Trying to find byte position where two files differ using file i/o read and write

Question

I'm trying to compare two files to find where they differ using read and write commands and I'm going off a book's example in some sense and think I got the compare function right but I obviously don't because the sprintf functions print out. My best guess is how I'm comparing the two files?:

(updated code)

 Byte pos where two files differ is:
 Byte value of file 1:  %o
 Byte value of file 2: %o

 void compare_two_binary_files(int f1, int f2)
 {
      write(1, sprintf, 10);
      ssize_t byte_read_f1, byte_read_f2, length, numRead, bob;
      char buf1[BUF_SIZE], buf2[BUF_SIZE], a[100], b[100], counter[100];
      int count = 0, b_pos1, b_pos2;
      while ((byte_read_f1 = read(f1, buf1, sizeof buf1) > 0) && (byte_read_f2 = read(f2, buf2, sizeof buf2) >0)) {
            ssize_t len =  byte_read_f1 <byte_read_f2 ? byte_read_f1 : byte_read_f2;
            if (memcmp(buf1, buf2, len) != 0){
                 ssize_t i;
                 sprintf(counter, "Byte pos where two files differ is:%lld\n", (long long) count + i);
                 sprintf(a, "Byte value of file 1: %hho\n", buf1[i]);
                 sprintf(b, "Byte value of file 2: %hho\n", buf2[i]);
                 break;
            }
            count += len;
 }

New code but still error with not going through the if "!=" statement because it simply doesn't go through it when I want it to. Help please:

 void compare_two_binary_files(int f1, int f2)
 {
    ssize_t byte_read_f1, byte_read_f2, length, i;
    char buf1[BUF_SIZE], buf2[BUF_SIZE], a[BUF_SIZE], b[100], counter[100];
    int count = 0;
    while ((byte_read_f1 = read(f1, buf1, BUF_SIZE) > 0) && (byte_read_f2 = read(f2, buf2, BUF_SIZE) >0)){
            if (byte_read_f1 != byte_read_f2){
                    sprintf(counter, "Byte pos where two files differ is:%lld\n", (long long)  count + 1);
                    sprintf(a, "Byte value of file 1:  %o\n", buf1[i]);
                    sprintf(b, "Byte value of file 2: %o\n", buf2[i]);
                    write(1, counter, 100);
                    write(1, a, 100);
                    write(1, b, 100);
            }
            i++;
            count++;
    }
 }

I'm here now. And it's not going into the if (memcmp(buf1, buf2, len) != 0){:

 void compare_two_binary_files(int f1, int f2)
 {
    //write(1, sprintf, 10);
    ssize_t byte_read_f1, byte_read_f2, length, numRead, bob;
    char buf1[BUF_SIZE], buf2[BUF_SIZE], a[100], b[100], counter[100];
    int count = 0, b_pos1, b_pos2, hey;
    while ((byte_read_f1 = read(f1, buf1, sizeof buf1) > 0) && (byte_read_f2 = read(f2, buf2, sizeof buf2) >0)) {
            ssize_t len =  byte_read_f1 <byte_read_f2 ? byte_read_f1 : byte_read_f2;
            if (memcmp(buf1, buf2, len) != 0){
                    ssize_t i;
                    for (i = 0; i<len; i++){
                            if (buf1[i] != buf2[i]) break;
                    }
                    sprintf(counter, "Byte pos where two files differ is:%lld\n", (long long) count + i);
                    write(1, counter, 100);
                    sprintf(a, "Byte value of file 1: %hho\n", buf1[i]);
                    sprintf(b, "Byte value of file 2: %d\n", buf2[i]);
                    write(1, counter, 100);
                    write(1, b, 100);
                    write(1, a, 100);
                    break;
            }
            count += len;
    }
 }

When I change the != of memcmp to == it works but gives me wrong numbers. I'm supposed to get byte location 4, file 1 to be 65 and file 2 to be 143, but get pos 1, file 1 is 163 and file 2 is 115

Read up on `read` first. 1. you are reading *n* bytes from both files into the same buffer -- when tested, they'd always be "the same". 2. you are comparing the number of read bytes. 3. you cannot compare char arrays using `==`. — Jongware, Jan 29 '14 at 22:49
@Jongware comparing the number of bytes read is good, its just that a subsequent `memcmp()` of properly read buffers is also needed. — chux - Reinstate Monica, Jan 29 '14 at 22:51
@chux: sure, but it seems the OP expects this somehow compares the bytes as well. OP: what chux means is that if `byte_read_f1 != byte_read_f2` then you don't have to compare -- the number of bytes read is different, so the files differ at that point. — Jongware, Jan 29 '14 at 22:54
@chux - So the buf, BUF part of the loop is fine, it's just the memcmp() you speak of? — Question_Guy, Jan 29 '14 at 22:55
@MMavipc - Did you really ding me a couple points because I didn't indent properly? Was my code that hard to understand without the proper indent of the last "}"? — Question_Guy, Jan 29 '14 at 23:11
@OSU_2016 I didn't ding you any points, I just left a comment. Someone else must have done it. The problem is your tab width, in most places you have it set to 8 spaces, but in 6 lines you have it set to 4 spaces — Avery3R, Jan 29 '14 at 23:15

chux - Reinstate Monica · Accepted Answer · 2014-01-30T22:39:55.083

1

Need separate buffers and use memcmp and/or a loop to compare the bytes.
All be aware that file lengths may differ.

// Untested code
char buf1[BUF_SIZE];
char buf2[BUF_SIZE];
while (((byte_read_f1 = read(f1, buf1, sizeof buf1)) > 0) && 
       ((byte_read_f2 = read(f2, buf2, sizeof buf2)) > 0)) {
  // find min length
  ssize_t len =  byte_read_f1 <byte_read_f2 ? byte_read_f1 : byte_read_f2;
  if (memcmp(buf1, buf2, len) != 0) {  // use memcmp for speed
    ssize_t i;
    for (i = 0; i<len; i++) {
      if (buf1[i] != buf2[i]) break;
    } 
    sprintf(counter, "Byte pos where two files differ is:%lld\n", (long long) count + i);
    sprintf(a, "Byte value of file 1: %hho\n", buf1[i]);
    sprintf(b, "Byte value of file 2: %hho\n", buf2[i]);
    break;
  }
  count += len;
  // If lengths differ
  if (byte_read_f1 != byte_read_f2) {
    sprintf(counter, "Byte pos where two files differ in length: %lld\n", (long long) count);
    // TBD code for a and b
    break;
    }
}

edited Jan 30 '14 at 22:39

answered Jan 29 '14 at 23:07

chux - Reinstate Monica

143,097
13
135
256

@ Chux or someone else - Can you tell me why write works like it does in my first post rather than putting the write statement within the if statement? – Question_Guy Jan 29 '14 at 23:26
I'm still getting output of %lld, %hho, etc – Question_Guy Jan 30 '14 at 00:01
@OSU_2016 Not clear on your request. `read()` is inside a `while` loop in both the original post and this answer. `read()` is not in an `if()` statement. – chux - Reinstate Monica Jan 30 '14 at 01:50
@OSU_2016 Please show your code that "still getting output of %lld, %hho". – chux - Reinstate Monica Jan 30 '14 at 01:51
Updated in original post – Question_Guy Jan 30 '14 at 02:10
@OSU_2016 1) Suggest that you do not change your original question, but rather append or note changes. 2) Perhaps your system is not compliant and does not understand format specifiers `"%o"` nor `"hho"`. Try a simple `"%d"` – chux - Reinstate Monica Jan 30 '14 at 03:45
Well I'm not too familiar with using this stackoverflow....so I'm not comfortable appending, but as you are implying, I could have added my new code – Question_Guy Jan 30 '14 at 04:07
And I tried output of %d and still nothin, also "write(1, sprintf, 100)" was supposed to say "write(1, "", 100)" – Question_Guy Jan 30 '14 at 04:18
My only guess is with the sprintf. I say that because when I mess with it I get garbage as output and I don't know how to use it exactly with the write command. I know write excepts a char buff and sprintf is that but not sure how to implement it. – Question_Guy Jan 30 '14 at 04:23
Try `sprinf(counter, "test"); write(1, counter, strlen(counter));` Some `sprintf()` functions incorrectly return the count converted rather than a `char *`. – chux - Reinstate Monica Jan 30 '14 at 04:26
@OSU_2016 Obviously we have a `write()` issue. Sorry I am not much more help. – chux - Reinstate Monica Jan 30 '14 at 04:38
I just gained a little progress of understanding I think.....So when I say write(....), I need to say write(1, the char name I gave for sprintf)? - NVM, I've been doing that I think....I'm just worn out right now from tryin to figure this out. – Question_Guy Jan 30 '14 at 04:40
I did your test in a seperate c program, by itself, and it worked – Question_Guy Jan 30 '14 at 04:40
@OSU_2016 GTG Good luck – chux - Reinstate Monica Jan 30 '14 at 04:57
write the values of byte_read_f1 and byte_read_f2. If 0, no data in file. if < 0, check `errno`. – chux - Reinstate Monica Jan 30 '14 at 04:59
1

@OSU_2016 Based on your more recent post http://stackoverflow.com/questions/21468435/using-file-i-o-to-read-byte-length I've corrected this one. – chux - Reinstate Monica Jan 30 '14 at 22:40
@OSU_2016 Only lines `while (((byte_read_f1 = read(f1, buf1, sizeof buf1)) > 0) && ((byte_read_f2 = read(f2, buf2, sizeof buf2)) > 0)) {` – chux - Reinstate Monica Jan 30 '14 at 23:57

Trying to find byte position where two files differ using file i/o read and write

1 Answers1