fix: resilient post-adoption sync fallback (v4.5.1) The inline snapshot fallback used a single fire-and-forget triggerSyncIfReachable plus a setTimeout retry that dies with the MV3 service worker. If the snapshot failed and the user navigated away, the vault stayed empty until the 10-minute alarm tick. - Snapshot receive timeout bumped 15s → 30s - setTimeout retry replaced with chrome.alarms one-shots at 30s and 2min (schedulePostAdoptionSync) — survives SW death - Both source and receiver session fallbacks use the alarm-based retry - Alarm listener in sync-alarm.ts handles the post-adoption prefix Gate: tsc clean, 3645 tests pass, Chrome/Firefox/Mobile builds pass.
$ koh steal kepr.uk/peach-browser@bcae1b330341
·
parent: dbac78ce978b
discussion
log in to leave a comment.