fix(hid): decode Contour ShuttleXpress/ShuttlePro v2 buttons from bytes 3-4. Principle XI.

[sq9fk]

Summary

Fixes #5927.

Both Contour parsers read the button bitmask from the wrong bytes. A raw hid_read() capture from a ShuttleXpress (VID 0B33 / PID 0020) shows buttons 1–4 in buf[3] bits 4–7 and button 5 in buf[4] bit 0. These are the ShuttlePro v2 button 5–9 positions. Because the parser only scanned bits 0–4 of buf[3], only physical button 1 was seen, and it was reported as button 5. ShuttleProV2Parser built its mask from buf[2], which is always 0, and buf[3], which shifted buttons 1–8 to 9–16 and never read 9–15.

• ShuttleXpress: btns = (buf[3] >> 4) | ((buf[4] & 0x01) << 4), so the existing scan loop and 1-based numbering stay unchanged.
• ShuttlePro v2: btns = buf[3] | (buf[4] << 8).
• Both parsers: require len >= 5. HidEncoderManager::m_buf is reused across reads, so a short report would otherwise decode a stale buf[4] as a phantom press.
• Layout comments: fixed (bytes 3–4, not 2–3).
• New tests/hid_device_parser_test.cpp: registered in tests/tests.cmake and built unconditionally, because the parsers have no hidapi dependency and only need HAVE_HIDAPI for their #ifdef. It uses the captured reports from the issue:
  • all 5 Xpress buttons press and release as 1–5,
  • a short report is ignored,
  • ring-only reports produce no event,
  • the jog still decodes, including wrap,
  • all 15 Pro v2 buttons press and release as 1–15.

Out of scope, and to be done in a follow-up PR as noted on the issue: when two buttons change in the same report, only the first edge is reported.

Constitution principle honored

Principle XI: the fix is demonstrated. hid_device_parser_test fails 3 checks when built against the parser from main (Xpress buttons, short-report guard, Pro v2 buttons) and passes all checks with this change.

Test plan

• Local build passes (cmake --build build, Windows 11, MSVC 19.44, Qt 6.8.3, Ninja)
• Behavior verified on real hardware: ShuttleXpress on a FLEX-6500 (fw 4.2.20). Key 1–5 mapped to distinct actions, each physical button 1–5 fires its own action, and the jog still tunes.
• ctest -R hid_device_parser_test passes; the same test fails 3 checks on the old parser.
• Existing tests pass (CI)
• Reproduction steps documented in ShuttleXpress: only 1 of 5 buttons works – parser reads the wrong button bits #5927
• ShuttlePro v2: layout inferred from the Xpress capture, not tested on hardware (I don't own one).

Checklist

• Commits are signed (docs/COMMIT-SIGNING.md)
• No new flat-key AppSettings calls (no settings touched)
• Code is clean-room: the layout was derived from my own USB HID capture of my device
• All meter UI uses MeterSmoother (no meter UI touched)
• Documentation: no user-visible doc change needed (the help page already lists five ShuttleXpress buttons)
• Security-sensitive changes: n/a

🤖 Generated with Claude Code

[aethersdr-agent]

Welcome to AetherSDR, @sq9fk! Thanks for your first pull request.

A few things that might help:

• Our CONTRIBUTING.md covers coding conventions and the PR process
• CI will run automatically — if it fails, I'll post a comment explaining what went wrong
• Jeremy (KK7GWY) reviews all PRs before merge
• Come say hello on the AetherSDR Discord — most of the community hangs out there, and merged contributors get the Contributors role

If you have questions, feel free to ask here or in Discussions.

— AetherClaude (automated agent for AetherSDR)

[jensenpat]

Verdict

No code blockers. Reviewed 4c908a50825bef7e3c1e89336bbfb191c9a2bde5. This is the right fix for #5927. A clean merge recommendation waits on the CI run that was sitting in action_required and is now in progress. This comment is not an approving review.

Issue fit

#5927 asks for all five ShuttleXpress buttons to report as 1–5. The suggested pack is (buf[3] >> 4) | ((buf[4] & 0x01) << 4), with len < 5. The triage note on the issue also asked for the same guard on ShuttlePro v2 (buf[3] | (buf[4] << 8)), corrected layout comments, and a parser test from the captured reports. This PR does those and nothing else.

Scope

File	Change
src/core/HidDeviceParser.cpp	Xpress and Pro v2 button bytes, len < 5, comments
tests/hid_device_parser_test.cpp	Captured Xpress reports, short report, jog/wrap, inferred Pro 1–15
tests/tests.cmake	hid_device_parser_test registered. No socket, no AETHER_SETTINGS_CONSUMERS entry (the test does not touch settings)

Shared HID parsing only. No radio command, no family branch, no settings key.

Verification

hid_device_parser_test was compiled and run from #5933 (4be1554c), which contains this parser change. All of this PR's checks passed there. Replacing the Xpress mask with the old buf[3] read failed the button checks and left the jog and Pro checks passing. The tree was restored and the test passed again.

The test does not open a socket. CI had not run (first-contributor action_required); those workflows are approved and running on this SHA. The per-PR gate does not execute a new ctest, so the local run is the coverage for this target.

Stack

#5932 and #5933 are stacked on this commit. Squashing this PR onto current main and then squashing #5932 conflicts in HidDeviceParser.cpp and tests/hid_device_parser_test.cpp. The three changes land cleanly as one squash of #5933. I will do that once #5933's checks are green, and close this PR as included, rather than merge it out from under the stack.

Non-blocking: ShuttlePro v2 is still inferred from the Xpress capture, as the issue says. No hardware here.

[jensenpat]

Approving 4c908a50825bef7e3c1e89336bbfb191c9a2bde5. CI is green on this SHA: build, check-macos, check-windows, Static checks, and Sanitizer option configures. The earlier comment stands for the parser fix and the local hid_device_parser_test run.

Landing this PR on its own. #5933 is blocked on static checks, so the stack tip is not the merge. #5932 will need a rebase onto main after this squash; squashing it as-is conflicts in the parser and the new test.