[PROS] call of overloaded 'abs(double&)' is ambiguous

VEX V5 Technical Support UNOFFICIAL Tech Support
1

The linter gives me the above error when I try to take the abs() of a double in this code on lines 4 and 5:


double PIDdo(PID *thing){
  updateSensors();
  thing->error  = thing->target - thing->sensor;
  if(abs(thing->error) < 200){
    if(abs(thing->error) < 8){
      //thing->integral = 0;
    }
    else{
      thing->integral += thing->error;
    }
  }
  else{
    thing->integral = 0;
  }
  //may need integral = 0 stuff
  thing->derivative = thing->error - thing->previous_error;
  thing->previous_error = thing->error;
  return (thing->Kp*thing->error + thing->Ki*thing->integral + thing->Kd*thing->derivative);
}

I tried using


fabs()

, but that seems to be only for C++…
What can I do to fix this?

2

[s]I think you should be using the


.

operator instead of


->

, as


thing

is already a pointer from your function parameter.[/s]

3

AFAIK the


.

operator cannot be applied to pointers directly and instead must be used like this:


(*pointer).foo

which


pointer->foo

is actually the shortcut of.
PROS actually notices this and the linter gives this error when the


.

operator is used in this circumstance:


request for member 'integral' in 'thing', which is of pointer type 'PID*' (maybe you meant to use '->' ?)
4

+1

5

Actually, adding


#include "math.h"

seems to make


fabs()

exist. I don’t have the ability to see if it actually works right now. Is this a safe thing to do?

6

This should be fine assuming it compiles +links.


fabs

is correct. If it doesn’t work, remember that its easy enough to implement an abs() function of your own using conditionals or bit inversion.

one example:


double fabs(double num) { //crappy fabs()
    if(num < 0.0) return num *= -1;
    else  return num;
}

Or you can try something more on the edge similar to* this :wink: : c - How is this implementation of fabs() working? - Stack Overflow
Just don’t expect your function library to be compatible with V5 once that comes out if you do that.

Indeed,


abs()

is your problem. I don’t believe C provides the overload for double in


abs(int)

, only C++ does this. Also, @sazrocks is correct that the


->

is not the issue at all here.

  • not this because cortex is not x86

edit (confirmed): C abs() function | C Arithmetic functions | Fresh2Refresh

7