| 10 Jun 2025 |
 teo (they/he) | In reply to @maralorn:maralorn.de
teo (they/he): I sadly didn’t get to look at your test PR during zurihac had too much on my plate. It is high on my list and I will look at it shortly. No worries! There's no particular rush | 09:38:19 |
 chreekat | In reply to @alex:tunstall.xyz
https://github.com/NixOS/nixpkgs/pull/414063 Fwiw I think this happens in the upstream GHC CI sometimes, too. Well, something like it, I guess. | 09:38:23 |
| chreekat | Maybe I'm thinking of https://grafana.gitlab.haskell.org/d/167r9v6nk/ci-spurious-failures?orgId=2&from=now-30d&to=now&var-types=ghc-config_file_conflict&var-types=hadrian_locked_file&var-runners=$__all&var-jobs=$__all&refresh=15m&timezone=browser , which looks like a race condition in hadrian (?) | 09:40:48 |
| chreekat | Error when running Shake build system:
at want, called at src/Main.hs:127:44 in main:Main
* Depends on: test:all_deps
at apply1, called at src/Development/Shake/Internal/Rules/Rerun.hs:41:5 in shake-0.19.8-5dc581e9d19dda2753fdbf2d3206f1e10a496c54d77902ed1f75a44d52207a78:Development.Shake.Internal.Rules.Rerun
* Depends on: test:ghc
at need, called at src/Rules/Test.hs:325:3 in main:Rules.Test
* Depends on: _build/stage1/lib/package.conf.d/integer-gmp-1.1-inplace.conf
at need, called at src/Rules/Register.hs:149:5 in main:Rules.Register
* Depends on: _build/stage1/libraries/integer-gmp/build/stamp-integer-gmp-1.1-inplace_dyn
* Raised the exception:
_build/stage1/libraries/integer-gmp/build/stamp-integer-gmp-1.1-inplace_dyn: withFile: resource busy (file is locked)
| 09:41:21 |
 Alex | In reply to @b:chreekat.net
Error when running Shake build system:
at want, called at src/Main.hs:127:44 in main:Main
* Depends on: test:all_deps
at apply1, called at src/Development/Shake/Internal/Rules/Rerun.hs:41:5 in shake-0.19.8-5dc581e9d19dda2753fdbf2d3206f1e10a496c54d77902ed1f75a44d52207a78:Development.Shake.Internal.Rules.Rerun
* Depends on: test:ghc
at need, called at src/Rules/Test.hs:325:3 in main:Rules.Test
* Depends on: _build/stage1/lib/package.conf.d/integer-gmp-1.1-inplace.conf
at need, called at src/Rules/Register.hs:149:5 in main:Rules.Register
* Depends on: _build/stage1/libraries/integer-gmp/build/stamp-integer-gmp-1.1-inplace_dyn
* Raised the exception:
_build/stage1/libraries/integer-gmp/build/stamp-integer-gmp-1.1-inplace_dyn: withFile: resource busy (file is locked)
This is in the make install when using the Make build script in 9.4.
Unless Hadrian copied or reused the problematic code, it shouldn't be affected by the same bug. | 13:59:37 |
| chreekat | Yeah | 14:40:18 |
 maralorn | alexfmpe: https://hydra.nixos.org/build/299137964/nixlog/1 do you understand this gtk stuff? This is keeping me from using hls 2.11 in reflex-plattform. I think this means that one of our system dependencies changed in a way that the generated Haskell does not typecheck anymore at usage site? | 14:57:06 |
| maralorn | More precisely it is keeping me from bumping r-p to a current haskell-updates commit which I wanna do to get hls 2.11 which has code actions again. | 14:57:50 |
 alexfmpe | Yeah whatever gi-soup generates bindings to must have changed | 15:12:51 |
| alexfmpe | haven't figured it out yet, nor could I understand how to rerun the stuff in https://github.com/haskell-gi/haskell-gi to try and reproduce upstream | 15:13:36 |
| alexfmpe | https://discourse.haskell.org/t/build-failing-for-ghcjs-project-the-symbol-jscheckscriptsyntax-does-not-fit-here/12185/12?u=alexfmpe | 15:14:34 |
| alexfmpe | in the meantime you should be able to workaround by using jsaddle-warp for reflex-dom | 15:14:35 |
| maralorn | Uh, probably an excellent idea anyway. | 15:27:18 |
| maralorn | Buuut, I guess we shouldn’t do that in nixpkgs, should we? 😄 | 15:30:38 |
| maralorn | Huh, I wonder what is the correct way to detect that we are targeting js. ghc.isGhcjs does not exist anymore. | 15:37:58 |
| alexfmpe | stdenv.hostPlatform.isGhcjs | 15:39:08 |
| alexfmpe | but in the reflex-platform case, I thought warp was the default jsaddle | 15:39:57 |
| alexfmpe | https://github.com/reflex-frp/reflex-platform/blob/2e0d31bc1b565eed02a3848cde0e0dbc4f33a2d4/project/default.nix#L120 | 15:39:58 |
| alexfmpe | er, or rather, in r-p there's a nix setting for it | 15:40:11 |
| alexfmpe | and in obelisk it was made the default | 15:40:16 |
| maralorn | Nope | 15:40:37 |
| maralorn | stdenv.targetPlatform.isJavascript | 15:40:55 |
| alexfmpe | what | 15:41:08 |
| alexfmpe | oh are you checking on the compiler or on the package? | 15:41:26 |
| alexfmpe | because for the package it's https://github.com/NixOS/nixpkgs/blob/50015e7b86465fd5bd10d79ffe9567b4483d6279/pkgs/development/haskell-modules/generic-builder.nix#L24 | 15:41:33 |
| maralorn | on the compiler | 15:41:37 |
| maralorn | Ah, I see. | 15:41:45 |
| alexfmpe | then it's shifted | 15:42:08 |
| alexfmpe | so there it's targetPlatform.isGhcjs | 15:42:15 |
| alexfmpe | because the compiler targets the platform the package it compiles is ran/hosted on | 15:43:27 |
