有人可以解释C使用#defineUMAX(a,b)((a)>(b)?(a):(b))指令的硬逻辑错误，该指令返回较低的值，在2个编译器中

我试图找出发生在 MUD c 代码库上的错误的原因。其中 #define UMAX(a, b) ((a) > (b) ? (a) : (b)) 的使用用于返回两个值的最大值。它有时会返回较低的值，甚至调试我也找不到原因。

我隔离了复制的相关代码并在两个编译器中运行它，https ://www.tutorialspoint.com/compile_c_online.php (GCC) 上的在线 C 编译器和使用 Visual Studio 2019。两者都有以下代码吐出负值：

dur = UMAX(0, 4 - number_fuzzy(level / 10));
printf("Dur = %in", dur);

基本上，UMAX 用于防止计算中出现非负值。我们将级别除以 10，然后使用模糊/随机函数将其随机更改为 -1/+0/+1。我们从 4 中减去这个值，如果该值小于零，我们使用 UMAX 返回 0。但是对于 50 到 59 级，这有时会产生负值，如下所示。但我从未见过它对 60-69 产生负面影响。我知道模糊随机是随机改变值，但 UMAX 守卫应该始终有效，而不仅仅是某些级别。

完整代码如下：

//==================================================================
//= SLAP ME ON https://www.tutorialspoint.com/compile_c_online.php =
//==================================================================

#include <stdio.h>
#include<time.h> // pd 5

#define UMAX(a, b)      ((a) > (b) ? (a) : (b))
#define OLD_RAND true
static  int     rgiState[2+55];

// Replacement
//https://p2p.wrox.com/c-programming/37911-my-coin-flip-game.html
void init_mm()
{
    int *piState;
    int iState;
    piState = &rgiState[2];
    piState[-2] = 55 - 55;
    piState[-1] = 55 - 24;
    piState[0] = ((int) time(NULL))&((1 << 30)- 1);
    piState[1] = 1;
    for(iState = 2; iState < 55; iState++)
    {
        piState[iState] = (piState[iState - 1] + piState[iState - 2])
        & ((1 << 30)- 1);
    }
    return;
}




long number_mm( void )
{
#if defined (OLD_RAND)
    int *piState;
    int iState1, iState2, iRand;
    piState             = &rgiState[2];
    iState1             = piState[-2];
    iState2             = piState[-1];
    iRand               = (piState[iState1] + piState[iState2]) & ((1 << 30) - 1);
    piState[iState1]    = iRand;
    if ( ++iState1 == 55 )
        iState1 = 0;
    if ( ++iState2 == 55 )
        iState2 = 0;
    piState[-2]         = iState1;
    piState[-1]         = iState2;
    return iRand >> 6;
#else
    //return random() >> 6;
#endif
}

// db.c line 6267
int number_bits( int width )
{
    return number_mm( ) & ( ( 1 << width ) - 1 );
}

// db.c line 6195
/* Stick a little fuzz on a number. */
int number_fuzzy( int number )
{
    switch ( number_bits( 2 ) )
    {
    case 0:  number -= 1; break;
    case 3:  number += 1; break;
    }
    return UMAX( 1, number );
}

int lixo(int a){
    return UMAX( 1, a );
    
}

int main()
{
    init_mm(); // Needed cause I guess this is for seeding for random number generation


int level = 50;
int dur = 0;
dur = UMAX(0, 4 - number_fuzzy(level / 10));
printf("Dur = %in", dur);
dur = UMAX(0, 4 - number_fuzzy(level / 10));
printf("Dur = %in", dur);
dur = UMAX(0, 4 - number_fuzzy(level / 10));
printf("Dur = %in", dur);
dur = UMAX(0, 4 - number_fuzzy(level / 10));
printf("Dur = %in", dur);
dur = UMAX(0, 4 - number_fuzzy(level / 10));
printf("Dur = %in", dur);
dur = UMAX(0, 4 - number_fuzzy(level / 10));
printf("Dur = %in", dur);
dur = UMAX(0, 4 - number_fuzzy(level / 10));
printf("Dur = %in", dur);

int test = 0;
test = UMAX(0, 4 - lixo(level / 10));
printf("Test = %in", test);
}

可以返回/打印以下内容：

Dur = 0
Dur = 0
Dur = 0
Dur = -1
Dur = 0
Dur = -1
Dur = 0
Test = 0

使用 Visual Studio 2019，我继续组装，如下所示：

    dur = UMAX(0, 4 - number_fuzzy(level / 10));
00971A89  mov         eax,dword ptr [level]  
00971A8C  cdq  
00971A8D  mov         ecx,0Ah  
00971A92  idiv        eax,ecx  
00971A94  push        eax  
00971A95  call        _number_fuzzy (0971181h)  
00971A9A  add         esp,4  
00971A9D  mov         edx,4  
00971AA2  sub         edx,eax  
00971AA4  jns         main+72h (0971AB2h)  
00971AA6  mov         dword ptr [ebp-100h],0  
00971AB0  jmp         main+93h (0971AD3h)  
00971AB2  mov         eax,dword ptr [level]  
00971AB5  cdq  
00971AB6  mov         ecx,0Ah  
00971ABB  idiv        eax,ecx  
00971ABD  push        eax  
00971ABE  call        _number_fuzzy (0971181h)  
00971AC3  add         esp,4  
00971AC6  mov         edx,4  
00971ACB  sub         edx,eax  
00971ACD  mov         dword ptr [ebp-100h],edx  
00971AD3  mov         eax,dword ptr [ebp-100h]  
00971AD9  mov         dword ptr [dur],eax  
    printf("Dur = %in", dur);
00971ADC  mov         eax,dword ptr [dur]  
00971ADF  push        eax  
00971AE0  push        offset string "Dur = %in" (0977B30h)  
00971AE5  call        _printf (09710D2h)  
00971AEA  add         esp,8  
    dur = UMAX(0, 4 - number_fuzzy(level / 10));

但是找不到原因。这很可能是由于我的组装技能低下。

一位朋友建议这可能与无符号整数有关。但是我对 Lixo 函数进行了编码，它不会产生负面影响。我知道这个错误很容易通过用 IF 替换 UMAX 来纠正，但由于技术上的好奇，我对这个错误非常感兴趣。此外，这个古老的 MUD 代码充满了 UMAX 调用，因此了解它发生的原因有些重要。

那么，比我更擅长 C 的人能解释为什么会发生这种行为吗？

回答

#define UMAX(a, b) ((a) > (b) ? (a) : (b))
dur = UMAX(0, 4 - number_fuzzy(level / 10));

dur = ((0) > (4 - number_fuzzy(level / 10))) ? (0) : (4 - number_fuzzy(level / 10)));

#define UMAX_ASSIGN(var, val1, val2) { 
    int var1 = val1; 
    int var2 = val2; 
    var = var1 > var2 ? var1 : var2; 
}

UMAX_ASSIGN(dur, 0, 4 - number_fuzzy(level / 10));

int UMax(int a, int b) { // use inline suggestion if desired.
    if (a > b) return a;
    return b;
}