2025/03/18

Newest at the top

2025-03-18 19:31:52 +0100Digit(~user@user/digit) (Ping timeout: 272 seconds)
2025-03-18 19:30:14 +0100Digitteknohippie(~user@user/digit) Digit
2025-03-18 19:27:48 +0100 <EvanR> maybe the wiki has the latest
2025-03-18 19:27:26 +0100biberu(~biberu@user/biberu) biberu
2025-03-18 19:27:06 +0100biberu(~biberu@user/biberu) (Quit: ZNC - https://znc.in)
2025-03-18 19:26:26 +0100 <EvanR> is it up to date
2025-03-18 19:26:20 +0100 <EvanR> geekosaur, yeah I've read it. Then I was like, what does GHC actually do now?
2025-03-18 19:23:22 +0100kh0d(~kh0d@212.200.181.151)
2025-03-18 19:21:59 +0100hattckory(~hattckory@bras-base-toroon4524w-grc-47-184-146-98-182.dsl.bell.ca) (Ping timeout: 260 seconds)
2025-03-18 19:18:16 +0100julie_pilgrim(~julie_pil@user/julie-pilgrim/x-1240752) julie_pilgrim
2025-03-18 19:17:21 +0100machinedgod(~machinedg@d108-173-18-100.abhsia.telus.net) machinedgod
2025-03-18 19:17:17 +0100sprotte24(~sprotte24@p200300d16f2b0f00f146771413e22502.dip0.t-ipconnect.de) (Read error: Connection reset by peer)
2025-03-18 19:12:42 +0100kh0d(~kh0d@109.111.226.14) (Remote host closed the connection)
2025-03-18 19:10:30 +0100alfiee(~alfiee@user/alfiee) (Ping timeout: 260 seconds)
2025-03-18 19:08:13 +0100jespada(~jespada@2800:a4:22a4:bf00:c8e8:ae3a:868c:376c) (Quit: My Mac has gone to sleep. ZZZzzz…)
2025-03-18 19:08:09 +0100lxsameer(~lxsameer@Serene/lxsameer) (Ping timeout: 244 seconds)
2025-03-18 19:06:01 +0100alfiee(~alfiee@user/alfiee) alfiee
2025-03-18 19:03:30 +0100 <haskellbridge> <Bowuigi> Also tangentially related but Perceus seems to be okay-ish without extensions https://www.reddit.com/r/ProgrammingLanguages/comments/1anyq4j/can_reference_counting_really_be_as… extensions make it good tho https://lptk.github.io/files/ctrc-2024-05-09.pdf
2025-03-18 19:02:50 +0100hattckory(~hattckory@bras-base-toroon4524w-grc-47-184-146-98-182.dsl.bell.ca)
2025-03-18 19:00:36 +0100 <geekosaur> see the "making a fast curry" paper re eval/apply vs. push/enter
2025-03-18 19:00:08 +0100 <haskellbridge> <Bowuigi> Specially if parts of it were generated with single-purpose optimizers, like https://www.fing.edu.uy/inco/proyectos/fusion/
2025-03-18 18:58:41 +0100krei-se(~krei-se@p508296bc.dip0.t-ipconnect.de) (Ping timeout: 252 seconds)
2025-03-18 18:58:28 +0100krei-se-(~krei-se@p3ee0f216.dip0.t-ipconnect.de) krei-se
2025-03-18 18:58:10 +0100peterbecich(~Thunderbi@syn-047-229-123-186.res.spectrum.com) (Ping timeout: 248 seconds)
2025-03-18 18:57:43 +0100 <haskellbridge> <Bowuigi> First order STLC doesn't look too different from various IRs like QBE or LLVM IR
2025-03-18 18:57:01 +0100 <haskellbridge> <Bowuigi> A random thought that happened to cross my mind based on the paper by Kovacs I linked
2025-03-18 18:56:54 +0100hattckory(~hattckory@bras-base-toroon4524w-grc-47-184-146-98-182.dsl.bell.ca) (Ping timeout: 276 seconds)
2025-03-18 18:56:48 +0100 <EvanR> making "what is GHC actually doing now" more of a question
2025-03-18 18:56:34 +0100 <EvanR> I'm getting the impression there's alot of new ideas too, not necessarily implemented
2025-03-18 18:55:51 +0100 <EvanR> re: magic on first order STLC, is this your original research or an actual thing happening
2025-03-18 18:55:17 +0100 <haskellbridge> <Bowuigi> I mentioned that because of the alternative compilation route via a first order language
2025-03-18 18:54:43 +0100 <monochrom> I don't think GHC does defunctionalization. It goes eval-apply or push-enter (I forgot which one is older and which one is current).
2025-03-18 18:53:58 +0100 <haskellbridge> <Bowuigi> Haven't seen books on full functional language compiler implementations that appeared after the 2000s
2025-03-18 18:53:13 +0100 <haskellbridge> <Bowuigi> Yeah those are really old
2025-03-18 18:52:54 +0100 <monochrom> or s/paper/papers, books, etc./
2025-03-18 18:52:28 +0100 <monochrom> I think EvanR has already seen the STG paper. But it doesn't cover any development since then (and there are a lot).
2025-03-18 18:51:23 +0100 <haskellbridge> <Bowuigi> To translate a language into first order, you need defunctionalization. Here's the best one I know https://dl.acm.org/doi/10.1145/3591260
2025-03-18 18:50:12 +0100krei-se(~krei-se@p508296bc.dip0.t-ipconnect.de) krei-se
2025-03-18 18:50:04 +0100 <haskellbridge> <Bowuigi> Nowadays, probably doing some magic on a first order STLC can get you nice results tho, namely, some superset of the object language of https://andraskovacs.github.io/pdfs/2ltt_icfp24.pdf should be simple enough to translate to assembly
2025-03-18 18:48:05 +0100 <haskellbridge> <Bowuigi> And https://www.microsoft.com/en-us/research/publication/implementing-lazy-functional-languages-on-sto… is a good resource on STG
2025-03-18 18:47:35 +0100hattckory(~hattckory@bras-base-toroon4524w-grc-47-184-146-98-182.dsl.bell.ca)
2025-03-18 18:46:53 +0100 <haskellbridge> <Bowuigi> The compiler itself is a good reference of Core https://hackage.haskell.org/package/ghc-9.6.1/docs/GHC-Core.html, which links to https://www.microsoft.com/en-us/research/publication/system-f-with-type-equality-coercions/
2025-03-18 18:46:42 +0100wildsalander(~wildsalan@81-175-155-97.bb.dnainternet.fi) (Quit: wildsalander)
2025-03-18 18:45:18 +0100 <haskellbridge> <Bowuigi> https://dl.acm.org/doi/10.1145/3674654 deals with unboxing and some machine stuff
2025-03-18 18:44:12 +0100 <haskellbridge> <Bowuigi> EvanR the GHC wisdom is scattered over the world but there are new papers that implement some of those concepts
2025-03-18 18:39:24 +0100krei-se(~krei-se@p3ee0f57d.dip0.t-ipconnect.de) (Ping timeout: 260 seconds)
2025-03-18 18:35:03 +0100kuribas(~user@ip-188-118-57-242.reverse.destiny.be) (Quit: ERC 5.5.0.29.1 (IRC client for GNU Emacs 29.3))
2025-03-18 18:32:14 +0100cipherrot(~znc-user@user/petrichor) (Ping timeout: 268 seconds)
2025-03-18 18:31:19 +0100euleritian(~euleritia@95.90.214.149)
2025-03-18 18:31:01 +0100euleritian(~euleritia@ip5f5ad695.dynamic.kabel-deutschland.de) (Ping timeout: 248 seconds)