PClint的局限

问题

线上出现一个bug，把一个很大数赋值给一个32bit数导致，下面做了一个测试，对于这种溢出问题

测试代码

#include <stdio.h>
#define KK 1
int main()
{ 
  unsigned int i,k,m; 
  unsigned int j= 1;

 k=0x12345678*100;
 i=j*0x12345678*100;
 m=KK*0x12345678*100;
  return 0;
}

pclint检查结果

--- Module: diy64.c (C)
     1  /* Enter your 64-bit C code here - you can delete this line.      */
     2  
     3  #include <stdio.h>
     4  #define KK 1
     5  int main()
     6  { 
     7    unsigned int i,k,m; 
     8    unsigned int j= 1;
     9  
                         _
    10   k=0x12345678*100;
diy64.c  10  Warning 648:  Overflow in computing constant for operation: 'multiplication'
    11   i=j*0x12345678*100;
                            _
    12   m=KK*0x12345678*100;
diy64.c  12  Warning 648:  Overflow in computing constant for operation: 'multiplication'
                  _
    13    return 0;
diy64.c  13  Warning 438:  Last value assigned to variable 'i' (defined at line 7) not used
diy64.c  13  Warning 438:  Last value assigned to variable 'k' (defined at line 7) not used
diy64.c  13  Warning 438:  Last value assigned to variable 'm' (defined at line 7) not used
        _
    14  }
diy64.c  14  Warning 550:  Symbol 'i' (line 7) not accessed
diy64.c  14  Warning 550:  Symbol 'k' (line 7) not accessed
diy64.c  14  Warning 550:  Symbol 'm' (line 7) not accessed

结论

• 只能针对确定性运算结果是常数提示，其中运算过程不能含有变量