test: Desktop deadline checks no longer race CI scheduling

[teknium1]

Desktop orphan-reap deadline tests no longer race CI scheduling before their first record.

• Freeze each existing budget fixture's clock until its first identity probe completes, then advance to the exact deadline; always restore the clock on test completion.
• Retain both invariants: confirmed PID mismatches are discarded, confirmed orphans are stopped, and unprocessed records survive budget exhaustion.
• Add an opt-in scheduling-pressure replay under evals/; production code and deadline behavior are unchanged.

Root cause: the fixtures allowed only 1 ms of real time before the first iteration, so a scheduling pause correctly preserved both records instead of processing the first.

Live repro: scheduling-controlled execution of the real ownership module on base 2237be355906fbe6065ce1815711eee52b2d646e and post-merge main 5280fe99872a58f1b9d25b00a79d240ecd035f7d reproduces the run 34171261707 assertion ([] instead of [62]). Both source and test files are byte-identical across those commits; this is preexisting, not caused by #105442. The deadline fixtures originated in #101929; existing-fix sweep found no superseding clock fix.

Validation

Check	Before	After
Same 5 ms real scheduling delay at ownership clock reads	2 failed, 17 passed	19 passed
Disable production deadline guard (mutation control)	—	Both deadline tests reject over-processing
Change >= deadline boundary to > (mutation control)	—	Both deadline tests reject over-processing
Full Electron platform suite, normal and pressure replay	—	Each: 151 files / 2,139 tests passed; 2 files / 6 tests skipped
Desktop typecheck, touched-file ESLint and Prettier	—	Passed
Windows footguns, compat-pointer gate, diff whitespace	—	Passed

The replay injects actual pauses, not fabricated timestamps; the corrected fixtures explicitly control time. Local dependency setup initially lacked desktop packages/Electron; a clean lockfile install and Electron install resolved that before the full green runs.

NODE_OPTIONS="--require=$PWD/evals/desktop-reap-clock-pressure.cjs" \
  npm run test:desktop:platforms -w apps/desktop -- electron/backend-ownership.test.ts

Infographic

[github-actions]

૮ >ﻌ< ა ci review

_{ran on 781211b — test: make desktop orphan-reap deadline checks deterministic}

⚠️ Warnings

OSV vulnerability scan · View job

28 known vulnerabilities found in pinned dependencies.

• CVE-2026-67213 in website/package-lock.json
• CVE-2026-82417 in scripts/whatsapp-bridge/package-lock.json
• CVE-2026-82417 in website/package-lock.json
• CVE-2026-75931 in package-lock.json
• CVE-2026-75931 in website/package-lock.json
• CVE-2026-83610 in package-lock.json
• CVE-2026-83610 in package-lock.json
• CVE-2026-71554 in uv.lock
• CVE-2026-73088 in package-lock.json
• CVE-2026-73088 in website/package-lock.json
• GHSA-8423-8fgw-73vq in uv.lock
• CVE-2026-70608 in package-lock.json
• CVE-2026-73089 in package-lock.json
• CVE-2026-73089 in website/package-lock.json
• CVE-2026-75975 in package-lock.json
• CVE-2026-75975 in website/package-lock.json
• CVE-2026-75899 in package-lock.json
• CVE-2026-75899 in website/package-lock.json
• CVE-2026-84371 in package-lock.json
• CVE-2026-56876 in package-lock.json

How to fix:

Review the findings in the Security tab. Update the affected dependencies if a patched version is available.

---

debug info

CI timings

CI timings · View report · View job

Wall time 5m10s vs 4m58s (+4.0%). 9 job(s) slower, 5 faster, 1 unchanged.

• Python tests / Run tests: +32.0s
• OSV scan / Scan lockfiles / osv-scan: -11.0s
• Python lints / Windows footguns (blocking): -10.0s
• OS-specific tests / Windows-only tests: +9.0s
• OS-specific tests / macOS-only tests: -7.0s