Debug: 0.01 km/h raw FTMS speed step on Horizon treadmill (not for merge)

[cagnulein]

Context

Related to the Gmail thread "Issue with controlling treadmill" — the user reports issues controlling their Horizon treadmill's speed.

This is a diagnostic build, not intended to be merged. Its only purpose is to probe how the treadmill firmware reacts to the smallest possible speed step, so we can correlate each button press with the treadmill's observed behavior and narrow down the control bug.

What it does

This build is always in debug mode (no configuration or env var needed — just install and use):

• Each tap on the home form's speed + button requests currentSpeed + 0.01 instead of the normal step (0.5/1.0 km/h).
• In horizontreadmill.cpp, bypasses the 1-decimal rounding and minStepSpeed threshold that would otherwise silently discard such a tiny delta before it reaches forceSpeed().
• Logs every outgoing BLE write with a timestamp and the raw hex bytes (qDebug), for both the standard FTMS branch and the two Horizon proprietary-protocol branches, so the exact command sent can be matched against what's observed on the treadmill console.
• Note: the Horizon proprietary protocol (gattCustomService) only has 0.1 km/h resolution, so a 0.01 step won't change the bytes on that branch — the log makes this explicit. The test is meaningful only if this specific unit uses the FTMS branch (gattFTMSService).

How to test

1. Build the Android debug APK via the android-build job in the CI GitHub Action (runs automatically on this PR, artifact name fdroid-android-trial), or build locally.
2. Install the APK and launch the app — no setup needed, the debug behavior is always on in this build.
3. Connect to the Horizon treadmill, tap the speed + button repeatedly, and note down what the treadmill actually does after each tap.
4. Pull the debug log (adb logcat on Android) and match each [QZ_DEBUG_SPEED_001_STEP] line (timestamp + raw bytes) to the corresponding observed treadmill reaction.

Test plan

• homeform.o and horizontreadmill.o compile cleanly (verified locally)
• Manual test against a real Horizon treadmill to confirm raw bytes are actually sent and observe firmware reaction

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.