Range of mills() on ATTiny1616?

[JeremyLaurenson]

I am using mills() to do some timing on this chip using megaTinyCore and I have a weird bug where if I try add more than 30 seconds to my time

setup()
{
 millisNow = millis();
  millisFullEnd = millisNow + 1000 * secondsToRun;
}

void loop() {
    millisNow=millis();
    if (millisNow > millisFullEnd) {
      cycleOn = false;
    }  // Did we finish the entire thing?
    
}

If I do a test:

#include <Wire.h>
#define s1pulse 0
#define s1direction 1
#define GREENLED 6
#define REDLED 5


unsigned long  long millisNow = 0;
unsigned long long  millisEnd = 0;
unsigned long long millisFullEnd = 0;

void setup() {

  pinMode(GREENLED, OUTPUT);
  pinMode(REDLED, OUTPUT);

  digitalWrite(GREENLED, HIGH);
  digitalWrite(REDLED, LOW);

  millisNow = millis();
  millisFullEnd= millisNow + 60 * 1000;
  

}



void loop() {
  millisNow = millis();
  if (millisNow > millisFullEnd) {
  digitalWrite(GREENLED, LOW);
  digitalWrite(REDLED, HIGH);
  
  }
}

I get an error on compile:

/private/var/folders/tc/fxg4yd1n2vg56pks9vmf_s680000gn/T/.arduinoIDE-unsaved202426-39231-dxgyes.1sbl8/sketch_mar6a/sketch_mar6a.ino: In function 'void setup()':
/private/var/folders/tc/fxg4yd1n2vg56pks9vmf_s680000gn/T/.arduinoIDE-unsaved202426-39231-dxgyes.1sbl8/sketch_mar6a/sketch_mar6a.ino:22:33: warning: integer overflow in expression [-Woverflow]
   millisFullEnd= millisNow + 60 * 1000;
                              ~~~^~~~~~
Sketch uses 1662 bytes (10%) of program storage space. Maximum is 16384 bytes.
Global variables use 133 bytes (6%) of dynamic memory, leaving 1915 bytes for local variables. Maximum is 2048 bytes.

Which smells to me like I am stepping on an integer overflow in the first example...

[hmeijdam]

If you don't want the warning

change

millisFullEnd= millisNow + 60 * 1000;

to

millisFullEnd= millisNow + 60 * 1000L;

[grandaspanna]

Strictly speaking, it's a warning, not an error, but the root cause is that it's treating the "60" and "1000" as ints (16-bit signed) and deciding that there's a risk of overflow. I'm not guru enough to know if the code it generates is correct.

There's a voice inside me that's paranoid about order of operations (and readability to some extent), so I'd pretty much always do "millisNow + ( 60 * 1000L)" :-)

[JeremyLaurenson]

Using 1000UL in the code, fixed the problem, so it is an Arduino C issue that I should've been aware of