[michakraus]

Summary

• Lifts the Aqua = "0.8.0 - 0.8.16" bound in test/Project.toml (from Bound Aqua below 0.8.17 in the test environment #288 / commit 5c28167a) back to plain Aqua = "0.8", and removes the now-stale explanatory comment.
• Aqua 0.8.18, released 2026-09-18, fixes the exact bug the bound worked around — an uninstalled weak dependency (EnzymeCore behind KernelAbstractions) no longer stops test_persistent_tasks's manifest walk.
• CHANGELOG entry added documenting the lift, immediately after the bullet that introduced the bound.

Pre-PR verification

Checked and clean

• Aqua.test_persistent_tasks(GeometricMachineLearning) passes with Aqua resolved to 0.8.18 and KernelAbstractions 0.9.42 — the same KernelAbstractions version present in the last green run and in the red one. Julia 1.13.0, fresh process, --check-bounds=auto.
• The failure scenario is genuinely exercised rather than absent: the resolved test environment has no EnzymeCore stanza in its manifest, and Base.locate_package for EnzymeCore returns nothing. Under 0.8.17 that value reaches error("Unable to locate \EnzymeCore` …"); 0.8.18 filters [weakdeps]out of the manifest entry'sdeps` and skips an uninstalled weak dependency.
• The whole test/aqua.jl testset is green — 10 passes, 8.2 s — run with the package set runtests.jl has loaded by that point (Zygote, GeometricIntegrators, HDF5). This covers the six enabled test_all checks and the Aqua.Piracy.hunt == 12 gate.
• Every factual claim in the new CHANGELOG entry is verified. The quoted sentence is verbatim from Aqua's own CHANGELOG.md for v0.8.18, and the 2026-09-18 release date matches.
• No stale comment survives: the explanatory comment is removed with the bound it explained, and test/Project.toml now carries no comment at all. No reference to 0.8.16/0.8.17 remains anywhere outside CHANGELOG.md.
• Aqua's own [deps]/[compat] are identical between 0.8.16 and 0.8.18, so lifting the bound moves no other package in the resolution.
• All of Aqua 0.8.16, 0.8.17 and 0.8.18 declare julia = "1.0", so the min CI job cannot be pushed back onto a broken version by a Julia-version constraint.
• Both changed files are NFC-normalised.

Considered and declined: a floor of Aqua = "0.8.18" would close the (CI-safe) risk of a stale local test/Manifest.toml staying on 0.8.17. Declined to keep the bound consistent with the tree-wide convention (plain "0.8", as every other package's Aqua entry reads, and as SimpleSplines' equivalent unpin — JuliaDEC/SimpleSplines.jl#21 — also settled on).

Not checked

• Full Pkg.test() — outside the scope of the pre-PR gate. test/aqua.jl is the only test file that uses Aqua, the diff changes no src/ and no other test file, and no other package moved in the resolution.
• Type piracy origin, type instability and allocations — no executable change in the diff.
• Doctests, lint and formatting — no docs/ change and no Julia source changed.

🤖 Generated with Claude Code

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 80.87%. Comparing base (89dd343) to head (48df468).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #298   +/-   ##
=======================================
  Coverage   80.87%   80.87%           
=======================================
  Files          80       80           
  Lines        3153     3153           
=======================================
  Hits         2550     2550           
  Misses        603      603           

☔ 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.