Liamzee> EvanR: but while doing the computation, you could potentially do free after use, but i'm not sure if the allocator works that way |
| 20:36:31 | <EvanR> | what do you mean |
| 20:37:30 | <EvanR> | free what |
| 20:39:21 | <haskellbridge> | <Liamzee> i mean linear types imply that a value is consumed, meaning that its allocation can be cleared |
| 20:39:53 | <haskellbridge> | <Liamzee> you might have a GC for the nonlinear haskell wrapping the computation, but the linear computation doesn't need a GC |
| 20:40:27 | <EvanR> | the meaning of "consumed" doesn't necessarily mean "deleted from existence" |
| 20:40:39 | <EvanR> | it just means you can't use it again |
| 20:41:02 | <haskellbridge> | <Liamzee> "not sure if the allocator works that way" |
| 20:41:08 | <EvanR> | in one example, that space was immediately reused by the operation you just did |
| 20:41:13 | <EvanR> | freeing it would break things in that case |
| 20:41:29 | <EvanR> | in another case, the object's ownership transferred elsewhere |
| 20:41:38 | <EvanR> | in another case, it wasn't a real object the begin with |
| 20:41:59 | <EvanR> | e.g. it could have been evidence of some sort that's now invalid |
| 20:43:10 | <EvanR> | linear anything is often much more general in scope |
| 20:43:47 | <haskellbridge> | <Liamzee> at least in theory, though, it should have better performance due to not needing to do as many allocations |
| 20:43:56 | <haskellbridge> | <Liamzee> and not needing the GC to be on to cover the data |
| 20:44:09 | <EvanR> | calling into a memory management system to "free" stuff constantly sounds like overhead of its own |
| 20:44:43 | <EvanR> | that's why Ur/Web waits until the end of the processing program to discard the entire block of memory |
| 20:45:34 | <EvanR> | it just sounds like you're conflating "no GC memory management" with linear types |
| 20:49:57 | × | L29Ah quits (~L29Ah@wikipedia/L29Ah) (Ping timeout: 248 seconds) |
| 20:50:15 | <haskellbridge> | <Liamzee> Web type? |
| 20:51:42 | <haskellbridge> | <Liamzee> i'm just being conservative because i was told that there's no automatic mutation implied by linear haskell |
| 20:51:48 | <haskellbridge> | <Liamzee> which is a feature that some people want |
| 20:55:10 | <EvanR> | "automatic mutation" ? |
| 20:55:15 | <merijn> | Liamzee: FYI, GHC's allocator is *stupidly* fast, though |
| 20:55:24 | <merijn> | EvanR: In place updates when reference count == 1 |
| 20:55:32 | × | jespada quits (~jespada@2800:a4:231e:8900:903f:fbe:20bf:5608) (Quit: My Mac has gone to sleep. ZZZzzz…) |
| 20:56:07 | <merijn> | Like, I'd be surprised if allocating memory took more than 5 instructions in GHC |
| 20:57:18 | <merijn> | It's basically "increment a number" and a comparison to check if you exceeded the heap |
| 20:57:29 | <merijn> | Maybe 1 or 2 instructions for an atomic CAS? |
| 20:57:31 | <EvanR> | allocation is fast and easy |
| 20:57:45 | → | jespada joins (~jespada@2800:a4:231e:8900:903f:fbe:20bf:5608) |
| 20:57:45 | × | jespada quits (~jespada@2800:a4:231e:8900:903f:fbe:20bf:5608) (Client Quit) |
| 20:58:42 | <merijn> | Liamzee: FWIW, I've had Haskell programs with allocation rates of over 4 Gb/s (and this is the average over a duration of 2-3 minutes) that never exceeded more than a few 10s of MB of resident memory |
| 20:59:14 | <tomsmeding> | one cute trick you can do with LinearTypes that I haven't seen much is `newtype State s a = Staet (s %1-> (Ur a, s))` |
| 20:59:38 | <tomsmeding> | you get `modify :: (s %1-> s) -> State s a -> State s a`, and it otherwise works essentially exactly like the normal state monad |
| 20:59:42 | <haskellbridge> | <Liamzee> one cute trick you can do with LinearTypes that I haven't seen much is using LinearTypes ;) |
| 21:00:14 | <EvanR> | napkin math says that implies you "run out of resident memory" 400 times a second |
| 21:00:15 | <tomsmeding> | so if you're using a State monad that you would like to do mutable updates with, it's actually very easy to achieve that now |
| 21:01:42 | <haskellbridge> | <Liamzee> I guess, I'll ask directly, is there something wrong with using LinearTypes as something wrapped over by non-linear Haskell to provide a low-cost performance gain without needing FFI? |
| 21:01:48 | <merijn> | EvanR: It was a program streaming a few GB worth of data from SQLite via conduit, it actually worked remarkably well in keeping low RES memory for the data size |
| 21:02:01 | × | kimiamania quits (~65804703@user/kimiamania) (Quit: PegeLinux) |
| 21:02:04 | <tomsmeding> | Liamzee: https://hackage.haskell.org/package/text-builder-linear for example |
| 21:02:24 | → | kimiamania joins (~65804703@user/kimiamania) |
| 21:02:29 | × | AlexZenon quits (~alzenon@178.34.150.194) (Ping timeout: 252 seconds) |
| 21:02:53 | <haskellbridge> | <Liamzee> yeah, i saw, I recently did a purview of the LH ecosystem :) |
| 21:03:28 | → | AlexZenon joins (~alzenon@178.34.150.194) |
| 21:03:34 | <haskellbridge> | <Liamzee> going full embedded with LH, at this time, is probably not a good idea, but developing wrapped linear Haskell libraries is probably a good transition to build up skills within the community and build up support for the LinearTypes extension |
| 21:04:56 | <haskellbridge> | <Liamzee> erm, not purview, review |
| 21:06:29 | → | malte joins (~malte@mal.tc) |
| 21:09:20 | <EvanR> | merijn, tens of megabytes sounds pretty good practically speaking |
| 21:09:44 | <EvanR> | otoh that would exhaust the first hard drive I had, much less the ram that went with it |
| 21:11:10 | <merijn> | EvanR: I mean, the raw data was a database of, like, 8 GB and 2 billion rows. That had a bunch of joins blowing up the data queried even more, then doing a full scan of that data. 10s of megabytes was much less than I was prepared to use :p |
| 21:11:35 | <merijn> | Considering the machine running the code had, like, 192 GB RAM :p |
| 21:15:01 | → | Unicorn_Princess joins (~Unicorn_P@user/Unicorn-Princess/x-3540542) |
| 21:20:27 | × | malte quits (~malte@mal.tc) (Remote host closed the connection) |
| 21:20:28 | × | bsima quits (~bsima@143.198.118.179) (Quit: ZNC 1.8.2 - https://znc.in) |
| 21:21:14 | → | bsima joins (~bsima@2604:a880:400:d0::19f1:7001) |
| 21:21:58 | × | nitrix quits (~nitrix@user/meow/nitrix) (Quit: ZNC 1.9.1 - https://znc.in) |
| 21:23:54 | → | nitrix joins (~nitrix@user/meow/nitrix) |
| 21:23:59 | → | malte joins (~malte@mal.tc) |
| 21:29:59 | × | wootehfoot quits (~wootehfoo@user/wootehfoot) (Read error: Connection reset by peer) |
| 21:40:39 | × | malte quits (~malte@mal.tc) (Remote host closed the connection) |
| 21:42:27 | → | L29Ah joins (~L29Ah@wikipedia/L29Ah) |
| 21:46:30 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 21:47:30 | × | j1n37- quits (~j1n37@user/j1n37) (Read error: Connection reset by peer) |
| 21:53:01 | → | j1n37 joins (~j1n37@user/j1n37) |
| 21:58:14 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 21:59:52 | × | chiselfuse quits (~chiselfus@user/chiselfuse) (Remote host closed the connection) |
| 22:00:26 | → | chiselfuse joins (~chiselfus@user/chiselfuse) |
| 22:00:47 | → | malte joins (~malte@mal.tc) |
| 22:05:12 | × | target_i quits (~target_i@user/target-i/x-6023099) (Quit: leaving) |
| 22:07:27 | × | malte quits (~malte@mal.tc) (Remote host closed the connection) |
| 22:07:28 | × | takuan quits (~takuan@d8D86B601.access.telenet.be) (Remote host closed the connection) |
| 22:08:51 | × | tromp quits (~textual@2001:1c00:3487:1b00:6095:11f3:6fa5:fb3d) (Ping timeout: 246 seconds) |
| 22:12:44 | × | toby-bro quits (~toby-bro@user/toby-bro) (Ping timeout: 260 seconds) |
| 22:17:45 | × | michalz quits (~michalz@185.246.207.205) (Remote host closed the connection) |
| 22:44:52 | → | emmanuelux joins (~emmanuelu@user/emmanuelux) |
| 22:50:11 | → | dudek joins (~dudek@2a02:a312:c9df:bf80:6431:ea:b097:7aad) |
| 23:02:46 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 23:03:50 | × | __monty__ quits (~toonn@user/toonn) (Quit: leaving) |
| 23:06:49 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 23:11:17 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 248 seconds) |
| 23:22:48 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 23:27:46 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds) |
| 23:31:10 | × | dudek quits (~dudek@2a02:a312:c9df:bf80:6431:ea:b097:7aad) (Quit: Leaving) |
| 23:31:50 | → | malte joins (~malte@mal.tc) |
| 23:34:40 | × | preflex quits (~preflex@user/mauke/bot/preflex) (Ping timeout: 252 seconds) |
| 23:34:55 | → | sintheta joins (~src@user/src) |
| 23:34:56 | → | preflex joins (~preflex@user/mauke/bot/preflex) |
| 23:35:24 | × | src quits (~src@user/src) (Ping timeout: 252 seconds) |
| 23:37:14 | × | tdammers quits (~tdammers@110-136-178-143.ftth.glasoperator.nl) (Ping timeout: 252 seconds) |
| 23:37:37 | → | tdammers joins (~tdammers@110-136-178-143.ftth.glasoperator.nl) |
| 23:38:51 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 23:41:38 | × | sudden quits (~cat@user/sudden) (Ping timeout: 252 seconds) |
| 23:43:37 | → | sudden joins (~cat@user/sudden) |
| 23:43:43 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 245 seconds) |
| 23:44:08 | → | merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl) |
| 23:51:54 | × | merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds) |
| 23:57:15 | → | ljdarj1 joins (~Thunderbi@user/ljdarj) |
All times are in UTC on 2025-03-28.