Fix GuoHe Q900 frame sync: require four consecutive 0xA5 bytes

[patrickrb]

Summary

GuoHeQ900Rig.checkHead() misidentifies the frame sync header, which intermittently corrupts frequency tracking on the GuoHe Q900 (and its Xiegu-family siblings that share this framing).

Every GuoHe frame begins with four consecutive 0xA5 sync bytes (see GuoHeRigConstant — PTT_ON, USB_MODE, READ_FREQ, etc. all start A5 A5 A5 A5 <len> …). The old checkHead counted 0xA5 bytes anywhere in the read buffer and returned the moment the 4th was seen:

if (data[i] == (byte) 0xa5) {
    count++;
    if (count == 4) return i + 1;   // never resets on a non-0xA5 byte
}

Root cause

A status frame (cmd 0x0b) carries two big-endian VFO frequencies at buffer[5..12], and those bytes are frequently 0xA5 (common HF frequencies land on it). USB-serial reads don't respect frame boundaries, so a read routinely delivers a prior frame's 0xA5-bearing tail followed by the next frame's sync run. The old counter reached 4 partway through those scattered/adjacent 0xA5 bytes and returned an index pointing into the sync run.

onReceiveData then read a 0xA5 as the length byte:

int len = data[startIndex] + 1;   // (byte)0xA5 + 1 == -90
buffer = new byte[len];           // new byte[-90] -> NegativeArraySizeException

The exception is swallowed by onReceiveData's try/catch, so it is not an app crash — but it aborts framing before clearBuffer(), leaving stale partial-frame state and silently dropping the frequency update. Symptom: the app's displayed frequency intermittently fails to track the GuoHe rig's VFO.

Fix

Reset the run counter on every non-0xA5 byte and return the first byte after a run of ≥4 consecutive 0xA5 — the true length byte:

if (data[i] == (byte) 0xa5) {
    count++;
} else if (count >= 4) {
    return i;      // first non-sync byte = length byte
} else {
    count = 0;     // sync must be consecutive
}

This also:

• correctly skips a stray leading 0xA5 (e.g. a previous frame's CRC low byte) that would otherwise make five in a row and, under the old code, still land on a 0xA5 length byte;
• returns -1 ("no complete header yet") instead of an out-of-range index when a sync run lands at the very end of a read.

checkHead is now package-private static (it referenced no instance state), so the framing logic is unit-tested directly without standing up the rig's Timer/connector.

Testing

New GuoHeCheckHeadTest (pure logic, no Robolectric needed) — 6 cases:

• well-formed frame → length-byte index
• non-consecutive 0xA5 payload bytes are not mistaken for sync
• realistic status-frame tail (0xA5 freq bytes) + next sync → finds the length byte, never a 0xA5

• 4 consecutive 0xA5 skips the whole sync run

• no header → -1
• sync run at buffer end → -1 (no overrun)

Four of these return the wrong index against the old implementation, so the suite fails-before / passes-after.

./gradlew :app:testDebugUnitTest --tests 'com.k1af.ft8af.rigs.*'
BUILD SUCCESSFUL   # GuoHeCheckHeadTest 6/6, all rigs tests green

Risk assessment

Very low. The change is confined to a single pure byte[] → int helper; no protocol/DSP/native code touched, no behavior change for correctly-framed input (the well-formed case returns the identical index as before). Strictly more robust for mis-framed input. No performance impact (same single linear pass).

Affected platforms

Android only — GuoHe Q900 CAT/frequency read path.

[Copilot]

Pull request overview

This PR fixes GuoHe Q900 (and related Xiegu-family) frame sync detection by updating GuoHeQ900Rig.checkHead() to only accept runs of 4+ consecutive 0xA5 bytes as the sync header, preventing mis-framing when 0xA5 bytes appear in payload (e.g., status-frame frequency fields).

Changes:

• Update checkHead(byte[]) to reset the counter on non-0xA5 bytes and return the index of the first non-0xA5 byte after a >=4 sync run (the length byte).
• Make checkHead static (package-private) so it can be unit-tested directly without instantiating the rig/timer.
• Add a dedicated unit test suite covering well-formed frames and common mis-framing splice scenarios.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
ft8af/app/src/main/java/com/k1af/ft8af/rigs/GuoHeQ900Rig.java	Fixes header detection to require consecutive sync bytes and documents the framing rationale.
ft8af/app/src/test/java/com/k1af/ft8af/rigs/GuoHeCheckHeadTest.java	Adds unit tests validating correct header detection across realistic stream/splice cases.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 36.62%. Comparing base (1ab0385) to head (6e29e92).
⚠️ Report is 5 commits behind head on dev.

Additional details and impacted files

@@            Coverage Diff            @@
##                dev     #624   +/-   ##
=========================================
  Coverage     36.62%   36.62%           
  Complexity      197      197           
=========================================
  Files           216      216           
  Lines         26885    26885           
  Branches       3294     3294           
=========================================
  Hits           9847     9847           
  Misses        16811    16811           
  Partials        227      227           

Flag	Coverage Δ
android	15.03% <ø> (ø)
native	9.93% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.