Recording status erroneously remains on

[keith_leitch]

I've been trying to catch a moment to log a problem across which I may have accidentally stumbled. I thought I would describe it here in case someone can explain it without a log.

I was running a polling process that used the CLI to check for current and pending recordings, and periodically restart the recording service if none were imminent. There are some deeply involved reasons that I wish to do this, and I can bore you with them in a different message if you wish.

In the meantime, here is what happened: I started to notice that my programs were often being cut off at the end. Looking at my own log, I saw that the polling process was restarting the service during post-padding time; not always, but often. Since Australian listings are ridiculously inaccurate (often as much as 20 minutes off), this would sometimes interfere with the program.

I tried to duplicate this from the command prompt, and couldn't. The CLI always accurately reported a recording in progress, even during postpadding time. However, now that I have disabled my polling process, the problem has gone away. So, it's a mystery.

[keith_leitch]

I apologise for this, as I now believe it was an error I had made. When I changed methods of populating my EPG, I didn't realise that the system had retained a memory of recurring recordings with no padding time. This left me with two recordings of several items, one with padding and one without. It seems that the recording service must have defaulted to the one without.

I will keep an eye on things, but it seems likely now that the CLI was blameless.