Binary I/O file reading (0x1a trouble)

[Danish]

My problem: I need to know how to read a block of n bytes from a file and copy them to a buffer in memory. I am trying to use open() and read(), and it works just fine as long as there is no 0x1a in the file. But I have them in there. Am I using the wrong function for this? I mean, getc() works, but I don't really want to use a for loop just to get the contents of a file (efficiency goes down the crapper). I'm thinking there's a flag that I could use for open that would override this, but lseek() sees past the 0x1a's for the eof. I'm really confused right now, so any help is appreciated. Thanks.

I'll post the answer, of course, if I find it.

- Dane

Edit: Okay, the problem I was having is that I was assuming getc() was the only way to get binary data. fread() does that too. I wonder if there's a way to use those lower-level functions, though? Well, either way, I'm happy.

[joe_bruin]

here is a program that uses read(2) to read binary data. it has no problems with 0x1a or any other character, and is much faster than multiple getc's.

#include <errno.h>
#include <fcntl.h>
#include <stdio.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>

int main()
{
  int len, fd, i;
  unsigned char buf[1024];

  fd = open("in.txt", O_RDONLY);
  if(fd == -1)
  {
    perror("open");
    return 1;
  }

  len = read(fd, buf, sizeof(buf));
  if(len == -1)
  {
    perror("read");
    // should probably check errno here
    return 1;
  }
  if(len == 0)
  {
    printf("eof reached\n");
  }
  else
  {
    for(i = 0; i < len; i++)
    {
      printf("%02x ", buf[i]);
      if(i % 8 == 7) printf("\n");
    }
  }
  close(fd);

  return 0;
}

[Danish]

Oh, thank you very much. I'll try this right away and see what happens. I guess I was just having problems understanding how to use read()...

Edit the first: It's still stopping after 0x1a; it detects that as EOF... I'm using MinGW, does that make a difference?

Edit the second: I tried your code, and made a file in.txt; I put in some text, and it read it fine. I opened up a hex editor, changed the fourth character to 0x1a, and it won't see past three characters. I'm still thinking this is a MinGW issue...?

Edit the third: AHA! Eureka. Change

fd = open("in.txt", O_RDONLY);

to

fd = open("in.txt", O_RDONLY | O_BINARY);

Just as I thought -- a flag in open()... well, thanks for the help.

And yeah, read() is faster than fread() afaict.