[GuillaumeLagrange]

Fixes the two release jobs that broke the 5.7.0 release, and makes every PR run the release build so the next such failure shows up before a release.

• aarch64 musl failed to link: libbpf calls gcc's outline-atomic helpers (__aarch64_ldadd4_sync). The -lgcc from target.aarch64-unknown-linux-musl.rustflags is dropped because dist sets RUSTFLAGS. The musl C code is now built with -mno-outline-atomics, so nothing is needed at link time. Only 3 atomics in libbpf change, all at attach time. Linking libgcc from memtrack's build script does not work: rustc places it before libbpf.
• macOS never started: dist 0.31.0 picks macos-14, which GitHub no longer runs. The release now builds on macos-latest, like CI. dist 0.34.0 defaults to macos-15; the override stays so a retired image cannot break a release again.
• CI missed both: musl-build used plain cargo build. It is replaced by pr-run-mode = "upload", which runs the dist build of every target on PRs without publishing. create-draft-release is skipped on PRs.
• cargo-dist 0.31.0 → 0.34.0: regenerated release.yml and re-applied its two hand edits. Reusable workflows are now called with the $/ self-repository syntax. The action pins in dist-workspace.toml were behind the ones dependabot set in release.yml, so regenerating would have downgraded them; they are synced, and the config now lists the hand edits to redo after a regen.

The first three commits come from #619, minus its release commit.

Refs COD-4123

[codspeed]

Merging this PR will not alter performance

2 warnings

⚠️ Walltime results may be unreliable

Some benchmarks used the Walltime instrument outside CodSpeed Macro Runners, which can make the results inconsistent.

Learn more about Macro Runners

See affected benchmarks

⚠️ 8 benchmarks are dominated by system calls

System calls cannot be simulated consistently and are left out of the measure. The reported value understates the real cost. Switch to the Walltime instrument to measure them.

Learn more about system calls

See affected benchmarks

✅ 53 untouched benchmarks
⏩ 6 skipped benchmarks¹

---

_{Comparing fix-release-and-macos-build (3a2bbac) with main (21b0cda)}

Footnotes

1. 6 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[greptile-apps]

Confidence Score: 5/5

[Medium impact] The PR appears safe to merge; no actionable defects were found.

Summary

This PR fixes the aarch64 musl link settings and moves macOS release builds to macos-latest.

• The ARM release builds use working compiler and runner settings.
• Pull requests now build release artifacts without publishing them.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  PR[Pull request] --> Plan[Plan release artifacts]
  Tag[Version tag] --> Plan
  Plan --> Build[Build target artifacts]
  PR --> Skip[Skip draft release creation]
  Tag --> Draft[Create draft release]
  Build --> Gate{Publishing a tag?}
  Gate -->|No| Stop[Keep artifacts on workflow run]
  Gate -->|Yes| Publish[Publish release artifacts]

Loading

_{Reviews (1) · Last reviewed commit: "ci: bump cargo-dist to 0.34.0" · Reviewed by Greptile}