| 11 Dec 2025 |
alexfmpe | iopq: https://github.com/commercialhaskell/stack/issues/6824 | 21:14:02 |
andromeda | I need to link the wayland headers at build time so the header library stops freaking out about not seeing things but at the same time dynamic linking has problems on nix :( | 21:30:52 |
andromeda | I'll just not do wayland fn, it's beyond my ability | 21:31:11 |
| 12 Dec 2025 |
iopq | yep that's the one | 06:06:19 |
iopq | https://github.com/jtt9340/aoc2025/commit/e5fdca266f62c6ac0145f17f1ba60802dfaf1698 is the solution | 08:42:37 |
alexfmpe | been getting this sort of thing lately
WARNING: /nix/store/5sfpjqy5rbipa3mdgiwal8k7zqfq5jxh-ghc-9.6.7/lib/ghc-9.6.7/bin/./ghc-9.6.7 is loading libcrypto in an unsafe way when building inside a shell like
nix-shell -p cabal-install ghcid haskell-language-server ghc openssl zlib pkg-config anyone's seen this? (on unstable) | 16:37:09 |
alexfmpe | * been getting this sort of thing lately
WARNING: /nix/store/5sfpjqy5rbipa3mdgiwal8k7zqfq5jxh-ghc-9.6.7/lib/ghc-9.6.7/bin/./ghc-9.6.7 is loading libcrypto in an unsafe way when building inside a shell like
nix-shell -p cabal-install ghcid haskell-language-server haskell.packages.ghc96.ghc openssl zlib pkg-config anyone's seen this? (on unstable) | 16:37:40 |
alexfmpe | I see this in .cabal
extra-libraries: crypto | 16:38:37 |
emily | I bet related to that dlopen change, maybe. | 16:39:06 |
emily | some discussion https://github.com/cl-plus-ssl/cl-plus-ssl/issues/114 | 16:39:17 |
emily | I assume this is macOS. | 16:39:20 |
alexfmpe | oops yeah aarch64-darwin | 16:40:51 |
alexfmpe | m1 specifically | 16:40:58 |
alexfmpe | hmm so the issue is likely to be from loading unversioned libcrypto? | 16:53:02 |
emily | my guess is that it's something related to https://gitlab.haskell.org/ghc/ghc/-/merge_requests/12264 / https://gitlab.haskell.org/ghc/ghc/-/merge_requests/14290 / blah | 16:54:06 |
emily | it might still be loading libcrypto wrong (we'd want that to be an absolute path for Nix presumably?) | 16:54:20 |
emily | but I'd guess this stuff is what exposed it | 16:54:28 |
alexfmpe | I'm getting this on ghc 9.6 though | 16:58:02 |
emily | ah. then I have no idea. git bisect --first-parent? :) | 16:58:53 |
alexfmpe | I mean, 9.8 too, just meant it seems to predate those merge request | 16:58:56 |
emily | right | 17:00:41 |
emily | I'm just assuming you didn't use to get this | 17:00:47 |
alexfmpe | not sure, first time I try to build this particular project | 17:01:09 |
alexfmpe | hmm hang on | 17:01:15 |
alexfmpe | I think changing this to crypto.3 fixed it | 17:01:53 |
alexfmpe | can't confirm right now because I did a clean afterwards and now github is dying on me or something | 17:02:11 |
emily | I feel like it should work with just crypto and we're holding it wrong or something. but maybe not. | 17:04:00 |
emily | have you tried 25.05, 24.11? | 17:04:08 |
alexfmpe | I'll see if I can get away with this stupid workaround just to complete the task I was aiming for, then circle back to investigate this properly | 17:10:17 |
emily | fair enough | 17:11:27 |