fix: detect stale PC disconnections - #114
Conversation
|
Review follow-up at b02640d:\n\n- Resolved the recovery-watchdog timing finding by coordinating health probes before/after background profile requests.\n- Classified pointer-profile native write failures separately so reconnecting begins immediately during both initial negotiation and recovery.\n- Added an absolute 22-second recovery presentation deadline so slow healthy probes cannot prolong the unavailable-state decision.\n- Added regression coverage for stale loss during a pending recovery response, initial/recovery write failures, and slow healthy probes.\n\nValidation: npm run validate passed (49 suites, 393 tests; Expo Doctor 21/21). Independent latest-head re-review and refreshed CI are in progress. |
|
Final handoff at b02640d:\n\n- Independent latest-head review: no actionable findings.\n- Local validation: npm run validate passed (49 suites, 393 tests; Expo Doctor 21/21).\n- CI run 32848600621: validate, Android native build, and iOS native build all passed.\n\nThe PR remains draft and unmerged pending explicit approval. |
Summary
Closes #113
Validation
pm run validate (pass: 49 suites, 389 tests, Expo Doctor 21/21)
px expo config --type public (pass; version 1.0.0/build 17 and native configuration unchanged)
Physical testing
Compatibility