[christiangnrd]

See JuliaGPU/GPUCompiler.jl#1042

Turns out we probably wouldn't even be testing macOS 1.10 anymore here anyways, so I added the macos and julia 1.10 jobs to test out. We should keep at most 1 of them before merging

[christiangnrd]

@maleadt this greatly increases the runtime of the macos tests, but I think it’s worth it to free up the metal runners on buildkite.

Do we want to keep one 1.10 run or should we stick to the 1.13 runs, and one of the 1.13 runs took 2h5 minutes, do you think a 2.5h timeout is reasonable?

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.30%. Comparing base (0739da8) to head (1bd6dba).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #578   +/-   ##
=======================================
  Coverage   89.30%   89.30%           
=======================================
  Files          18       18           
  Lines        1748     1748           
=======================================
  Hits         1561     1561           
  Misses        187      187           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[christiangnrd]

Still broken on 1.10? https://github.com/JuliaGPU/OpenCL.jl/actions/runs/39816117372?pr=578

Once the job times out, I'll make the buildkite job 1.10 only and move over the 1.13 jobs. We could probably just remove the 1.10 job since if we follow the pattern we only test other julia versions on x64 linux