bug/crash with scheduler api

[KipK]

I'm in holiday I can't track it but seems I saw a bug if only one schedule is set ( for example a "disabled" schedule with id: 1 , without an "enabled" one )

Can't see if it's a crash or bug from here, but then common status data from create_rapi_json are not published on MQTT, I think it's the same with http, event_send is not called. Either it crashes before or something else.

[jeremypoulter]

I did have an issue with something similar, thought I fixed it, will have a look.

[KipK]

I see some events.getnext() in scheduler.cpp without checking if there's such next event first, could this be the culprit ?

[KipK]

While working on the Timers part of the new GUI, I get this issue as the UX I'm doing is creating each timer separated.

I confirm the OpenEvse-wifi crash if there's only one timer.
Module crash in the wollowing 30 secs.

It doesn't happend if there's more, whatever the enable/disable parity is respected or not.

[KipK]

@jeremypoulter , do you have any hints on this issue yet ?
I have no debugger here, and no access to the esp32 in production to fix that on my side.

[jeremypoulter]

Sorry, I have not had a chance to look at it yet

[KipK]

@jeremypoulter Just a hint as I'm looking at it, it's not the fact that we have only one timer, it seems to need booth an active and disabled timer. I've tried with 2 enabled or disabled timers, and it also crash the firmware.
But I still don't know where it segfault, I need a stacktrace