Skip to content

v9.9.12

Latest

Choose a tag to compare

@Lexonight1 Lexonight1 released this 31 Aug 20:38
· 1 commit to main since this release

v9.9.12

Fedora installs could fail outright, and that is fixed. The RPM shipped its
own copy of the sounddevice library. Fedora now ships one too — and both
claimed the same files, which RPM refuses to allow. If you already had
python3-sounddevice on your system for any reason, installing or upgrading
TRCC failed with a file conflict. The package now depends on Fedora's copy
instead of carrying its own.

Your cooler may have been telling you it was unsupported when it is not.
Panels reporting PM 66 — the ELITE VISION, LF14 and LD7 — printed a warning in
the log asking you to file a bug report about a device that is fully
catalogued. The picture was always correct; only the accusation was wrong.

The week-start setting had no effect. On LC2 panels you can choose whether
the week begins on Sunday or Monday, from the app, the terminal or the API. The
choice was saved and reported back correctly, and the clock ignored it — it was
never passed to the part that draws the display. It works now.

Auto-start can be repaired without turning it off and on again. If TRCC was
moved or reinstalled, the auto-start entry kept pointing at the old location and
silently stopped working. trcc system autostart refresh re-points it. It will
not switch auto-start on if you have it off.

New: trcc system memory-slots. Prints your DIMMs — size, type, speed,
manufacturer — and, on Linux, the memory timings the LC1-style panel displays.
Useful to paste into a bug report about that panel, which previously had no way
to show what it was reading.

Also in this release, for people driving TRCC from outside the app: the
autostart and system-status endpoints now report the same answers as the app
itself rather than computing their own, and four trcc system autostart
subcommands that existed but appeared in no documentation are now in man and
the CLI reference.

Two fixes for hardware nobody here can test. A panel that identifies itself
as 0x36 was being rejected by the ALi/F5 wire even though the official app
accepts it, and the LY1 variant (0416:5409) read its panel-model byte from
the wrong slot. Both were read out of the official app's own code and both are
covered by tests, but neither has run on the hardware it targets. If you own one
of those coolers, we would be glad to hear whether this release changes
anything.