[DeusExTaco]

Summary

I reproduced the header overflowing a 320px viewport, then kept its existing actions reachable without changing their routes, labels or handlers.

• Use two compact header rows below 640px, retaining the 48px desktop row.
• Keep username text bounded without shrinking the user icons or losing the full accessible name.
• Anchor Add and activity menus within the available header/viewport space. The activity popup is deliberately header-relative at all widths.
• Add four regression cases per browser for both themes and normal/long usernames, covering 320–1280px bounds, hit targets, keyboard order and menus. All existing header test bodies remain unchanged.

Narrow test-only CI correction

I also reproduced the scheduler test's one-second progress-watchdog sensitivity under controlled real Linux database latency. Its three progress waits now use a test-local five-second bound; the original real SQLite setup, ordering/concurrency assertions, 50ms exclusion observation and cleanup are unchanged. Controlled first/second/drain delays pass, while concurrent work and dropped scheduling still fail. This does not prove the exact cause or timing of the earlier remote TimeoutError; the available remote log omitted its detailed traceback.

Validation

• Complete uninstrumented Linux owning suites for the scheduler correction: 59 passed. Full local validation was repeated for the final test-only correction, with all 18,936 lane-qualified test IDs/outcomes retained and no target warnings.
• Genuine assertion-level RED for the original overflow, then focused Chromium/Firefox GREEN.
• Owning route suite: 7 passed; sidebar suite: 25 passed per browser.
• Full local make ci-full: passed, including 15,824 main tests, 91.67% coverage against the unchanged 90% gate, 26 accessibility tests, 742 passed / 2 skipped per browser, and all 7 Docker smoke tests.
• Existing Docker runtime guards and Grype High threshold passed. This was the normal cached Linux ARM64 build, not a fresh multiarch build; lower-severity findings remain.
• Isolated manual checks covered real keyboard navigation and action destinations, light/dark, reduced motion, narrow layouts and expanded/collapsed desktop sidebar states.

Scope and follow-ups

No production Python, JavaScript handlers, dependency versions, migrations, shared fixtures, worker settings or security policy changed. PR #165 remains deferred. The live 8585 instance and recovery worktree were not changed.

Existing broader content clipping is outside this header fix. Manual observations of donation-dialog focus returning to BODY and an invisible loading indicator during a delayed request are retained without attribution or an unrelated repair. This is not a whole-application accessibility certification.

Full exact-head remote checks and final review are required before merge.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	⚠️ Failed 2026-10-16T08:00:40.775398Z	0f09247	Draft marked ready

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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.