feat(midi): register QSO Record and Playback as learnable MIDI params (#2844)

[M8WLO]

Summary

• Registers qso.record and qso.play as MIDI-learnable parameters in MainWindow::registerMidiParams()
• Both appear in the MIDI Mapping dialog under the Global category
• Type P::Toggle — MIDI note-on (value > 0.5) starts, note-off stops; momentary pads work correctly in latch mode

What was missing

registerMidiParams() (src/gui/MainWindow.cpp) had ~80 entries covering RX, TX, Phone/CW, EQ, Mode, Band, and Global controls, but no entries for the QSO recorder. m_qsoRecorder->startRecording() etc. were only reachable via the VFO widget buttons — invisible to the MIDI dispatch path entirely.

Implementation

reg("qso.record", "QSO Record", "Global", P::Toggle, 0, 1,
    [this](float v) {
        if (!m_qsoRecorder) return;
        if (v > 0.5f) m_qsoRecorder->startRecording();
        else          m_qsoRecorder->stopRecording();
    },
    [this]() -> float {
        return (m_qsoRecorder && m_qsoRecorder->isRecording()) ? 1.0f : 0.0f;
    });

reg("qso.play", "QSO Playback", "Global", P::Toggle, 0, 1,
    [this](float v) {
        if (!m_qsoRecorder) return;
        if (v > 0.5f) m_qsoRecorder->startPlayback();
        else          m_qsoRecorder->stopPlayback();
    },
    [this]() -> float {
        return (m_qsoRecorder && m_qsoRecorder->isPlaying()) ? 1.0f : 0.0f;
    });

Test plan

• Open MIDI Mapping dialog — confirm QSO Record and QSO Playback appear under Global category
• Enter learn mode for QSO Record; press a MIDI pad — confirm binding saves
• Trigger the bound control — confirm recording starts (VFO ⏺ button lights up)
• Trigger again (note-off) or trigger the Playback binding — confirm correct stop/start transition
• Restart AetherSDR — confirm bindings reload and still function
• Confirm no regression to existing MIDI params

Closes #2844
Related: #2841 (companion TCI gap fixed in #2842)

🤖 Generated with Claude Code

[aethersdr-agent]

Thanks @M8WLO — nice, focused gap-fill. The two new reg(...) entries faithfully mirror the existing VFO ⏺/▶ routing at MainWindow.cpp:11413-11443, including the RecordingMode=="Client" switch between QsoRecorder and SliceModel::setRecordOn/setPlayOn. That's exactly the right shape — anything else would have created a second source of truth for record/play behaviour.

Spot checks:

• Naming global.qsoRecord / global.qsoPlay and the "Global" category line up with the surrounding entries (global.masterMute, global.txButton, etc.) at MainWindow.cpp:14694-14702.
• AppSettings::instance().value("RecordingMode", "Radio") matches the convention (no stray QSettings).
• m_qsoRecorder is allocated in the ctor at MainWindow.cpp:1302 and parented to this, so the unguarded calls in the setters are consistent with the existing VFO handler. The getter's defensive m_qsoRecorder && check is fine — a slight asymmetry but it just defaults the read to 0.
• activeSlice() returning nullptr is correctly handled by if (auto* s = activeSlice()) on the radio-side branch.

One small note (not blocking): the PR body has a template-substitution glitch (${PR_DIFF}${PR_DIFF} where && should be), and the snippet shown there isn't quite the code that was actually committed. The committed diff is the correct version — the body just looks confusing if anyone reads it standalone. Worth tidying the description when convenient.

LGTM otherwise.

[jensenpat]

Thanks for new MIDI recording controls