[ET10x00] Power Timers & Start Of Summer Time

[EMJB]

I have recently noticed my power timer was not working as expected, and on investigation the displayed time was 60 minutes later than I had originally set.

It would appear that when summertime started the timer continued to fire at the same GM time, and the displayed time was offset by 60 minutes, whereas I would have expected the displayed time to be unchanged and the firing to occur 60 minutes earlier in terms of GMT. However I did upgrade from 4.1.27 to 5.0.008 at around the time I noticed this, so perhaps it was something to do with the upgrade.

EMJB

 

[birdman]

It would appear that when summertime started the timer continued to fire at the same GM time, and the displayed time was offset by 60 minutes

Seems possible if the timer just resets itself to add 86400 seconds rather than calculate the hh:mm offset for the next day. However, if that were the case I'd have expected others to have seen it by now....

What are the timer details (type, time, repeat frequency etc.).

 

[EMJB]

@ birdman: Sorry for not replying sooner - have been tied up with other matters. It is a power timer to switch to deep standby late every evening, and wake to normal standby early evening to allow an EPG update & the AutoTimer plug-in to run.

However I think I now understand what caused the problem, based on the premise that the timer operates on Unix Time, and a 60 minute change is introduced at the start and end of summer time to give constant clock times. I upgraded to the (then) latest Vix version soon after the start of summer time, but my last backup was before the start of summer time. When restoring the settings from that backup it would have picked up a winter Unix time and skipped the process that updated to summer time.

If my theory is correct, perhaps people need to be encouraged to do a new backup just after the summertime changes to avoid this problem, especially if it can occur with other settings.

EMJB

 

[birdman]

If my theory is correct...

Then there's a bug - either in the logic or in the code.