hamradiotools

Nexus 0.15.0

TempoFast & TempoDeep, panels you can remove, DXKeeper, and two silent data-loss bugs found. Released 21 July 2026. The current release is 1.24.0. Download it · Every version

Fixed — two ways QSOs were quietly being lost

  • A QSO rejected by LoTW was stamped "sent" and never retried.

    Nexus invokes TQSL with -x -a compliant, which sets ignore_err, so a record TQSL refuses is skipped silently and unidentified. Exit 9 (some suppressed) was mapped to Pending and exit 8 (none processed) unconditionally to Duplicate — both count as sent — and one outcome is stamped across the whole batch. The rejected QSO therefore left the unsent list permanently while never reaching LoTW. Exit 9 is now Rejected, and exit 8 stays Duplicate only when the stderr shows no rejection. Re-offering an accepted QSO costs nothing (LoTW dedupes); losing one is forever. This was never mode-specific — it could swallow any rejected record.

  • POTA park references never reached HRDLog, or anything else keying on POTA_REF.

    Exports wrote only SIG/SIG_INFO, the older overloaded convention that WWFF and special events also use. ADIF 3.1.4 added dedicated POTA_REF/MY_POTA_REF precisely to disambiguate it. Now both go out. The giveaway that this was an oversight rather than a choice: our own importer already read the dedicated fields. We were reading modern and writing legacy.

Added — panels you can actually remove

  • A panel can now be removed outright

    , not merely popped out to another window. ⊞ Panels in the Operate header: untick and it is gone — no placeholder, no window, and the decode lists and roster grow into the space. It stays gone across restarts. Removable: waterfall, Band Activity, Call Roster, Rx Frequency, Stations, Tx Messages. Because the component truly unmounts, a removed waterfall also stops its 120 ms spectrum poll — a small performance win, not only a space win. Undo last change and Reset layout ship in the same menu, so there is no state you can strand yourself in. Layout is per-surface, so a popped-out Operate window keeps its own arrangement.

  • DXKeeper (DXLab Suite) integration.

    Settings ▸ Integrations. Each logged QSO is pushed to DXKeeper's TCP Network Service. Note the field asks for the Base Port (default 52000), matching DXKeeper's own config panel — DXKeeper listens on base +1, and nothing listens on the base itself, which is why "use port 52000" is such a common report. The hint shows the resolved port live. Uploads default OFF, since Nexus already pushes to LoTW/eQSL/ClubLog/QRZ and enabling both would upload every QSO twice to four services.

  • State and Country are editable in Log this QSO.

    Both were always auto-filled from the QRZ lookup and written to the record — they were simply never shown, so correcting a misheard state meant logging the QSO and then editing it in the Logbook.

Changed — FT1 is now TempoFast, DX1 is now TempoDeep

  • The two native protocols are renamed throughout: on screen, in the logbook, in the source tree, and in the build.

    Nothing about the on-air protocols changed — grep confirms neither name ever appeared in a transmitted payload, so a station worked before the rename is unaffected.

  • TempoFast QSOs now upload to LoTW as MODE=MFSK + SUBMODE=TEMPOFAST.

    The ADIF Mode enumeration is closed, so the previous bare <MODE:9>TempoFast was rejected outright by TQSL ("Invalid MODE") — a TempoFast QSO could not have been confirmed anywhere. MFSK is the honest family, not a flag of convenience: TempoFast is 4-CPM h=1/2 BT=0.3, the same continuous-phase FSK family as FST4, which already lives under MFSK. Your local logbook still records TempoFast, because MFSK would erase the distinction from TempoDeep. Verified against live LoTW config.xml v11.34: MFSK resolves to the accepted DATA group.

  • Band-edge tones moved from Digital to Rig settings.

    The cue already fired on phone and CW identically — it was only grouped under Digital by accident.

  • POTA/SOTA spots are sortable

    (workable-now, activator, reference, band, mode), and the Sort / Band / Program / Mode filters now survive leaving and returning to the view.

Fixed — other

  • POTA/SOTA default sort was inverted

    , putting the least workable activators on top. The arrow glyph also disagreed with the list on that one key.

  • Closed a latent .bss overflow in the FT8 a7 path.

    ft8::decode_frame documented itself as "a7-inert" while passing a7_final = true, so its decode counter grew unbounded; msg0 is byte-adjacent to jseq in .bss. Unreachable in production, but one future call site away from memory corruption.

Every version