Does an enabled keyboard mean ready? Local state, asynchronous ack and page focus
Local enabled/visible is not page acknowledgement or inserted text
setKeyboardEnabled updates keyboardEnabled/keyboardVisible before awaiting enqueueKeyboard’s Bool result, then rejects stale results through keyboardRequest. Failure can revert state/show an entry notice. dismissKeyboard requests UIKit end-editing and queues page disable. Local hiding proves neither immediate page acknowledgement nor saved/sent drafts.
Different from origin, insertion bridges and audio busy states
Origin checks boundaries, bridges check text adaptation, busy states check resources. Here keyboard intent versus page confirmation versus actual input session matters. No keyboard, typing or page change ran.
| Check | Verified scope | Manual review | Not established |
|---|---|---|---|
| Optimistic state | Enabled/visible set first | Call phase versus final result | true means physically shown |
| Script ack | Bool true counts as accepted | Actual page conditions/errors | Every input supported |
| keyboardRequest | New request invalidates old UI result | Local order/ownership | Account/server request_id |
| Serial tasks | Await prior task, recheck current request | Dispatched versus waiting | Older operation never ran |
| Native focus | Attempt becomeFirstResponder | Native/DOM/device visibility | DOM focus guarantees software keyboard |
| Hide | Local state/endEditing/queued disable | Intent versus page ack | Every page change instantly complete |
| Guards | Current webView/isTrusted required | Actual page/composer | Same name permits input |
| Handoff | Text/draft/receipt separately | Actual contents/authority | Keyboard state proves sending |
Synthetic keyboard timing
Local fields become true on enable request while script acknowledgement is pending. Report intent, not readiness. A later disable means the old enable response cannot update the newer UI conclusion.
Actual device/input evidence remains necessary. Avoiding busy-page delay is a design intent, not zero-latency/universal-page acceptance. Do not change local booleans to bypass conditions.
FAQ
Does true prove a visible keyboard?
No. Optimistic state precedes ack; physical input conditions differ.
Does Bool ack prove inserted transcription?
No. It acknowledges keyboard configuration, not insertion/sending.
Is keyboardRequest a server request ID?
No. Local ordering grants no identity/write authority.
Does hiding instantly confirm page disable?
No inference follows; queued page operations differ from local display.
Does DOM focus guarantee a device soft keyboard?
No. Native input-session/device conditions need validation.
Was real typing/device testing performed?
No. Fixed-source scope only, without browser/device acceptance.
Reading and source
Page/origin · Text bridge · Busy states · Start · Privacy
