| 21 Sep 2025 |
 emily | how does it fail anyway? | 21:32:38 |
 Grimmauld (any/all) | i have no clue yet lol | 21:36:15 |
| Grimmauld (any/all) | i mean it fails, it expects to hit a limit and just doesn't | 21:36:32 |
| emily | so it succeeds unexpectedly? | 21:36:44 |
| Grimmauld (any/all) | libxml changed how it counts xml objects, so either their reproducer is now too small or libxml broke something | 21:36:51 |
| emily | that's a bit concerning if the limit is there for security | 21:36:58 |
| Grimmauld (any/all) | if libxml broke something, then at least arch will have billion laughs once again lol | 21:37:10 |
| emily | I see | 21:37:01 |
| Grimmauld (any/all) | (did i say i find the libxml-native testsuite quite lacking?) | 21:37:41 |
| Grimmauld (any/all) | there is a reason why my goal was to build towards libreoffice | 21:38:11 |
| Grimmauld (any/all) | the open document formats are basically xml, and pull all the fun svg and html stuff in deps | 21:38:29 |
| Grimmauld (any/all) | LO has actually good tests, but building towards it sucks | 21:39:00 |
| Grimmauld (any/all) | i mean i also had to patch like three other things. Two of those were const-ness issues, where fields either became const or lost their const. Thats dumb api breakage for very little reason... | 21:42:04 |
| Grimmauld (any/all) | truly not a fun update | 21:42:14 |
 ElvishJerricco | https://github.com/NixOS/nixpkgs/pull/444946
I only just realized that since it only affects musl it probably doesn't need to target staging, is that right Alyssa Ross? (But if it does need to, it probably should be staging, not staging-next at this point?)
| 22:45:15 |
| 22 Sep 2025 |
 Alyssa Ross | Yeah that's fine for master | 04:56:04 |
 K900 | So uh | 05:23:41 |
| K900 | Did anyone get the dotnets | 05:23:45 |
| K900 | I think I'm down to that and like three other things now | 05:24:06 |
| K900 | And I have a new candidate for biggest commit with cmake 4 fixes hidden in it: https://github.com/SoftFever/OrcaSlicer/commit/883607e1d4a0b2bb719f2f4bcd9fd72f8c2174fa | 05:32:58 |
| K900 | Oh god I found the dotnet expression | 05:46:52 |
 Winter | condolences | 05:48:49 |
| K900 | OK I think I figured out how to shove the patch in there | 06:21:54 |
| K900 | OK I think this actually does work | 06:38:54 |
 ofalvai | To be honest, I was building off a few days old staging and encountered the failure there. But now I see the failure in Hydra as well: https://hydra.nixos.org/build/308144224 | 06:42:30 |
 Vladimír Čunát | The error doesn't happen on x86_64-linux, so that can be confusing. | 06:47:02 |
| K900 | /build/dotnet-9.0.110/eng/detect-binaries.sh: line 135: 5091 Segmentation fault /build/dotnet-9.0.110/.dotnet/dotnet run --project /build/dotnet-9.0.110/eng/tools/BinaryToolKit -c Release --property:RestoreSources=/nix/store/pir670rnvffbwf5jigw5zgm5c442z8px-dotnet-stage0-sdk-9.0.110-artifacts clean /build/dotnet-9.0.110 -o /build/dotnet-9.0.110/artifacts/log/binary-report -ab /build/dotnet-9.0.110/eng/allowed-sb-binaries.txt -l Debug -p CustomPackageVersionsProps=/nix/store/pir670rnvffbwf5jigw5zgm5c442z8px-dotnet-stage0-sdk-9.0.110-artifacts/PackageVersions.props
| 07:01:53 |
| K900 | So uhhh | 07:01:55 |
| K900 | Maybe we pin LLVM 20 lol | 07:01:59 |
| K900 | I'm going to try building it with LLVM 20 pinned | 07:11:56 |
