ci: cache check-features

[carneiro-cw]

PR Type

Enhancement, Tests

---

Description

• Add cargo-hack@0.6.39 to CI tools setup

• Introduce "Check all features" caching step

• Prime cache for all feature combinations

• Preserve existing build and test caching

---

Diagram Walkthrough

flowchart LR
  A["Set up dependencies"] --> B["Build Debug binaries"]
  B --> C["Check all features"]
  C --> D["Save cache"]

Loading

File Walkthrough

	Relevant files
Configuration changes	build-debug-cache.yml Enhance CI caching for feature checks                                        .github/workflows/build-debug-cache.yml Installed cargo-hack@0.6.39 alongside existing tools Added "Check all features" step to cache feature checks Cleared RUSTFLAGS to avoid coverage flag mismatches Retained existing test and build cache steps +11/-1

---

[cloudwalk-review-agent]

Summary

Looks good. This change cleanly adds cargo-hack installation and primes the shared debug cache with cargo hack check --each-feature, which is aligned with the stated goal of avoiding repeated feature-combo rechecks in PR jobs. The explicit RUSTFLAGS: "" is a good safeguard to keep cache fingerprints compatible with the non-coverage check-features workflow.

No concrete correctness, security, or deploy-safety issues found in the provided diff/context.

[github-actions]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 2 🔵🔵⚪⚪⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ No major issues detected

[github-actions]

PR Code Suggestions ✨

Explore these optional code suggestions:

Category	Suggestion	Impact
General	Split tool installation steps The install-action step may not support installing multiple tools via a single space-separated string. Split this into two steps so each tool is installed explicitly to avoid silent failures. .github/workflows/build-debug-cache.yml [56-57] -with: - tool: cargo-llvm-cov@0.6.14 cargo-hack@0.6.39 +- name: Set up cargo-llvm-cov + uses: taiki-e/install-action@cfdb446e391c69574ebc316dfb7d7849ec12b940 + with: + tool: cargo-llvm-cov@0.6.14 +- name: Set up cargo-hack + uses: taiki-e/install-action@cfdb446e391c69574ebc316dfb7d7849ec12b940 + with: + tool: cargo-hack@0.6.39 + Suggestion importance[1-10]: 6 __ Why: Splitting the tool input into two explicit steps prevents potential silent failures if the install-action doesn’t support space-separated tool lists, improving reliability.	Low
Possible issue	Include workspace in feature check Without --workspace, cargo hack only checks the root crate’s features. Add --workspace so all workspace crates get .rmeta artifacts cached and avoid missing targets. .github/workflows/build-debug-cache.yml [79] -run: cargo hack check --each-feature --keep-going +run: cargo hack check --workspace --each-feature --keep-going Suggestion importance[1-10]: 5 __ Why: Adding --workspace ensures that feature checks and cached .rmeta artifacts include all crates in the workspace rather than just the root crate.	Low

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 2e89c3c769

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cloudwalk-review-agent]

Summary

LGTM. The PR cleanly separates check-features caching from the existing shared debug cache by introducing a dedicated cache key/workflow (check-features-release) and updating check-features.yml to restore from that key with an OS + Cargo.lock hash suffix.

The new build-check-features-cache.yml primes cache using the same command/env (just check-features, RELEASE=1, CARGO_PROFILE_RELEASE_DEBUG=0) as the consumer job, which is the right way to maximize artifact reuse and avoid fingerprint divergence. I did not find concrete correctness, security, or CI-safety regressions in the provided diff.