Electron STARTUP() blinking red mode led

The document (example) for STARTUP() suggests that any code will run. STARTUP() worked great for me until I added my own wait that called Time.now().
See snippet. If I comment out my waitAwhile, in setPins() everything is great otherwise blinking red. I call waitAwhile in loop() to show that it works fine. I’m using the DEV IDE and whatever the default system version is 4.8?
UPDATE: I removed the Particle.process() (leaving just {} and no more errors. That may be logical; but, the documentation should be clearer.

#include "Particle.h"
#include <math.h>

void waitAwhile(unsigned long delay);
void setPins()
  pinMode(D4, OUTPUT); // control valve relay HIGH means open
  digitalWrite(D4, HIGH);
  waitAwhile(1); // this the killer; comment out this statement and it works great.

STARTUP(setPins()); //get the pins setup ASAP

void setup()
void  loop()
  int i = 10;
  while (i>0)
    i --;
    System.enterSafeMode(); // exit
 void waitAwhile(unsigned long delay)
   unsigned long t0 =Time.now() + delay;
   while (Time.now() <= t0) {Particle.process();}

Any code that does not require objects for which you instructed the compiler to have your code executed before their initialisation (like Time)

That's an assumption based on two lines of code?
That's like deriving a fourth order function from two arbitrary sample points.

Particle.process() is tightly related to the network/cloud connection and the docs clearly state

And if networking setup code should be in setup() anything depending on this being done already could hardly hope to work before that, could it?

BTW: Time also "needs" cloud connection to give you the correct time

Using millis() or even delay() might be the better way.

1 Like

thanks for explanation. My comment was aimed at obtaining clarity and think you have provided an improvement. If processing() prior to setup(), is fatal it should be blocked even if not logical.