shell - The code is correct, but it does not work

Question

Everything is normal. But it gives an error. I could not solve it.

try.sh:

#!/bin/sh

website="http://lastofdead.xyz"
ipaddress=$( ifconfig | grep -v 'eth0:' | grep -A 1 'eth0' | \
             tail -1 | cut -d ':' -f 2 | cut -d ' ' -f 1)
mlicense=$(curl -s $website/lodscript/special/lisans.php?lisans)

if [ "$mlicense" = "$ipaddress" ];
then 
    echo "Positive"
else
    echo "Negative"
fi

lisans.php:

<?php 
if(isset($_GET['lisans'])) {
echo "188.166.92.168" . $_GET['lisans'];
}
?>

Result:

root@ubuntu:~# bash -x s.sh
+ website=http://lastofdead.xyz
++ cut -d ' ' -f 1
++ cut -d : -f 2
++ tail -1
++ grep -A 1 eth0
++ grep -v eth0:
++ ifconfig
+ ipaddress=188.166.92.168
++ curl -s 'http://lastofdead.xyz/lodscript/special/lisans.php?lisans'
+ mlicense=188.166.92.168
+ '[' 188.166.92.168 = 188.166.92.168 ']'
+ echo Negative
Negative

https://i.stack.imgur.com/jNMNq.jpg

Maybe try to change single = with -eq operator in your condition and remove " characters? — Canser Yanbakan, Feb 26 '17 at 13:01
Personally I would confirm if there is any whitespace around ip addresses — grail, Feb 26 '17 at 13:19
Maybe not related but `if [ "$mlicense" = "$ipaddress" ];` will always give true and the semi-colon is not necessary. — Amin NAIRI, Feb 26 '17 at 13:21
The only thing I can think of is that the result of the `curl` call ends with a CRLF pair, resulting in the value of `mlicense` ending with a carriage return. I would expect that to show up clearly in the output of `bash -x`, though. — chepner, Feb 26 '17 at 13:23
Are you *sure* the code you show above is the code that is actually being executed? — chepner, Feb 26 '17 at 13:34

score 4 · Accepted Answer · edited May 23 '17 at 12:00

Oh, thank you very much for posting the code. Well, having tried to see myself the differences between the two strings, it turned out that the problem was something called Byte Order Mark (BOM). For more info about this, see this answer: https://stackoverflow.com/a/3256014.

The string returned by curl, when piping it to an hex dump, shows this:

$ curl -s 'http://lastofdead.xyz/lodscript/special/lisans.php?lisans'
188.166.92.168
$ curl -s 'http://lastofdead.xyz/lodscript/special/lisans.php?lisans' | xxd -c 17 -g 1 -u
0000000: EF BB BF 31 38 38 2E 31 36 36 2E 39 32 2E 31 36 38  ...188.166.92.168

Can you see them? Those three bytes, 0xEF,0xBB,0xBF, are the UTF-8 representation of the BOM, and that's what makes the strings differ. The question page linked above shows some ways to strip it, like using grep, or you could pipe the curl output to cut -c 2-, or even through a simple substitution after the curl line: mlicense="${mlicense:1}". And, to make sure that what we strip is the BOM, we could use two lines of substitution: bom="$(echo -en '\xEF\xBB\xBF')"; mlicense="$(echo -n "${mlicense#${bom}}")", or even turn them into one: mlicense="$(echo -n "${mlicense#$(echo -en '\xEF\xBB\xBF')}")".

shell - The code is correct, but it does not work

1 Answers1