[staging-next] Darwin stdenvBootstrap fixes 2025-10#448523
[staging-next] Darwin stdenvBootstrap fixes 2025-10#448523vcunat merged 2 commits intoNixOS:staging-nextfrom
Conversation
Fix this config name for consistency with non-bootstrap platform config. This also fixes a build failure of the bootstrap tools on aarch64-darwin, where a test in LLVM 21 assumes it's spelt arm64, not aarch64. Also saves on having to rebuild LLVM just for bootstrap tools, now that the config is consistent.
nixpkgs-ci
bot
added
10.rebuild-darwin: 0
dramforever
force-pushed
the
darwin-bootstrap-fixes-2025-10
branch
from
cd13af0 to
8bfd3c8
Compare
This comment was marked as outdated.
This comment was marked as outdated.
Since NixOS#420382, we set an absolute default CLANG_RESOURCE_DIR for LLVM versions >= 21, which breaks the bootstrap tools because the absolute path was removed by nuke-ref while building and isn't there while bootstrapping. Pass -resource-dir explicitly while testing Clang to make it work.
dramforever
force-pushed
the
darwin-bootstrap-fixes-2025-10
branch
from
8bfd3c8 to
b0ff7e6
Compare
| @@ -22,7 +22,7 @@ | |||
| supportedSystems ? builtins.fromJSON (builtins.readFile ../../ci/supportedSystems.json), | |||
| # The platform triples for which we build bootstrap tools. | |||
| bootstrapConfigs ? [ | |||
| "aarch64-apple-darwin" | |||
| "arm64-apple-darwin" | |||
There was a problem hiding this comment.
has the triple we use for aarch64 darwin been inconsistent between pkgs vs bootstrap for ages? :c
aarch64-darwin's pkgs.hostPlatform.config does give arm64-apple-darwin
There was a problem hiding this comment.
Yes. This has finally come to bite us.
There was a problem hiding this comment.
So I've ended up centralizing what we know here #444862 (comment)
There was a problem hiding this comment.
It seems like we're using darwinArch to turn aarch64 -> arm64 in some places, maybe we can clean those up in a followup if they were only needed due to this inconsistency.
darwinArch = cpu: if cpu.name == "aarch64" then "arm64" else cpu.name;
|
cc @reckenrode: This will require fixes in darwin bootstrap next time the bootstrap tarballs are refreshed. Unfortunately I have no idea how. |
|
Thanks. I think we'd better unblock now on Hydra. But feel free to improve further, anyone. |
…ple-darwin Follow-up to NixOS#448523
3 participants
Note: I haven't tested these.Also, I don't know what needs to be done for the--resource-dirthing to not run into it while bootstrapping Darwin stdenv next time.For the
{aarch64 => arm64}-apple-darwinrename, it seems to avoid rebuilding LLVM at all, which looks promising.I haven't tested the--resource-dirthing at all.Edit:
stdenvBootstrapTools.{x86_64,arm64}-apple-darwin.{build,test}worksThings done
passthru.tests.nixpkgs-reviewon this PR. See nixpkgs-review usage../result/bin/.Add a 👍 reaction to pull requests you find important.