ci: cache workspace crates and share rust caches across all workflows

[desmondcheongzx]

Extract the inline restore-mtime script into a reusable composite action (.github/actions/restore-mtime/) and apply the full cargo cache treatment to all 10 jobs that use Swatinem/rust-cache:

• fetch-depth: 0 on checkout (full history for mtime restoration)
• restore-mtime composite action (deterministic source file timestamps for cargo fingerprinting)
• cache-workspace-crates: "true" (preserve workspace crate artifacts in cache — the key fix from ci: cache workspace crate artifacts in integration build #6261)
• shared-key by cargo profile (consolidate caches instead of one per job)
• save-if: main-only (single source of truth, avoids PR cache pollution and LRU eviction)

Cache keys are shared by cargo profile:

• Linux-dev-build: unit-test, rust-tests, doctests, test-imports, style, docgen (build-docs.yml)
• Linux-integration-build: integration-test-build, profile-daft, property-based-tests
• Linux-dev-bench-build: benchmark-codspeed (buildjet, separate cache provider)

Previously each job maintained its own cache via prefix-key, resulting in ~8 separate cache entries for the same profile. Consolidating to 3 shared keys frees significant space against the 10 GB repo cache limit.

On the integration-test-build job (#6261), this approach reduced Build wheels from 36 min to 10 sec on warm cache. Expect similar improvements for unit-test (~8 min build) and benchmark-codspeed (~25 min compile + link).

See #6244, #6246, #6261.

[greptile-apps]

Greptile Summary

This PR consolidates Rust build caching across all CI workflows by extracting the mtime restoration logic into a reusable composite action and standardizing cache configuration. Previously, each job maintained its own cache via prefix-key, resulting in ~8 separate cache entries for the same cargo profile. This PR consolidates to 3 shared cache keys grouped by cargo profile:

• Linux-dev-build: unit-test, rust-tests-platform, doctests, test-imports-platform, style, docgen (6 jobs)
• Linux-integration-build: integration-test-build, profile-daft, property-based-tests (3 jobs)
• Linux-dev-bench-build: benchmark-codspeed (1 job, separate buildjet cache provider)

Key improvements:

• Extracted inline mtime restoration script into .github/actions/restore-mtime/ composite action with proper macOS/BSD compatibility
• Applied uniform cache treatment to all 10 jobs: fetch-depth: 0, restore-mtime action, cache-workspace-crates: "true", shared cache keys by profile, and save-if: main-only
• Expected significant build time reductions on warm cache (integration-test-build saw 36min → 10sec improvement in ci: cache workspace crate artifacts in integration build #6261)
• Frees cache space against the 10 GB repo limit by eliminating duplicate cache entries

Confidence Score: 5/5

• This PR is safe to merge with minimal risk
• The changes are well-structured CI/CD improvements that follow established patterns. The mtime restoration script has been properly extracted into a reusable composite action with cross-platform compatibility (macOS/BSD detection added). All cache configurations are consistently applied across workflows with appropriate conditionals for Windows. The consolidation strategy is clearly documented and matches the PR description.
• No files require special attention

Important Files Changed

Filename	Overview
.github/actions/restore-mtime/action.yaml	New composite action that restores file mtimes from git history with proper macOS/BSD compatibility detection
.github/workflows/build-docs.yml	Updated to use shared cache key Linux-dev-build, added fetch-depth: 0, restore-mtime action, and cache-workspace-crates
.github/workflows/daft-profiling.yml	Updated to use shared cache key Linux-integration-build, added fetch-depth: 0, restore-mtime action, and cache-workspace-crates
.github/workflows/pr-test-suite.yml	Replaced inline mtime restoration script with composite action, updated 7 jobs to use shared cache keys (dev-build, integration-build, dev-bench-build) with workspace crate caching
.github/workflows/property-based-tests.yml	Updated to use shared cache key Linux-integration-build, added fetch-depth: 0, restore-mtime action, and cache-workspace-crates

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    subgraph "Cache Key Strategy"
        A[10 CI Jobs] --> B{Cargo Profile}
        
        B -->|dev profile| C[Linux-dev-build]
        B -->|integration profile| D[Linux-integration-build]
        B -->|dev-bench profile| E[Linux-dev-bench-build]
        
        C --> C1[unit-test]
        C --> C2[rust-tests-platform]
        C --> C3[doctests]
        C --> C4[test-imports-platform]
        C --> C5[style]
        C --> C6[docgen]
        
        D --> D1[integration-test-build]
        D --> D2[profile-daft]
        D --> D3[property-based-tests]
        
        E --> E1[benchmark-codspeed]
    end
    
    subgraph "Cache Optimization"
        F[fetch-depth: 0] --> G[restore-mtime action]
        G --> H[Deterministic timestamps]
        H --> I[cache-workspace-crates: true]
        I --> J[shared-key by profile]
        J --> K[save-if: main only]
        K --> L[Reduced cache pollution]
    end

Loading

_{Last reviewed commit: 64c5ac2}

[greptile-apps]

_{5 files reviewed, 1 comment}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

touch -d is GNU-specific and will fail on macOS runners (BSD touch). The unit-test job in pr-test-suite.yml runs on macos-latest for main branch commits.

BSD touch uses different syntax: touch -t [[CC]YY]MMDDhhmm[.SS]

Suggested change

	| while IFS=$'\t' read -r ts file; do [ -f "$file" ] && touch -d "@$ts" "$file"; done
	| while IFS=$'\t' read -r ts file; do [ -f "$file" ] && touch -d "@$ts" "$file" 2>/dev/null || touch -t "$(date -r "$ts" +%Y%m%d%H%M.%S)" "$file"; done

[greptile-apps]

_{5 files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 73.45%. Comparing base (f52e4f6) to head (64c5ac2).
⚠️ Report is 5 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #6264      +/-   ##
==========================================
- Coverage   73.45%   73.45%   -0.01%     
==========================================
  Files        1001     1001              
  Lines      133124   133120       -4     
==========================================
- Hits        97792    97778      -14     
- Misses      35332    35342      +10     

see 12 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.