You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
v3.26.2 — ensure_channel waits for the requested encoding
ensure_channel() now waits for radiod to report the requested
encoding.create_channel() sends OUTPUT_ENCODING as a separate
command after the create, and ensure_channel() returned the first status
it saw, which could predate that command. A fresh F32 channel then came
back labelled S16, and RadiodStream decoded 4-byte floats as 2-byte
ints: noise-like samples on the 1/32768 grid, rms about 0.58 of full
scale, with "completeness 100%". codar-sounder ran this way on AC0G-ND and
UACNJ and heard no CODAR all day; with the encoding settled, the same
channel showed the Rutgers 13 MHz radars at 32 dB. ensure_channel() now
re-polls until the reported encoding matches the request, within its timeout. If radiod still reports another encoding at the deadline, it
logs an error and returns radiod's value, so a stream decodes what is on
the wire. Checked live on UACNJ's radiod: a new F32 channel first reported
encoding 2 and settled to 4.
The comments that blamed radiod for "silently downgrading F32 to S16"
(control.py, stream.py) now name the stale status instead. The bee1
2026-05-15 T6/TSL3 incident they cited was very likely the same race.