Question

I’m running on a scaled down version of CentOS 5.5 without many tools available. No xxd, bc, or hd. I can’t install any additional utilities, unfortunately. I do have od, dd, awk, and bourne shell (not bash). What I’m trying to do is relatively simple in a normal environment. Basically, I have a number, say 100,000, and I need to store its binary representation in a file. Typically, I’d do something like …

printf '%x' "100000" | xxd -r -p > file.bin

If you view a hex dump of the file, you’d correctly see the number represented as 186A0.

Is there an equivalent I can cobble together using the limited tools I have available? Pretty much everything I’ve tried stores the ascii values for the digits.

Answer 1

You can do it with a combination of your printf, awk, and your shell.

#!/usr/bin/env awk
# ascii_to_bin.awk
{
    # Pad out the incoming integer to a full byte
    len = length($0);
    if ( (len % 2) != 0) {
        str = sprintf("0%s", $0);
        len = len + 1;
    }
    else {
        str = $0;
    }
    # Create your escaped echo string
    printf("echo -n -e \"");
    for(i=1;i<=len;i=i+2) {
        printf("\\\\x%s", substr(str, i, 2));
    }
    printf("\"");
}

Then you can just do

$ printf '%x' "100000" | awk -f ascii_to_bin.awk | /bin/sh > output.bin

If you know your target binary length you can just do a printf "%0nX" (n is the target size) and remove the (len % 2) logic.