Is there any example program that responds to button click without going to external-wakeup-only mode?

Hi,
So far I found only peripheral examples that callarch_ble_ext_wakeup_on()before configuring the button and going to sleep.
i.e. the program is not intended to work (advertise or respond to timers) while it waits for a button click.

Furthermore, when the wkupct callback is called it does something like:

> static void app_button_press_cb(void)
> {
> if (GetBits16(SYS_STAT_REG, PER_IS_DOWN))
> {
> periph_init();
> }
>
>如果(arch_ble_ext_wakeup_get ())
> {
> arch_set_sleep_mode(app_default_sleep_mode);
> arch_ble_force_wakeup();
> arch_ble_ext_wakeup_off();
> app_easy_wakeup();
> }
> }


Notice how the call toarch_ble_ext_wakeup_off()turns off a boolean flag (namedsleep_ext_force), so future calls toarch_ble_ext_wakeup_get()will returnfalse. Therefore, if there are many interrupts in a short interval (app_button_press_cb()is called many times in a row), the block of code wrapped inside the secondifstatement will be executed only once.

We have a project in which there are many interrupts in a short interval, to which we want to respond only once. However, we want the program to keep "working" while waiting for the "button-clicks" (advertise and run timers) - for example, count the number of button-clicks and update the advertising payload with this number.
We are experiencing a very strange behavior - program resets, program stop working and waste battery, etc. We are not sure where the bug is.
Is there an example program that do something similar?
Should I follow a similar pattern, i.e. do something like:

> bool waiting_for_button __attribute__((section("retention_mem_area0"), zero_init));
> static void app_button_press_cb(void)
> {
> if (GetBits16(SYS_STAT_REG, PER_IS_DOWN))
> {
> periph_init();
> }
>
> if (waiting_for_button)
> {
> arch_set_sleep_mode(app_default_sleep_mode);
> arch_ble_force_wakeup();
> waiting_for_button = false;
> app_easy_wakeup();
> }
> }

i.e. Should I use a different boolean flagwaiting_for_buttonwhich does not affect the sleep likearch_ble_ext_wakeup_on()?
Is there a risk in listening for a button click without going to external-wakeup-on mode ? Like interrupting the program if the button is clicked in exactly the same moment it is executing?

I know that thewkupct_enable_irqaccepts two parameters number-of-events and debouncing-time. We will use higher number-of-events to avoid too many wakeups. I want to be sure that even with a low number-of-events the program does not reset or goes crazy.

Regards,
Oren Zomer