Open tehn opened 6 years ago
short term solution.
sudo systemctl stop norns-crone
sudo systemctl start norns-crone
sudo systemctl stop norns-crone
doesn't work.
we'll have to do
killall sclang
killall supernova
sudo systemctl stop norns-crone
doesn't work.
@tehn Why doesn't that work?
@simonvanderveldt @tehn it works for me
three more points:
it might be cleaner to shutdown SC from within SC (s.quit
, ("kill "++thisProcess.pid).unixCmd
). this would also allow a slick fadeout of the audio instead of a hard stop.
when restarting, will have to go through the handshake procedure as in audio.restart()
. beware of cairo thread safety for the audio connect animation (nothing else can be using the screen)
and of course the real difficulty is detecting timesync. @artfwo proposed a manual method of having a thread in matron
that checks the system time delta. (this, i could help with.) other possible solutions involve adding systemd services, which is beyond me.
oh indeed it works, i made a mistake in my launching earlier.
slick fadeout
!!!
would be nice on shutdown as well.
I've run into reproducible screen lock ups, button latency and audio drop outs when I turn on wifi. I'm worried that all these hard resets might make the filesystem on the eMMC get corrupt. So I've been doing a reboot via SSH, which doesn't actually reboot the screen or wifi.
I wonder if disabling NTP will fix this problem? Now that I have a shell, I'm tempted to do "normal linux stuff".
@lazzarello when you turn on wifi and NTP sync's you MUST reboot crone. otherwise you'll experience problems.
i'm preparing an update which disables crone on wifi connection, and implemented a menu option to reboot crone.
fyi: i'm running into problems with the new wifi logic.
failed
thanks for the report. i need to prioritize a wifi overhaul.
the only thing that changed is that upon wifi activation crone is shut down (to be manually re-activated after connection is established.) i should add a message when activating wifi that explains this.
i'm not sure why it would take two attempts, and why wifi is still saying failed, however.
@tehn Why was this closed? I don't think we have a solution for this yet, right?
this was closed in the great-ticket-purge at the march meetup. likely as it wasn't reported as a problem for some time.
have there been issues with this? (AFAIK, no?)
It still happens every now and then that someone runs into this. Maybe we should keep this open keep track of it? Although realistically there isn't much we'll be able to do about it ourselves.
Mentioned by @lazzarello: ” I wonder if disabling NTP will fix this problem? ”
This is probably one way to address this. Or have people manually sync upon confirming that the script/engine is restarted or something (weird ”fix” but would probably work anyhow).
And this is messy still - any engine and script using a high rate polls stall upon time change. I think there’s reason to reopen this.
ah, i forgot about the poll-heavy crash on startup w/ wifi attached.
sclang crashes when time/date change by a substantial amount