[ProtonDev-sys]

Release

Release v1.3.0 after the eleven optimization, feature and reliability PRs #31–#41, all now merged. CMake, the executable resources and the administrator manifest have matching version metadata.

Compared with current main c037b4020ce6ca355de268e6835e2de369fa0dca, the release tree changes only eight version/documentation files. The final follow-up commit d059d2d46d5f303e90cf2fa06b43f8b51b0e2872 changes only docs/PERFORMANCE-ROUND3.md and docs/releases/v1.3.0.md; every compiled source, build configuration and version input is identical to the fully tested candidate below.

Documentation

• Cover scoped search, CSV export and shortcuts, sampling recovery, process identity, atomic history/export saves, unavailable-data feedback and Python tooling.
• Retain exact before/after source and binary identities, observed ranges, responsiveness budgets, network results and evidence hashes.
• Keep the earlier paired study separate from the complete release candidate: main timing ranges overlap, observed diagnostic memory increased, and the study executable grew by 21.5 KiB. The later verified candidate is 966,144 bytes, 4 KiB smaller than the original build; no new timing or production-memory claim is inferred from that size.
• Record the 38-view native screenshot review and three corrected GPU rendering fixtures, together with platform limitations.
• Use tag-specific absolute links so references work in the published GitHub release notes.

Verification

The complete release-candidate Windows workflow succeeded for PR head 499e98c8e4c0f7b388193f92b03d6bda95638d53, tested merge 4f57bc78c4aaa47f232b57d92abb945bedf562f2, tree 64024132208ea172e192d6480d1a5a82aaf4edcf:

• 12/12 CTests; 357 passing named native checks; 21/21 suite scenarios.
• 31 passing diagnostic reports, including 25 UI reports with healthy sampling and passing keyboard, sorting, coalescing and tab-stress checks.
• All 18 retained suite UI reports independently pass the existing responsiveness budgets; no thresholds were relaxed.
• Windows: 31 Python tooling tests and 19 workflow guards. Linux: 29 tooling tests, two explicit Windows-only skips, 19 guards and portable search/CSV regressions.
• 300 retained PNGs decode successfully. All 75 GPU availability fixtures exactly match the three visually reviewed light/dark/High Contrast images.
• Candidate executable SHA-256 fc21c6d38b8459fcba59509d6997ad39d7823abdbf18573bb3139ca2ada229ef agrees with its package checksum.

The earlier paired study, history correction and GPU correction also completed successfully. The paired study includes the 38 reviewed captures and owned Windows UDP attribution test.

The final follow-up changes only the two evidence documents; its compiled inputs match the verified candidate. The merged main commit has now passed its own complete workflow and publication audit below.

Final main verification and publication

Windows run 38213050710 completed successfully for merged main 68201184b3afaa7a67ea6b367554bd7d25873a79. All four jobs passed, including publication. Its tree b3602e0a7e4ddf61abb1db338091897c5985487c exactly matches the reviewed release tree.

• 12/12 CTests in 189.74 seconds; 357 unique passing named native checks in two identical reports; 21/21 suite cases.
• All 31 retained diagnostic JSON reports pass. All 25 UI reports pass sampling health, keyboard, sorting, coalescing and tab-stress gates, with no failure label.
• Every one of the 18 suite UI reports passes the existing budgets in an offline check of the retained reports. No scenarios were rerun for this audit.
• All 300 PNGs pass integrity validation and fully decode. All 75 GPU fixture files match the three visually reviewed images exactly.

Responsiveness metric	Largest main value	Existing budget
UI update p95	5.0341 ms	50 ms
Cached sort p95	1.4570 ms	50 ms
Tab switch p95	26.7148 ms	50 ms
Message queue p95	63.4458 ms	100 ms
Letter navigation p95	0.8063 ms	100 ms
First sample	405.199 ms	2,000 ms

v1.3.0 is published. Its tag and release target both identify the exact merged commit above. The downloaded main artifact's release.json matches that commit and version; its release notes exactly match the committed and published notes.

The published TaskManager.exe is 966,144 bytes, 4 KiB smaller than the original 970,240-byte build. Its actual SHA-256, package checksum entry and GitHub release asset digest all agree:

fc97a75d94f393ac485d91bda361cd028e37821c72df7b179e920f259953620f

The checksum file's own digest also agrees with the published asset. The final main package ZIP SHA-256 is 2612205b4c1b1ae6c81a80c4290ed390739114f9f7a9b71381d73d30d46b8f64; the verification ZIP SHA-256 is 77104f043b17a13100b8d4499db50bf266206d105ae775ff947305a0faec0bd9. The final binary identity and absolute responsiveness checks remain separate from the earlier paired timing study.

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

• Configuration used: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: a1ef852b-602d-44b8-b466-23f288f247b1

---

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

---

_{Comment @coderabbitai help to get the list of available commands.}