Floating point math is imprecise because of the challenges of storing such values in a binary representation. Even worse, floating point math is
not associative; push a float or a double through a series of simple mathematical operations and the answer will be
different based on the order of those operation because of the rounding that takes place at each step.
Even simple floating point assignments are not simple:
float f = 0.1; // 0.100000001490116119384765625
double d = 0.1; // 0.1000000000000000055511151231257827021181583404541015625
(Results will vary based on compiler and compiler settings.)
Therefore, the use of the equality (==) and inequality (!=) operators on float or double values
is very often an error.
The accepted solution is to use or write a float comparison library:
- Either using a comparison taking into account the magnitude of the numbers being compared and an epsilon value (which may be based on the
capability of the floating point epsilon (FLT_EPSILON)). This comparison will often be absolute for very small values, and relative for larger ones
- Or using the notion of units in the last place
This rule checks for the use of equality/inequality tests on floats, doubles and long doubles.
Noncompliant code example
float myNumber = 3.146;
if ( myNumber == 3.146 ) { //Noncompliant. Because of floating point imprecision, this will be false
// ...
}
if (myNumber <= 3.146 && mNumber >= 3.146) { // Noncompliant indirect equality test
// ...
}
if (myNumber < 4 || myNumber > 4) { // Noncompliant indirect inequality test
// ...
}
