Calculation Error? 60*60*1000 is NOT 3600000 ???!

Projects Programming
1

I did already some Ardunio projects, though I am not a hardcore programmer.
Yesterday I discovered an interesting effect which I thought of posting here as this might
occur to more people, I guess.

My code was like this:

unsigned long T0; // Time stamp
unsigned long Rest_L; // remaining time in min
unsigned long Dauer; // Duration
i

void loop(void) {
...

T0 = millis(); // is set when special condition occurs and used as reference for calculation later

.....

Rest_L= ((T0+(60601000) - millis() )/60000); //calculate remaining time in minutes

This didn't work out.
Then I tried to get closer to the problem and isolated this:

Dauer = 60603600;

printed then Dauer to the serial monitor.
And it was't 3600000 to my surprise.

Then I substituted the "60601000" by 3600000 in my comparison equation and it worked.

Rest_L= ((T0+3600000 - millis() )/60000); //calculate remaining time in minutes

I find this behaviour a bit surprising, but probably harcore programmers will laugh.
Nevertheless several newbies might fall into the same trap.
So I thought it might be useful to post it.

Cheers, Peter

2

Not this one again

60UL * 60UL etc

3

Yes, this is what I thought as well would be necessary,
So, values not specifically declared are "int"? And this creates the overflow?
Right?

As said, for a newbie this is not easy to get.

Cheers, Peter

4

all the math that can be done at compile time are done with integer on 16 bits unless told otherwise by using l or ul in your numbers

t = 60*60*1000ul; will do the right thing (you need at least 1 long integer in the formula to get the math done with more bits)

5

understood.
I kind of foresaw this as I declared my result value already as unsigned long.
Thought this would do it.
But now I know.

Cheers, Peter

6

Yep - sometimes our wishes or expectations are let down by the cruel reality of the pre-processor or compiler

that being said, the compiler warns me that I'm doing something stupid if I write

unsigned long t = 60*60*1000;

at compile time you would get

warning: integer overflow in expression [-Woverflow]
 unsigned long t = 60 * 60 * 1000;
                           ^

so always look at warning and fix them is a good practice.