File: inkcontroller/src/main/java/com/inksdk/ink/OnyxInkController.kt:154-189
onBeginRawDrawing calls recordDispatch twice (for EVENT_KERNEL_TO_JVM then PEN_KERNEL_TO_JVM). When tsMs <= 0L, each call increments dispatchProbeCount and logs "dispatch probe: tp.timestamp=0 (sdk did not populate)".
Result: the very first event with a null timestamp produces two identical probe lines; after one stroke and a half (DOWN + PEN-DOWN + first MOVE) we hit count=3 and emit the disable line. Not incorrect, but noisier than the comment ("auto-detect on the first event") suggests.
Suggested fix: gate the increment + log on metric == EVENT_KERNEL_TO_JVM (the always-called one), or track "probed this event already" out-of-band.
File:
inkcontroller/src/main/java/com/inksdk/ink/OnyxInkController.kt:154-189onBeginRawDrawingcallsrecordDispatchtwice (forEVENT_KERNEL_TO_JVMthenPEN_KERNEL_TO_JVM). WhentsMs <= 0L, each call incrementsdispatchProbeCountand logs"dispatch probe: tp.timestamp=0 (sdk did not populate)".Result: the very first event with a null timestamp produces two identical probe lines; after one stroke and a half (DOWN + PEN-DOWN + first MOVE) we hit count=3 and emit the disable line. Not incorrect, but noisier than the comment ("auto-detect on the first event") suggests.
Suggested fix: gate the increment + log on
metric == EVENT_KERNEL_TO_JVM(the always-called one), or track "probed this event already" out-of-band.