Fix FT-2000 CAT: SWR meter dead during TX (only first of two coalesced replies parsed)

[patrickrb]

Summary

Yaesu38Rig ("YAESU FT-2000 series") loses its SWR meter reading during TX. On every meter poll the rig is asked for two meters back-to-back — RM4; (ALC) then RM6; (SWR) — and the transport routinely delivers both replies coalesced in a single onReceiveData chunk ("RM4nnn;RM6nnn;"). The old parser handled only the first ;-terminated command and re-buffered the remainder with its terminator; the next poll's clearBufferData() then wiped that carry-over, so the second (SWR) reply was permanently dropped.

Net effect for FT-2000 users: the SWR bar never moves during transmit, and the high-SWR safety alert never fires.

This is the same root cause already fixed for the FT-891 in #665, on a distinct, still-affected rig.

Root cause

onReceiveData was a one-command-per-read parser:

Yaesu3Command yaesu3Command = Yaesu3Command.getCommand(buffer.toString());
clearBufferData();
buffer.append(s.substring(s.indexOf(";") + 1)); // rest kept WITH terminator, parsed never

Any second complete command in the same chunk was left in the buffer with its trailing ; and then destroyed by the next readMeters()/readFreqFromRig() clearBufferData(). Because the SWR read is always the second of the two, it is the one consistently lost.

Fix

onReceiveData now drains every complete ;-terminated command from the accumulated buffer and carries only the unterminated tail into the next read. The framing decision is extracted into a pure, package-private static helper so it can be unit-tested without the rig's Timer-scheduling constructor:

• splitCommands(String) -> Frames — pure String → {List<String> commands, String remainder}.
• processCommand(String) — the existing per-command FA/FB/RM handling, unchanged in behavior.
• Collapsed a latent double Yaesu3Command.getFrequency() call.

The > 1000-char no-terminator overflow guard is preserved.

Kept self-contained (no new shared class) deliberately: #665 introduces a shared CatLineSplitter but is still open, so adding my own copy would collide. Once #665 merges, the remaining identical Yaesu gen-3 siblings (FTDX-450 / DX10 / Wolf-SDR) can adopt the shared helper as follow-ups.

Testing

• New Yaesu38RigTest — 8 pure-JVM cases (no Robolectric; the helper is Android-free): single command, two coalesced meter replies both drained (the regression), complete + partial tail, no terminator, empty input, leading terminator, split-across-reads reassembly via remainder carry, three-commands-plus-tail.
• ./gradlew :app:testDebugUnitTest — full suite green; Yaesu38RigTest 8/8.

Risk

Low and localized to one rig's read path. No protocol/DSP change; frame semantics are unchanged for the common single-command case, and the fix only adds recovery of previously-dropped coalesced commands. No hardware was available, so validation is via the unit tests and the identical-to-#665 root cause.

Related

• Same root cause as Fix FT-891 CAT: only the first of two coalesced replies parsed (SWR reading lost, next parse poisoned) #665 (FT-891); this covers the FT-2000. Identical siblings Yaesu38_450Rig, YaesuDX10Rig, Wolf_sdr_450Rig remain follow-ups.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 36.79%. Comparing base (ac169cc) to head (e8ba664).
⚠️ Report is 4 commits behind head on dev.

Additional details and impacted files

@@             Coverage Diff              @@
##                dev     #667      +/-   ##
============================================
+ Coverage     36.73%   36.79%   +0.06%     
- Complexity      197      207      +10     
============================================
  Files           216      218       +2     
  Lines         26888    26994     +106     
  Branches       3287     3311      +24     
============================================
+ Hits           9877     9933      +56     
- Misses        16785    16829      +44     
- Partials        226      232       +6     

Flag	Coverage Δ
android	15.29% <ø> (+0.25%)	⬆️
native	9.93% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.
see 8 files with indirect coverage changes

🚀 New features to boost your workflow:

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

[Copilot]

Pull request overview

Fixes FT-2000-series (Yaesu gen-3) CAT reply parsing so coalesced ;-terminated replies (notably ALC + SWR meter polls arriving as "RM4...;RM6...;" in a single read) are fully drained and processed, preventing the SWR meter from going stale during TX.

Changes:

• Updated Yaesu38Rig.onReceiveData to split accumulated input into all complete ;-terminated commands plus an unterminated remainder, and to process each command in order.
• Extracted per-command handling into processCommand(String) and framing logic into a pure helper splitCommands(String) with a small Frames carrier type.
• Added a new pure-JVM unit test class Yaesu38RigTest covering coalesced replies, partial tails, leading terminators, and multi-command inputs.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
ft8af/app/src/main/java/com/k1af/ft8af/rigs/Yaesu38Rig.java	Drains and processes all complete CAT commands per receive chunk, preserving only the unterminated tail to avoid dropping coalesced replies.
ft8af/app/src/test/java/com/k1af/ft8af/rigs/Yaesu38RigTest.java	Adds unit tests that lock in correct framing behavior, including the ALC+SWR coalesced-reply regression.

---

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