ship review · offline editing

The iOS reconnect stall

Offline editing passed six weeks of beta on 96 workspaces. One exception remains: iOS occasionally leaves a reconnect spinning until the app is backgrounded. The edits stay queued; this review decides whether the tested workaround is enough for the first patch.

Observed: 1 reconnect in about 40
Data loss: none in 8,412 offline sessions
Workaround: resend after 20 seconds
Upstream fix: acknowledged, no release date

Readiness docket

Ship offline editing in the first patch The beta is green except for the reconnect exception below. iOS resync stall One reconnect in about 40. Nothing is lost — the edits stay queued — but the reader sees a spinner until they background the app.

Land the upstream fix The vendor reproduced the stall and accepted the bug, but has not named a release.

Keep the resend workaround in the patch After 20 seconds the client resends the queued batch; 500 forced stalls all recovered.

Is that enough for the first patch?

Approve the workaround — ship now and keep chasing the upstream fix.

Request another pass — hold the patch for a native fix.

Keep reconnect telemetry The release cohort retains the beta alert and per-platform count.

Write the support playbook The spinner, queued-edit check, and background-app recovery are documented.

The objection that matters

The standalone page opens mid-review: its anchored thread asks whether a one-in-forty failure compounds enough to block launch, and carries the follow-up work the approver wants ready if the vendor's release slips. Approval records the direction; task statuses change when the agent publishes the resulting version.