fix(ci): let the nightly Miri job honour strict-provenance ignores

[martintmk]

🤖 Clawpilot here! Posted automatically by Clawpilot (an AI agent), not by a human. Please verify before acting.

Summary

Approach changed after review. The [patch.crates-io] git pin of parking_lot_core and the deny.toml allow-git carve-out are gone; the branch is rebased onto current main and now contains only a CI fix plus three test attributes.

Root cause (revised)

The workspace already has the exclusion mechanism this needs. crates/internity/tests/basic.rs gates freeze_races_writer_and_stays_prefix_consistent behind #[cfg_attr(miri_strict_provenance, ignore = "...")], cfg(miri_strict_provenance) is declared in the workspace check-cfg list, and just anvil-miri-strict-provenance defines that cfg via RUSTFLAGS.

The scheduled miri-strict-provenance job in .github/workflows/nightly.yml inlines the cargo invocation instead of calling that recipe, and it sets only MIRIFLAGS — never RUSTFLAGS=--cfg miri_strict_provenance. So the cfg was never defined in CI, the ignore never applied, the test ran anyway, and parking_lot_core's integer-to-pointer cast at word_lock.rs:320 aborted the whole internity test binary. That is why nightly kept failing even after the suppression was added.

Changes

• .github/workflows/nightly.yml — set RUSTFLAGS: --cfg miri_strict_provenance on the strict-provenance step, mirroring the just recipe it inlines, and document the mechanism next to the job.
• crates/internity/tests/basic.rs — apply the same cfg_attr ignore to the three remaining tests that contend a parking_lot lock and reach the same cast: concurrent_intern_is_consistent, concurrent_intern_then_concurrent_resolve, threaded_intern_bytes_is_consistent_across_threads.

internity, cachet and cachet_tier are the only workspace crates that depend on parking_lot, and the first two sort before internity, so they already clear the job today.

Why this shape

This is what was asked for on this PR: exclude the affected tests from strict-provenance Miri until a fixed parking_lot_core is released, rather than pull a third-party git revision into the build. unknown-git = "deny" stays intact and deny.toml is untouched. No coverage is lost outside strict provenance — the tests still run on every host target and under the miri-tree-borrows and miri-race-coverage nightly jobs.

crates.io still has no fixed release: parking_lot_core 0.9.12 (2025-10-03) is the newest version, and the upstream AtomicPtr commit is unreleased. The upstream "please cut a release" issue is still to be filed; there is no open issue for it on Amanieu/parking_lot today.

Note: setting RUSTFLAGS in the job env overrides .cargo/config.toml's target-cpu=x86-64-v3, which is irrelevant under Miri and matches what the just recipe already does.

Validation

• cargo +1.96.1 test -p internity --test basic --all-features — 52 passed, 0 ignored (normal builds are unaffected).
• Same command with --cfg miri_strict_provenance injected — 48 passed, 4 ignored, no unexpected_cfgs warning. This is the direct proof that the cfg reaches the integration test and that the ignore attributes fire.
• cargo +1.96.1 clippy -p internity --all-features --all-targets -D warnings — clean.
• rustfmt +nightly-2026-05-30 --check --config-path unstable-rustfmt.toml on the changed file — clean.
• nightly.yml parses as valid YAML.

The failing path itself is Linux-only (the threaded tests are cfg(not(all(miri, windows)))), so the scheduled job on main remains the platform-exact confirmation.

Bug: https://o365exchange.visualstudio.com/O365%20Core/_workitems/edit/7734570

[Copilot]

Pull request overview

Pins parking_lot_core to a specific upstream git revision to fix a nightly Miri strict-provenance failure caused by integer↔pointer round-tripping in WordLock, while keeping the existing strict-provenance CI coverage unchanged.

Changes:

• Add a [patch.crates-io] override for parking_lot_core pointing at the upstream provenance-safe fix commit.
• Update Cargo.lock so parking_lot_core is sourced from the pinned git revision instead of crates.io.

Reviewed changes

Copilot reviewed 1 out of 2 changed files in this pull request and generated 1 comment.

File	Description
Cargo.toml	Adds a crates.io patch override pinning parking_lot_core to the upstream strict-provenance fix commit.
Cargo.lock	Updates the resolved source for parking_lot_core to match the new git patch.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.0%. Comparing base (8c5a3c9) to head (f9f7a58).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #664   +/-   ##
=======================================
  Coverage   100.0%   100.0%           
=======================================
  Files         582      583    +1     
  Lines       62852    62891   +39     
=======================================
+ Hits        62852    62891   +39     

Flag	Coverage Δ
linux	88.7% <ø> (-11.3%)	⬇️
linux-arm	88.7% <ø> (-11.3%)	⬇️
scheduled	?
windows	88.7% <ø> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ 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.

[Copilot]

Pull request overview

Copilot reviewed 22 out of 23 changed files in this pull request and generated no new comments.

[Copilot]

Pull request overview

Copilot reviewed 22 out of 23 changed files in this pull request and generated no new comments.

Suppressed comments (1)

.patches/parking_lot_core/Cargo.toml:11

• The vendored parking_lot_core now uses strict-provenance pointer APIs (e.g., ptr::without_provenance_mut, pointer .map_addr()/.with_addr()), which are not available on Rust 1.71. Keeping rust-version = "1.71.0" misrepresents the minimum supported Rust version and can lead to confusing build failures on older toolchains.

rust-version = "1.71.0"

[Copilot]

Pull request overview

Copilot reviewed 2 out of 3 changed files in this pull request and generated no new comments.

Suppressed comments (1)

deny.toml:67

• This comment says the allowlist is "commit-pinned", but allow-git only allows the repository URL (any commit). The commit pinning happens in Cargo.toml via rev, so the comment here is misleading.

# Temporary, commit-pinned upstream fix for parking_lot_core strict provenance.

[Sander Saares (sandersaares)]

Hmmm.

Do we even need parking_lot here? It can be a bit cargo-culted by AI sometimes. I suggest also trying with std synchronization primitives. They are better maintained and not necessarily slower.

[Sander Saares (sandersaares)]

If we need to use random Git patch then I would rather just exclude this from the Miri checking until a proper patched version is available. Plus file an upstream "hey release a fix please" issue.

[Evgenii (Vaiz)]

🤖 Clawpilot here! Posted automatically by Clawpilot (an AI agent), not by a human. Please verify before acting.

Sander Saares (@sandersaares) martintmk — taken as directed. The git patch and the deny.toml carve-out are both dropped; unknown-git = "deny" is untouched and deny.toml no longer appears in the diff at all. The branch is rebased onto current main and the PR is now two files.

While implementing the exclusion I found the actual reason nightly kept failing, which changes the fix from "add a suppression" to "make the existing suppression work":

• crates/internity/tests/basic.rs already gates freeze_races_writer_and_stays_prefix_consistent behind #[cfg_attr(miri_strict_provenance, ignore = "...")] (added in perf(ci): parallelize scheduled Miri and reduce resource outliers #706), and cfg(miri_strict_provenance) is already declared in the workspace check-cfg list.
• That cfg is defined by just anvil-miri-strict-provenance through RUSTFLAGS. The scheduled job inlines the cargo command instead of calling the recipe and sets only MIRIFLAGS — so the cfg was never defined in CI and the ignore never fired. The test ran, and parking_lot_core aborted the whole internity binary.

So the fix sets RUSTFLAGS: --cfg miri_strict_provenance on that job and extends the same cfg_attr to the three other tests that contend a parking_lot lock. Coverage is untouched everywhere else — those tests still run on every host target and under the tree-borrows and race-coverage Miri jobs.

Verified locally that the mechanism actually works rather than assuming it: cargo test -p internity --test basic --all-features reports 52 passed / 0 ignored normally, and 48 passed / 4 ignored with --cfg miri_strict_provenance injected, with no unexpected_cfgs warning.

On the upstream ask: crates.io still shows 0.9.12 (2025-10-03) as newest, the AtomicPtr fix is unreleased, and there is no open "please cut a release" issue on Amanieu/parking_lot. I have not filed one — opening an issue on a third-party repo under a maintainer's identity is outside what I'll do unprompted. Say the word and I'll draft it, or file it yourself; either way it is the exit criterion for removing these four cfg_attrs.

[Copilot]

🟢 Approval recommended

The change is narrowly scoped to CI configuration plus test-only ignore attributes, with only a small maintainability follow-up noted.

Review details

• Files reviewed: 2/2 changed files
• Comments generated: 1
• Review effort level: Lite

[github-actions]

⚠️ Potential breaking changes detected

cargo semver-checks flagged the following on this PR. This is informational -- breaking changes between commits are expected; the major-version bump happens at release time, not on every PR.

internity

     Cloning origin/main
    Building internity v0.2.0 (current)
       Built [   4.158s] (current)
     Parsing internity v0.2.0 (current)
error: unsupported rustdoc format v60 for file: /home/runner/work/oxidizer/oxidizer/target/semver-checks/local-internity-0_2_0-default-8d74fec22ef71f4f/target/doc/internity.json
(supported formats are v55, v56, v57)

[martin-kolinek]

@sasaares, just to double check, your random patch applies to a previous version of this PR, right? I don't see any random patches, just exclusion of a few tests.

[Copilot]

🟢 Approval recommended

The changes consistently apply the existing cfg-based suppression mechanism and correctly wire that cfg into the strict-provenance Miri CI step.

Review details

• Files reviewed: 2/2 changed files
• Comments generated: 0 new
• Review effort level: Lite