Newest at the top
| 2025-01-13 18:00:47 +0100 | <haskellbridge> | <magic_rb> r-sta Thats llvm-hs i think which is not really functional anymore :( |
| 2025-01-13 18:00:31 +0100 | <geekosaur> | that's the IR. you can carry it to another system with no ghc but has LLVM and run opt and llc there |
| 2025-01-13 18:00:10 +0100 | <r-sta> | https://www.hcesperer.org/posts/2017-07-28-writing-a-small-llvm-compiler-frontend-in-haskell.html |
| 2025-01-13 17:59:31 +0100 | <haskellbridge> | <magic_rb> I can always do a healthy amount of Text concatenations |
| 2025-01-13 17:58:14 +0100 | <r-sta> | fair |
| 2025-01-13 17:57:42 +0100 | <haskellbridge> | <magic_rb> No i meant writing my own codegen in haskell for my own shitty little language |
| 2025-01-13 17:57:24 +0100 | <r-sta> | nothing about llvm IR tho |
| 2025-01-13 17:57:18 +0100 | <r-sta> | and a file called main.ll is created. |
| 2025-01-13 17:57:17 +0100 | <r-sta> | ghc -keep-llvm-files main.hs |
| 2025-01-13 17:56:34 +0100 | <r-sta> | https://stackoverflow.com/questions/52040651/compile-haskell-programs-to-llvm-ir |
| 2025-01-13 17:56:33 +0100 | <r-sta> | i can find this; |
| 2025-01-13 17:56:28 +0100 | machinedgod | (~machinedg@d108-173-18-100.abhsia.telus.net) (Ping timeout: 252 seconds) |
| 2025-01-13 17:55:50 +0100 | <haskellbridge> | <magic_rb> My other option is fixing llvm-hs |
| 2025-01-13 17:55:37 +0100 | <haskellbridge> | <magic_rb> Was just about to say GHC does that |
| 2025-01-13 17:55:24 +0100 | <geekosaur> | (LLVM doesn't want people using IR directly. never mind that ghc does, they pay the price for it) |
| 2025-01-13 17:55:14 +0100 | <haskellbridge> | <magic_rb> Youre not :) |
| 2025-01-13 17:54:58 +0100 | <geekosaur> | I thought you weren't supposed to do that |
| 2025-01-13 17:54:53 +0100 | <r-sta> | sounds like a parsecy sort of thing maybe? |
| 2025-01-13 17:54:23 +0100 | <haskellbridge> | <magic_rb> Ive taking a course on compiler construction, unfortunately in python so im thinking if i can rewrite it haskell. Do we have a package which would allow me to spit out textual llvm IR easily? |
| 2025-01-13 17:53:58 +0100 | <r-sta> | 7* |
| 2025-01-13 17:53:47 +0100 | <r-sta> | its got code i wrote like 5 years ago |
| 2025-01-13 17:53:41 +0100 | <r-sta> | its the bigest project ever! |
| 2025-01-13 17:53:12 +0100 | <r-sta> | ill see if it works |
| 2025-01-13 17:53:08 +0100 | <r-sta> | i mean, repl should do that anyway right? but the run call would have maybe left *something* extra... |
| 2025-01-13 17:52:41 +0100 | <r-sta> | but maybe it will stop short of main.exe |
| 2025-01-13 17:52:30 +0100 | acidjnk_new | (~acidjnk@p200300d6e7283f35081469c6fc5c461d.dip0.t-ipconnect.de) (Ping timeout: 246 seconds) |
| 2025-01-13 17:52:28 +0100 | <r-sta> | it is doing that |
| 2025-01-13 17:52:04 +0100 | <geekosaur> | well, `cabal repl` may just compile everything again. ghci's the one being annoying here with its preference for compiled stuff |
| 2025-01-13 17:51:42 +0100 | <r-sta> | im having to deal with like 10 min compilation times, its a pain |
| 2025-01-13 17:50:50 +0100 | <r-sta> | ok so i just delete dist-newstyle |
| 2025-01-13 17:50:41 +0100 | <geekosaur> | you need to force it to load the source files instead |
| 2025-01-13 17:50:28 +0100 | <geekosaur> | that's the point, you used v2-run which compiled it, now ghci is insisting on using the compiled one |
| 2025-01-13 17:49:29 +0100 | <r-sta> | how is there one which is not compiled!? |
| 2025-01-13 17:49:00 +0100 | <geekosaur> | but the compiled one can't be debugged |
| 2025-01-13 17:48:51 +0100 | <geekosaur> | if it finds a compiled module it uses it, unless you prefix with an asterisk |
| 2025-01-13 17:47:37 +0100 | <r-sta> | i guess im missing a cabal flag |
| 2025-01-13 17:47:25 +0100 | <r-sta> | No modules are loaded with debugging support. |
| 2025-01-13 17:47:12 +0100 | <r-sta> | but if i try to write :break 2 at the ghci prompt, it says; |
| 2025-01-13 17:46:05 +0100 | <r-sta> | thats changing `cabal v2-run' to `cabal v2-repl' |
| 2025-01-13 17:45:45 +0100 | <r-sta> | so i think its *trying* to break on exception |
| 2025-01-13 17:45:32 +0100 | <r-sta> | Attempting to reconstruct a stack trace... |
| 2025-01-13 17:45:31 +0100 | <r-sta> | Access violation in generated code when executing data at 0x7ff687e1c1a0 |
| 2025-01-13 17:45:30 +0100 | <r-sta> | now it gives |
| 2025-01-13 17:45:27 +0100 | <r-sta> | hmm |
| 2025-01-13 17:42:58 +0100 | <geekosaur> | there's nothing to beak to in a compiled program |
| 2025-01-13 17:42:40 +0100 | <r-sta> | ok, i understand this i think |
| 2025-01-13 17:42:26 +0100 | <geekosaur> | yes, ghci only |
| 2025-01-13 17:42:19 +0100 | r-sta | (~r-sta@sgyl-37-b2-v4wan-168528-cust2421.vm6.cable.virginm.net) |
| 2025-01-13 17:41:28 +0100 | <geekosaur> | `-fbreak-on-exception` only works in ghci, I think? but `v2-run` compiles |
| 2025-01-13 17:38:00 +0100 | r-sta | (~r-sta@sgyl-37-b2-v4wan-168528-cust2421.vm6.cable.virginm.net) (Client Quit) |