Fix: Reset ESP32 RTC clock on software reset

[rightup]

Updated ESP32RTCClock::begin() to treat software reset (ESP_RST_SW) the same as power-on reset (ESP_RST_POWERON), ensuring the clock resets to the default time (15 May 2024, 8:50pm) when triggered via the CLI reboot command or any other software-initiated reset.

No personal issues with a good old-fashioned power cycle, but some have reported their clock getting accidentally set far into the future. Probably less funny when the device is zip-tied 8m up a tree ... not exactly reset-button friendly!

[rightup]

Hey, just a note: I used Fix: in the commit message, but it’s possible the original behavior was by design, so feel free to treat this as a merge suggestion rather than a bug fix, depending on intent.

Either way, resetting the clock on soft reboot feels like the safer option in the field, especially when devices are zip-tied to trees 😅

[ripplebiz]

Yeah, this one is a little tricky. It was intentional to keep the ESP system time across soft resets, and I seem to recall it being handy in various contexts. But am struggling to rummage through my memory and figure out if this really should be changed...

[rightup]

Totally fair, I figured this might’ve been intentional, which is why I hesitated on calling it a “fix”.

I haven’t personally seen the time set incorrectly, but I’ve noticed a few user reports where the clock jumped a few hours into the future. If that does happen, the only way to reset it is via a power cycle or physically pressing the reset button — which isn’t ideal when the device is mounted somewhere creative.

No problem at all — happy to drop this if it risks interfering with other use cases. I know timing in embedded is always tricky… there’s a reason even Doc Brown had issues with it. 🤣

[recrof]

if this is not the best solution, can we introduce another command that will reboot with clock reset?