函数未正确从标准输入读取

Question

所以我一直在尝试编写一个从标准输入读取数据的虚拟计算机程序。标准输入不断以（%d %s %d）的形式出现，数据基本上代表了一个有特定任务的程序，看起来像这样：

01 READ 60
02 LOAD 60
03 SUB 61
04 STOR 60
05 BRNG 15
06 READ 70
07 LOAD 70
08 ADD 80
09 STOR 80
10 LOAD 60
11 SUB 61
12 STOR 60
13 BRNG 15
14 BRAN 6
15 WRIT 80
16 HALT 99
61 SET 1
80 SET 0

有一个函数作为编译器工作，应该逐行读取这些数据，但出于某种原因，每当此函数运行时，我都会得到 “不正确格式的程序现在将退出”。一段时间以来，我一直在试图弄清楚为什么它会这样做，但没有任何线索。

程序逐行读取stdin流，当没有更多stdin时它应该停止并检查格式是否正确，并且确实是正确的格式，为什么它不能正确读取？我个人的猜测是由于某种原因它没有逐行阅读，但我不确定是不是这样！

PS：我第一次制作这个程序时使用的是文件指针，它工作得很好，但是当我将流切换到标准输入时，它开始给出不正确的格式错误！

int compile ( int memory [], int* instructionCounter , int* instructionRegister ,int*operationCode ,int* operand){
    char s[80]; /* The buffer */
    *instructionRegister=0;
    *operationCode=0;
    while(((*instructionRegister)=scanf("%d %s %d", operationCode,s,operand)) != NULL){ /*Reads data line by line then stores the integer returned by fscanf to instructionRegister pointer so that I can check for formating */
       if((*instructionRegister) ==3 ){ /*Checks for improper format by comparing the current instructionRegister count to 3, returns improper format otherwise */
            if(*operand >9999|| *operand <0){  /* Checks for word overflow in compiler, makes sure that digits do not exceed 9999 */
                printf("attempts to place a word in memory that is larger than 4 digits, or attempted to pass in a negative value\n ");
                exit(0);
            }
            /*Compares the string section of the code and checks if it matches the following words and then it converts it to it's 4 digit value by adding into it the operand */
            if(strcmp(s,"READ") == 0) {
                memory[*operationCode] = 10 * 100 + *operand;
            }
            else if(strcmp(s,"WRIT") == 0) {
                memory [*operationCode] = 11 * 100 + *operand;
            }
            else if(strcmp(s,"LOAD") ==0){
                memory [*operationCode] = 20 * 100 + *operand;
            }
            else if(strcmp(s,"PRNT") ==0){
                memory [*operationCode] = 12 * 100 + *operand;
            }
            else if(strcmp(s,"STOR") ==0){
                memory [*operationCode] = 21 * 100 + *operand;
            }
            else if(strcmp(s,"SET") ==0){
                memory [*operationCode] = *operand;
            }
            else if(strcmp(s,"ADD") ==0){
                memory [*operationCode] = 30 * 100 + *operand;
            }
            else if(strcmp(s,"SUB") ==0){
                memory [*operationCode] = 31 * 100 + *operand;
            }
            else if(strcmp(s,"DIV") ==0){
                memory [*operationCode] = 32 * 100 + *operand;
            }
            else if(strcmp(s,"MULT") ==0){
                memory [*operationCode] = 33 * 100 + *operand;
            }
            else if(strcmp(s,"MOD") ==0){
                memory [*operationCode] = 34 * 100 + *operand;
            }
            else if(strcmp(s,"BRAN") ==0){
                memory [*operationCode] = 40 * 100 + *operand;
            }
            else if(strcmp(s,"BRNG") ==0){
                memory [*operationCode] = 41 * 100 + *operand;
            }
            else if(strcmp(s,"BRZR") ==0){
                memory [*operationCode] = 42 * 100 + *operand;;
            }
            else if(strcmp(s,"HALT")==0){
                memory [*operationCode] =9999;
            }

            else {   /* Prints back to the user that the compiler did not recognize one of them commands as it was going through it */
                printf ("This is an unknown command, commands are case sensitive, program will now exit \n");
                exit(0);

            }
        }
        else{    /* Returns improper format if instructionRegister does not match 3*/
            printf("Improper Format, program will now exit \n");
            exit(0);
        }

    }
    /* Checks if the instruction data contains a HALT, if not it would terminate */
    while(*instructionCounter<100){
        if (memory[*instructionCounter] == 9999){
            return 1;
        }
        else
            (*instructionCounter)++;
    }
    printf("Halt was not found, program will now exit");
    exit (0);
}

此函数的预期目的是逐行读取数据并将其存储在内存虚拟单元中。它应该检查这些行的格式是否为%d %c %d，以便它可以提取数据并将其相应地存储在内存中。

Answer 1

你的问题是这一行：

while(((*instructionRegister)=scanf("%d %s %d", operationCode,s,operand)) != NULL){

scanf() 函数返回

• 成功转换的次数，介于 0 和 3 之间，或
• EOF 以防在任何成功转换之前发生输入失败（文件结束或读取错误）。 EOF 宏是一些负数，通常是 -1。

您将返回值与NULL 进行比较，即0（或(void *)0 或类似）。使用您提供的输入，while 循环将继续读取，直到scanf() 返回EOF。这不会跳出while 循环，但下一次测试会失败：

    if((*instructionRegister) ==3 ){ 
        /* big snip */
    else{
        printf("Improper Format, program will now exit \n");
        exit(0);
    }

当然，解决方案是将结果与EOF 进行比较，而不是NULL。

---

在您的scanf() 调用中也存在潜在的缓冲区溢出。如果字符串长度超过 79 个字符，%s 转换说明符将导致 scanf() 写入超出 s[] 缓冲区的末尾。 （别忘了它会也写一个终止\0。）

您可以通过添加最大字段宽度来限制数量：

scanf ("%d %79s %d", /*...*/)

如果输入字符串长度超过 79 个字符，则输入将无法匹配 "%d" 修饰符之前的空格，scanf() 将返回 2 以指示不完整的结果。