A two-week release driven by your bug reports. APRS now hears what your TNC hears, clicking a spot lands the radio exactly on frequency the first time, and a new Clock layout fills the screen with a big digital clock between nets.
🕐 Headline: Clock Layout and Idle Clock
A layout that is a big digital clock and nothing else: UTC very large with seconds, local time with its zone beneath, the date, and your callsign and grid in the corner — in your theme and font. Tap the big time to make local the large one; it remembers. Pick it like any layout, add it to a Scene Rotation, or let it appear on its own: Settings → Display → Idle Clock shows it after 2 to 60 minutes with no input, and the first touch, mouse movement or key press brings your previous layout straight back. For shack TVs and club displays during down time.
📡 APRS Hears Everything Your TNC Hears (#1194)
Neither of our APRS parsers understood Mic-E, the default beacon format on Yaesu and Kenwood radios, so every FT-5D, FT-3D, FTM-400 and TH-D74 heard over RF was silently dropped — and that was only half of it. Third-party frames (igates re-transmitting internet traffic over RF) were dropped whole, object reports had an offset bug that read the timestamp as the latitude, and compressed positions were never supported. All of it decodes now, along with items and position ambiguity; objects are keyed by their own name so a digipeater's repeater object no longer overwrites the digipeater. The server and the Rig Bridge TNC path share one parser with a guard test so they cannot drift. Verified against the reference aprslib parser and against an eleven-hour DireWolf capture from Michael, KE8PLM: 164 stations decoded, 163 in OpenHamClock, and the last one fixed since. Reported by Jason Taggart and KE8PLM.
🎛️ Clicking a Spot Lands the Radio on Frequency, First Time (#1195)
Clicking a POTA, SOTA or cluster station used to leave Yaesu rigs close but not on the spot, and a second click fixed it. We were sending the frequency and the mode at the same time; Yaesu radios move the dial by about 1.4 kHz when the mode changes, so whenever the mode landed after the frequency the radio ended up off by that much. The tune now does what operators' own macros do: mode first, then frequency, then the frequency once more after the mode settles. Reported by KE8PLM (FTDx10, FT-991A), confirmed by onzarob (FT-950). Fixed from the symptom rather than on a bench radio — tell us if your rig still lands short.
📊 The Band Plan Stops Crying Wolf (#1193)
Tune to 14.105 and the privilege bar said "Outside Amateur Extra privileges". The band plan's SSB segments on 20, 15 and 10 metres began below where US phone privileges actually start, so those digital stretches were flagged for every class. They are Data segments now, the bar draws adjacent same-type segments as one block, the tooltip names the emission type ("No Phone privileges here") instead of implying your licence is short, and 60 metres is modelled as it really is under the February 2026 rules — the 5351.5–5366.5 kHz band plus four channels — instead of one continuous range that said yes at 5340 kHz.
🌤️ Kiosk Weather Stays Current (#1197)
On an always-on display the DE weather panel could go hours stale: the browser fetches Open-Meteo directly, and the poll interval was a two-hour leftover from a server cache that no longer exists on that path. It is 15 minutes now. Diagnosed and fixed by brstiefel1.
🔧 WSJT-X Decodes No Longer Freeze with Rig Bridge (#1200, #1201)
Self-hosters running Rig Bridge could see the WSJT-X decode list freeze while the server kept receiving fresh decodes: the hook flipped to "local mode" on the first Rig Bridge message of any kind and then stopped polling for good. It now switches only on real WSJT-X data and keeps the fallback alive. Also gone: a console error on every VOACAP heatmap fetch. Both by SQ8BWM, in his first two PRs.
📥 Rig Listener Downloads Are Back
The rig-listener Windows, Mac and Linux executables had quietly stopped attaching to releases since 26.4.0: the Intel Mac build was pinned to a GitHub runner retired in December, so every release sat in a queue for a day and gave up. Fixed, and the 26.8.0 release was back-filled. Intel Mac builds end when GitHub retires its last x86 runner in late 2027.
🙏 Thanks
Michael, KE8PLM, for the DireWolf comparisons and three precise reports; Jason Taggart for the Mic-E decoder that started it; SQ8BWM; brstiefel1; onzarob.
73 de K0CJH
Docker image: ghcr.io/accius/openhamclock:26.9.0
docker pull ghcr.io/accius/openhamclock:26.9.0
Docker image: ghcr.io/accius/openhamclock:26.9.0
docker pull ghcr.io/accius/openhamclock:26.9.0