Operator/Bit Issue

Hi,

We are having a really unique problem that has stumped the whole programing team, all of the mentors, and the team leader. We are trying to write some code in MPLab to calculate the number of encoder ticks needed to reach six feet. We have found that fourteen encoder clicks is the same as one inch. The problem comes in when we try to multiply 72(six feet in inches) by 14(clicks in an inch). When this happens we get -16 which is 32 bits off. We are using the variable type long. Below is the code that we have written:

long atemp=0;
atemp=14*72;//should be 1008
printf("atemp=%ld
", atemp);

This code did not work and gave us a value of -16 for atemp. We also tried this:

long atemp=0;
atemp=72+72+72+72+72+72+72+72+72+72+72+72+72+72;
printf("atemp=%ld
", atemp);

This also did not work and gave us a value of -16. Adding smaller numbers works fine and multiplying smaller numbers also works. We have pretty much exhausted any ideas.
Thanks,
Julian

C18 diverges from the ISO standard in that when all operands will fit in a char, it will do the math in a char. If you cast one operand to an int, it will do the math the way you want. You can also enable integer promotion in the project file.

Or write

atemp=14L*72L;

which will make sure the numbers are treated as longs.

Also, i’m not sure if you can print longs…

I believe that both casting them to an int or using %ld should work.

I believe that both casting them to an int or using %ld should work.

int wont print the whole long though (only the low 8 bits), and i can’t remember whether %ld actually works.

I think %ld does work in later revisions of the default code (kevins?)

Also, an alternative method of getting the math the way you want it:

 atemp = (long)14*(long)72; 

Yay for C-style casting… in my opinion, SO much easier than C++ casts.

You code is working. The problem is with the printf statement.
so you can keep your code.

Thanks to everyone for the help. Typecasting made the code work. The printf statement was fine.
Julian