[tcpreplay-dev]

What

The K-Pod Options page now shows all sixteen K-Pod button macros, and macros can be imported from a K4 backup file.

K-Pod page

Sixteen slots as a table — tap column left, hold column right, Button / Name / CAT Command. Read-only: MacroDialog (Fn > MACROS) stays the only editor, so there is no second writer on the same QSettings array and two open views cannot clobber each other. The page reloads on macrosChanged, so an edit made in Fn > MACROS shows up here immediately.

The bindings themselves already worked — HardwareController emits K-pod.<n>T / K-pod.<n>H on tap and hold, MacroController dispatches whatever RadioSettings holds — but the only place to see them was MacroDialog's flat list of every macro slot in the application.

Rest of the page was rearranged to fit the dialog's default width: status and enable share a row, Device Summary is two property/value pairs per row instead of seven full-width rows, and the page scrolls.

Import from K4 Backup

No K4 CAT command reads or writes macro text. MS is Mic Setup (MSabcde;); neither the Programmer's Reference D5 in resources/k4_cat_commands.json nor Elecraft's published rev D4 documents any macro accessor.

The K4's backup utility does export them — Fn > hold BACKUP writes K4_SN<serial>/k4macros.json, one of its six selectable categories being "PF Key and K-Pod Macros". The new button reads that file.

It is a file picker and nothing more: no volume scanning, no startup probing, no guessing where a stick is mounted. Before writing, the import compares every slot against what QK4 holds and asks for confirmation, naming each macro that would be overwritten, defaulting to No. Slots absent from the file are untouched. It covers every slot the backup carries — 16 K-Pod, PF1-4, Fn.F1-8, REM ANT.

Why the slot map is explicit

The file's names are close enough to MacroIds to be dangerous: Kpod.1T against K-pod.1T, REM.ANT against REM_ANT, while Fn.F1-F8 and PF1-PF4 match exactly. A mistranslation would write macros into slots nothing dispatches and look entirely successful. Hence a lookup table rather than prefix surgery, unknown slots skipped with a log line, and tests asserting no raw backup name survives into the store.

Also

RadioSettings::setMacros() for the bulk path — one file write and one macrosChanged for the whole import instead of one of each per macro.

Testing

• tests/test_k4macrobackup.cpp, 12 cases covering slot translation, empty-command slots, unknown slots, verbatim commands, malformed JSON
• Full suite 15/15, build clean with no new warnings
• Parser verified end-to-end against a real backup stick: 17 assigned macros, all IDs translated correctly

🤖 Generated with Claude Code

[tcpreplay-dev]

Screenshot:

[tcpreplay-dev]

Closing for now — will rework this branch. Branch retained.