CWE-1077 (使用不正确的比较运算符比较浮点值) — Vulnerability Class 4

CWE-1077 represents a logic error where developers use direct equality operators to compare floating-point numbers, ignoring inherent precision loss from binary representation. This weakness arises because arithmetic operations often introduce minute rounding discrepancies, causing two mathematically equivalent values to differ slightly in memory. Attackers typically exploit this by crafting inputs that trigger these subtle divergences, bypassing security checks or causing unexpected program termination. To mitigate this risk, developers must avoid strict equality tests for floats. Instead, they should implement tolerance-based comparisons using a small epsilon value, checking if the absolute difference between numbers falls within an acceptable threshold. This approach ensures robust handling of imprecise results, maintaining logical integrity and preventing exploitation through numerical edge cases.