Outputting raw bytes from a file, first byte is corrupted

0
votes

So I have written a small program that reads the contents of a file into a char array(because fstream seems to only support char pointers). What I want to do is send the raw bytes to the console. AFAIK char is an 8 bit data type so it should not be too hard. However if I just print members of the array, I get the characters corresponding to the ASCII values, so I am using a static cast. This works fine, except the first byte does not seem to get cast properly. I am using a PNG file as the test.bin file. PNG files always begin with the byte sequence of 137,80,78,71,13,10,26,10. However the fist byte is printed incorrectly. I have a feeling it has to do something with the value being over 127. However, I cannot change the read buffer data type to anything else (like unsigned char, or unsigned short int), because foo.read() from fstream only supports char destination buffers. How do I get fstream to read the raw bytes into a usable unsigned type?

My code:

#include <iostream>
#include <fstream>
#include <sys/stat.h>

#define filename "test.bin"

void pause(){
    std::string dummy;
    std::cout << "Press enter to continue...";
    std::getline(std::cin, dummy);
}


int main(int argc, char** argv) {
    using std::cout;
    using std::endl;
    using std::cin;
    // opening file
    std::ifstream fin(filename, std::ios::in | std::ios::binary);
    if (!fin.is_open()) {
       cout << "error: open file for input failed!" << endl;
       pause();
       abort();
    }
    //getting the size of the file
    struct stat statresults;
    if (stat(filename, &statresults) == 0){
        cout<<"File size:"<<statresults.st_size<<endl;
    }
    else{
        cout<<"Error determining file size."<<endl;
        pause();
        abort();
    }
    //setting up read buffer and reading the entire file into the buffer
    char* rBuffer = new char[statresults.st_size];
    fin.read(rBuffer, statresults.st_size);

    //print the first 8 bytes
    int i=0;
    for(i;i<8;i++) {
        cout<<static_cast<unsigned short>(rBuffer[i])<<";";
    }



    pause();
    fin.clear();
    fin.close();
    delete [] rBuffer;
    pause();
    return 0;
}
3
c++arraysfile-iocastingunsigned
Can you see with the debugger what is the actual value? – Eugene Sh.
The value of rBuffer[0] is -119. – uLoop
How about if you seek to the beginning of the file? fin.seekg (0, fin.beg); – Juan Leni
Can you verify (with hex editor) that your file has the correct value? – Eugene Sh.
It's not necessary to provide in flag to an ifstream nor is it necessary to clear it before closing it. – Neil Kirk

3 Answers

3
votes

-119 signed is 137 unsigned (both are 1000 1001 in binary).
This gets sign-extended into the short 1111 1111 1000 1001, which is 65,417 unsigned.
I assume this is the value you're seeing.

To read into an unsigned buffer:

unsigned char* rBuffer = new unsigned char[statresults.st_size];
fin.read(reinterpret_cast<char*>(rBuffer), statresults.st_size);
0
votes

How about trying something other than fin.read()?

Instead of:

char* rBuffer = new char[statresults.st_size];
fin.read(rBuffer, statresults.st_size);

You could use:

unsigned char* rBuffer = new unsigned char[statresults.st_size];
for(int i = 0; i < statresults.st_size; i++)
{
    fin.get(rBuffer[i]);
}
0
votes

You likely want to be using unsigned char as your "byte". You could try something like this:

using byte = unsigned char;

...

byte* buffer = new byte[statresults.st_size];
fin.read( reinterpret_cast<char*>( buffer ), statresults.st_size );

...

delete[] buffer;