Queue correlated, volatile wake requests for the BTstack owner and expose
read-only completion state. Keep the controller chord, captured identity,
pairings, profiles, HD settings and existing wake burst unchanged.
Add Python API and wake CLI with bounded waits, explicit failure outcomes
and no automatic broadcast retries. Avoid importing SDL just to obtain
fixed sensor IDs so USB automation has clean output and no SDL dependency.
Windows libusb cannot control hub roots. Expose only INFO/WAKE on each
native child's existing vendor Interface 1 with isolated control state;
retain hub/HID drivers, descriptors and identities. Add safe sibling-aware
Windows discovery and a packaged Windows libusb runtime. Document manual
WinUSB binding to Interface 1 only and migrate examples to find_wake_pico.
Validate with 713 tests, 10 firmware/probe builds, Windows x64 wheel
resolution on Python 3.9/3.11, and a real child-interface wake burst on
Linux (2.03 s, no controller, no failures). Device/configuration descriptors,
serials and persistent state matched before/after. Physical Windows
hardware/driver operation remains unverified on this Linux workstation.
Preserve native frequency/amplitude timelines for DualSense PCM output,
remove the unsupported 64-frame path, and retain compatibility rumble for
other controllers. Use the validated 300 MHz sampling phase and restrict
non-bondable Classic discovery autoconnect to previously paired peers.
Buffer native-hub UART stdout and release USB IRQs around port-reset
callbacks. Add observer liveness and pre-SETUP root-response observations
without changing recovery behavior. Cover transport and haptics boundaries.
Record the user-accepted 0.108 trial: controls remained responsive and
rumble felt fine. Instrumentation changes timing; the disconnect root
cause and long-term reliability remain unqualified.
Validation: 624 tests, 11 affected firmware/probe builds, and on-device
concurrent USB and HD-auto-start checks. Private captures, generated
images, and unrelated working-tree files are intentionally excluded.
Add per-profile native solo routing, stick/click swapping, SL/SR output mappings, and digital left-stick directions with analog priority. Extend both codecs to schema 11 while preserving legacy profiles, expose the controls and preview in Studio, and carry rail outputs through Pro emulation. Verified with 541 tests, firmware builds, editor checks, and saved-profile preservation during native 0.89 deployment.
Keep native backend locking IRQ-permitting and service USB hardware from SRAM during BOOTSEL sampling. Add bounded input/control flight recording, root IN and EP0 snapshots, and masked-window handoff coverage. Record PC qualification and the remaining Switch stability limitations.
Pause Wii pointing work with firmware 0.33-wii-trace installed. Preserve the pinned libogc IR pipeline, calibrated native IMU, software BOOTSEL, and standard camera sensitivity trial.
Tracking instability, tracking loss, and ineffective vertical movement remain unresolved. Level-2 camera filtering is not hardware-qualified. Nine targeted regression tests pass; firmware and persistent storage were verified after flashing.
Implement the standalone USB protocol probe and Bluetooth-backed right Joy-Con bridge with its own persistent virtual pairing identity. Preserve complete ordered native reports, including opaque motion data, and match the console feature set. Relay built-in vibration cues only after genuine source acknowledgement and expose safe BOOTSEL pairing control. Include native capture diagnostics and focused protocol, packet-lifecycle, and cue regressions. Native mouse operation confirmed on Switch with bridge 0.24; private captures and firmware backups remain outside the commit.
