[khanra17]

Summary

• compile and link the real Tauri executable for all five existing OS/architecture targets without release-only packaging
• prepare frontend assets and CGO-free Go sidecars once, then reuse them across the native matrix
• add exact-output and incremental TypeScript, Go, and Rust caches, including useful cache saves after late failures
• add a fast CI-only Cargo profile while preserving release profile behavior
• retain clear platform-specific job names and GitHub step summaries

Timing comparison

Measured on GitHub-hosted runners against the latest successful upstream Build Matrix run. Runner load and cache state can vary.

Scenario	Wall time	Difference
Upstream matrix	34m53s	baseline
Optimized final validation	4m26s	about 87% faster / 7.9x
Optimized one-line changes in frontend, Go, and Rust	4m56s	about 86% faster / 7.1x

Notable one-line-change cache results:

• frontend check/build: 51s to 13s
• Go sidecars: 2m52s to 6s
• Windows Rust compile: 6m07s to 2m08s
• shared preparation: 4m31s to 1m06s

The upstream baseline includes a 33m49s universal macOS verification build. Universal installers remain covered by the release workflow; PR verification still compiles and verifies both macOS architecture executables.

Verification

• Final atomic-commit run: https://github.com/khanra17-org/cockpit-tools/actions/runs/30347787742
• Incremental cache experiment: https://github.com/khanra17-org/cockpit-tools/actions/runs/30347043683
• Upstream baseline: https://github.com/jlcodes99/cockpit-tools/actions/runs/30339148636
• Final run completed successfully with zero annotations.

Commits

• build: avoid redundant universal sidecar builds
• ci: add fast verification build profile
• ci: streamline pull request build matrix