Using integer arithmetic to calculate a value for assignment to a floating-point variable may lead to loss of information. This can be avoided by converting one of the integers in the expression to a floating type.
When converting integers to floating-point values, and vice versa, it is important to carry out proper range checks to avoid undefined behavior. (See rule FLP34-C. Ensure that floating point conversions are within range of the new type.)
Noncompliant Code Example
In this noncompliant code example, the division and multiplication operations take place on integers and are then converted to floating point. This causes floating-point variables d
, e
, and f
to not be initialized correctly because the operations take place before the values are converted to floating-point values. The results are truncated to the nearest integer or may overflow.
short a = 533; int b = 6789; long c = 466438237; float d = a / 7; /* d is 76.0 */ double e = b / 30; /* e is 226.0 */ double f = c * 789; /* f may be negative due to overflow */
Compliant Solution (Floating Point Literal)
In this compliant solution, the decimal error in initialization is eliminated by ensuring that at least one of the operands to the division operation is floating point.
short a = 533; int b = 6789; long c = 466438237; float d = a / 7.0f; /* d is 76.14286 */ double e = b / 30.; /* e is 226.3 */ double f = (double)c * 789; /* f is 368019768993.0 */
Compliant Solution (Conversion)
In this compliant solution, the decimal error in initialization is eliminated by first storing the integer in the floating-point variable and then performing the arithmetic operation. This ensures that at least one of the operands is a floating-point number and that the subsequent arithmetic operation is performed on floating point operands.
short a = 533; int b = 6789; long c = 466438237; float d = a; double e = b; double f = c; d /= 7; /* d is 76.14286 */ e /= 30; /* e is 226.3 */ f *= 789; /* f is 368019768993.0 */
Exceptions
FLP33-EX1: It may be desirable to have the operation take place as integers before the conversion (obviating the need for a call to trunc()
, for example). If this is the programmer's intention, it should be clearly documented to help future maintainers understand that this behavior is intentional.
Risk Assessment
Improper conversions between integers and floating point values may yield unexpected results, especially loss of precision. Additionally, these unexpected results may actually involve overflow, or undefined behavior.
Rule |
Severity |
Likelihood |
Remediation Cost |
Priority |
Level |
---|---|---|---|---|---|
FLP33-C |
low |
probable |
low |
P6 |
L2 |
Automated Detection
Tool |
Version |
Checker |
Description |
---|---|---|---|
9.7.1 |
|
|
|
Splint |
3.1.1 |
|
|
Compass/ROSE |
|
|
can detect violations of this rule. Any assignment operation where the type of the assigned-to value is float or double, but all the expressions to the right of the assignment are integral is a violation of this rule |
Related Vulnerabilities
Search for vulnerabilities resulting from the violation of this rule on the CERT website.
Related Guidelines
CERT C++ Secure Coding Standard: FLP33-CPP. Convert integers to floating point for floating point operations
The CERT Oracle Secure Coding Standard for Java: NUM16-J. Convert integers to floating point for floating-point operations
ISO/IEC 9899:1999 Section 5.2.4.2.2, "Characteristics of floating types <float.h
>"
MITRE CWE: CWE-681, "Incorrect Conversion between Numeric Types
MITRE CWE: CWE-682, "Incorrect Calculation"
Bibliography
[Hatton 1995] Section 2.7.3, "Floating-point misbehavior"
05. Floating Point (FLP) FLP34-C. Ensure that floating point conversions are within range of the new type