fix(logitech): G305 LIGHTSPEED — connect via Nano receiver and hide unsupported controls

[jazzstack]

Summary

Two fixes make the G305 LIGHTSPEED usable: it never connected before (probing latched onto its receiver's HID++ 1.0 error reply), and once connected it showed gaming-surface / LightForce controls the mouse does not have. Both are verified on hardware.

Fix 1 — device connect

The G305 connects through a Nano Lightspeed receiver (046d:c53f) that was missing from LOGITECH_RECEIVER_PRODUCT_IDS, so device-index probing ran the wrong way around.

• Root cause: resolveDeviceIndex() probes two HID++ device indexes to learn whether a mouse answers on its own endpoint (0xFF) or through a receiver pairing slot (0x01). For product IDs not on the receiver list it tried 0xFF first. The G305's receiver is a HID++ 1.0 device and answers that probe with a HID++ 1.0 error reply (0x8F, "invalid command") rather than silence. The old probe counted any reply — including an error reply — as "a mouse is here", so OpenMouse locked onto device index 0xFF and every later request failed with "The mouse rejected that setting (HID++ 1.0: invalid command)."
• Fix: resolveDeviceIndex() now only treats a successful root feature query as proof of a HID++ 2.0 mouse; a receiver's HID++ 1.0 error reply no longer counts as "answered", so probing falls through to the pairing slot and finds the mouse. Only genuine HID++ 2.0 mice answer the root probe successfully, so direct-connect mice (G402, G403 HERO, etc.) are unaffected. Also added 0xc53f (Nano Lightspeed 1.1, G305) and 0xc543 (Lightspeed 1.2) to LOGITECH_RECEIVER_PRODUCT_IDS so the receiver slot is probed first.

Fix 2 — hide surface / LightForce controls the G305 does not have

The G305 exposes Mode Status 0x8090, but only the power-mode half is meaningful: the status1 byte that would carry the gaming-surface and LightForce fields is reserved and reads 0. That decoded as Gaming surface "Auto" and LightForce "Optical" and offered both controls (plus the sensor card).

• Fix: keyed on the firmware model id 407400000000 (captured from the device's connection diagnostics), added to MODE_STATUS_POWER_ONLY_MODEL_IDS the same way the G309 was handled in ff6e1cd. gamingSurfaceMode and lightforceSwitchMode are now reported null and both cards stay hidden.

Files changed

• src/devices/logitech/hidpp.ts — strict index probe; receiver PIDs; G305 model id.
• src/devices/vendors.ts — receiver PIDs for the picker.
• src/devices/logitech/hidpp.test.ts — G305 mode-status assertion.
• src/devices/logitech/TESTING.md — receiver identifiers and G305 checklist.

Verification

• Confirmed on hardware with a G305 LIGHTSPEED (receiver 046d:c53f): connects and reads DPI, polling rate, and battery correctly; gaming-surface, LightForce, and lift-off cards stay hidden.
• npm run check (build + 197 tests) passes.

[jazzstack]

verified on hardware by discord user Gandhi (284314794788519938)