This rule is part of MISRA C++:2023.
Usage of this content is governed by Sonar’s terms and conditions. Redistribution is
prohibited.
Dir 0.3.1 - Floating-point arithmetic should be used appropriately
[ISO/IEC/IEEE 60559]
Category: Advisory
Amplification
A tool should highlight all suspicious uses of floating-point arithmetic.
Rationale
The safe use of floating-point arithmetic requires a high level of numerical analysis skills and in-depth knowledge of the compiler and target
hardware.
The incorrect use of floating-point may lead to problems related to the presence of positive and negative zero, loss of precision, NaNs (quiet and
signalling), infinities, overflow, underflow, catastrophic cancellation, etc.
Example
bool myIsNaN( double d )
{
return d == std::numeric_limits< double >::quiet_NaN(); // Always returns false
}
void f()
{
float f1 = 1E38f;
float f2 = 10.0f;
float f3 = 0.1f;
float f4 = ( f1 * f2 ) * f3;
float f5 = f1 * ( f2 * f3 ); // Values in f4 and f5 are significantly different
}
Copyright The MISRA Consortium Limited © 2023
