Question

I wrote this binary reader after a tutorial on the internet. (I’m trying to find the link…)

The code reads the file byte by byte and the first 4 bytes are together the magic word. (Let’s say MAGI!) My code looks like this:

std::ifstream in(fileName, std::ios::in | std::ios::binary);
char *magic = new char[4];

while( !in.eof() ){
   // read the first 4 bytes
   for (int i=0; i<4; i++){
      in.get(magic[i]);
   }

   // compare it with the magic word "MAGI"
   if (strcmp(magic, "MAGI") != 0){
        std::cerr << "Something is wrong with the magic word: " 
                  << magic << ", couldn't read the file further! " 
                  << std::endl; 
        exit(1);
    }

   // read the rest ...
}

Now here comes the problem, when I open my file, I get this error output:
Something is wrong with the magic word: MAGI?, couldn't read the file further! So there is always one (mostly random) character after the word MAGI, like in this example the character ?!
I do think that it has something to do with how a string in C++ is stored and compared with each other. Am I right and how can I avoid this?

PS: this implementation is included in another program and works totally fine … weird.

Answer 1

strcmp assumes that both strings are nul-terminated (end with a nul-character). When you want to compare strings which are not terminated, like in this case, you need to use strncmp and tell it how many characters to compare (4 in this case).

if (strncmp(magic, "MAGI", 4) != 0){

When you try to use strcmp to compare not null-terminated char arrays, it can’t tell how long the arrays are (you can’t tell the length of an array in C/C++ just by looking at the array itself – you need to know the length it was allocated with. The standard library is not exempt from this limitation). So it reads any data which happens to be stored in memory after the char array until it hits a 0-byte.

By the way: Note the comment to your question by Lightness Races in Orbit, which is unrelated to the issue you are having now, but which hints a different bug which might cause you some problems later on.