The SDK transport can now ask the main loop whether it's truly idle before saying so.
What's wrong with this entry?
The transport used by the SDK and bridge gained a way to register a liveness probe, so when the last pending permission or dialog request resolves it announces idle rather than running if the main loop reports it has nothing to do.
- Print/headless mode registers the probe from its idle gate.
- If the probe throws, the failure is recorded under the existing
per_task_stop_sparingevent with reasonliveness_probe_throwand the session is reported as not idle. - Present on this build wherever the SDK or bridge transport is used, with no flag.
setMainLoopLiveness
Strings lifted out of the shipped bundle, so the claim above can be checked against them.