When EKOS is connected to a remote indiserver, and that one vanishes, EKOS (and with it KStars) will hang indefinitely, not reacting to any input other than sending a terminate signal.
Restarting a new indiserver instance at the remote location will not bring it back to life.
This is with the latest kstars (3.3.5-706_g669835dae), so it's not due to the 'Sorry' dialog issue as suspected
here
No, this is already the current git, so it's not the 'Sorry' dialog, but something else.
Also, with this bug it is not using CPU. It just seems to wait for some answer from indiserver, without any(?) timeout. At least I had left it sitting there for at least half an hour
Bummer. I cannot reproduce it anymore. Wonder how I goofed up this time.
I had started KStars on the mount computer, and there started a local indiserver. I connected to that remotely from my laptop, and then ended the session on the mount computer, which left the mount computer kstars running, but hang the laptop one. But as I said, now I also get the disconnect notice and everything continues running.
I confirm this issue, but I think it depends on how your remote server vanishes. I observed multiple times that network requests would just hang there blocking, causing Ekos to freeze. It also happened on indiserver side when my wifi connection was instable, causing even local stack requests to stuck. Though I have no real clue where to look at, I didn't reproduce the situation with 4.x kernels, only 3.x ones. I also observed that my older 3.x gentoo was unable to restart the wifi chip of the remote mini-PC when rebooting while the newer 4.x lubuntu 18 was able to. Of course, that is very very difficult to reproduce.
Yesterday I had the case that when covering the telescope I accidently unplugged power to the INDI computer. Symptom in KStars was the same: It just hung dead. I rebooted the computer, and once it was up (reachable by network) KStars fired up the 'lost connection' dialog.
But I think the problem is exotic enough to give it low priority