c - 当 y 为 0 时，模运算符 x % y 的语义是什么？

Question

假设我尝试执行以下操作：

y = 0;
z = x % y;

这个语义是定义明确的、平台相关的还是未定义的？我主要询问 C/C++，但对各种编程/脚本语言（Java、perl、sh 等）的答案很感兴趣

我之所以这么问，部分是因为定义模运算有不同 的可能方法：作为除法运算的余数；作为商群的大小等。

Answer 1

C 的行为是未定义的。

来自C11 6.5.5 乘法运算符，p5

/ 运算符的结果是第一个操作数除以第二个操作数的商；% 运算符的结果是余数。在这两种操作中，如果第二个操作数的值为零，则行为未定义。

Answer 2

It's well defined for JavaScript:

The result of an ECMAScript floating-point remainder operation is determined by the rules of IEEE arithmetic: [...]

If the dividend is an infinity, or the divisor is a zero, or both, the result is NaN.

---

Now about the other languages. The common approach (Java, C#, Python, Ruby) is to throw some kind of ZeroDivisionError at you when you attempt to evaluate somenum % 0 expression.

For Perl, it's a bit more interesting:

use Data::Dumper;
print Dumper 0 % 0;
print 'Something else';

Now, this code results in Illegal modulus zero error; but had you put 0 / 0 instead, you would have seen Illegal division by zero message. Both are errors (stop execution of the remaining code), of course, not warnings.

---

Now PHP chooses a bit different stance on this:

var_dump(0 % 0); // it's the same for any numeric dividend
// Warning: Division by zero in ...
// bool(false)

As you see, you get false (sic) as a result, but warning is triggered. It's ignorable, though; have you set error_reporting level to E_ERROR, you wouldn't have even seen it.

Answer 3

在 Java 中，如果您尝试编译

public static void main(String[] args) {
    int x = 10,y,z;
    y = 0;
    z = x % y;
    System.out.println("Z: " + z);
} 

您将收到此消息：

Exception in thread "main" java.lang.ArithmeticException: / by zero
    at locationConfiguration.LocationConfigurator.main

因此，您将无法进行模零。