Strange effect with Nulltermination char inside of string-CodePudding

Simple Code example: Godbolt-Compiler Explorer

#include <cstdint>
#include <iostream>

using namespace std;

using u16 = uint16_t;

int main(){

  //char str[] = "0123456789\0abcdefghijklmnopqrtsuvwxyz";
  char str[] = "0123456789\00123456789\00123456789";

  auto str_len = (sizeof(str)-1);
  cout << "stringlength: " << str_len << "\n";

  for(int ii=0; ii<str_len;   ii){
    cout << u16(str[ii]) << ",";
  }
  cout << "\n";

  return 0;
}

Output is:

stringlength: 28 48,49,50,51,52,53,54,55,56,57,1,50,51,52,53,54,55,56,57,1,50,51,52,53,54,55,56,57,

which is the same array content as when viewed with a debugger...
Here Termchar, 0 and 1 replaced with a single char with value '1'

but when I change for the other definition of char str[] (the commented out one...)

I get 48,49,50,51,52,53,54,55,56,57,0,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,116,115,117,118,119,120,121,122,

which is correct.

So did I hit a strange unicode thing, or how to explain that effect?

PS: I use that char xyz[] definition, so (sizeof()-1) can be used as compile time strlen() calculation...

CodePudding user response：

The easiest way to stop this from being interpreted as an octal escape sequence would be:

char str[] = "0123456789\0" "0123456789\0" "0123456789";

Or just use \000 as the null character:

char str[] = "0123456789\0000123456789\0000123456789";

CodePudding user response：

One solution - but with an ugly syntax is:

#include <cstdint>
#include <iostream>

using namespace std;

using u16 = uint16_t;

int main(){

  char str[] = "0123456789\0" "0123456789\0" "0123456789";

  auto str_len = (sizeof(str)-1);
  cout << "stringlength: " << str_len << "\n";

  for(int ii=0; ii<str_len;   ii){
    cout << u16(str[ii]) << ",";
  }
  cout << "\n";

  return 0;
}