How to fix IP filter regular expressions written using shell script command in Linux?

Question

It doesn't work in shell script, but in javascript, it works fine, what's wrong with that?

I just want to filter out all IPs that are not in the range of 2~254, for example, ignore the 255

shell script

I want to filter out the IP 192.168.18.255 only using the standard regular expressions.

1. current useful input IPs: 192.168.18.195 192.168.18.255

2. wanted output IP: 192.168.18.195

#!/usr/bin/env bash

IPs=$(ifconfig | grep -oE '(192\.168\.1?[0-9][0-9]?|2[0-4][0-9]|25[0-5])\.(1?[0-9][0-9]?|2[0-4][0-9]|25[0-5])')
echo $IPs
# 192.168.18.195 192.168.18.255 ❌

IPs1=$(ifconfig | grep -oE '192\.168\.(1?[0-9][0-9]?|2[0-4][0-9]|25[0-5])\.([2-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-4])')
echo $IPs1
# 192.168.18.195 192.168.18.25 ❌

IPs2=$(ifconfig | grep -oE '192\.168\.(1?[0-9][0-9]?|2[0-4][0-9]|25[0-5])\.([2-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-4])$')
echo $IPs2
# ❌

# wanted, ignore 192.168.18.255 ❓
# IPs=$(ifconfig | grep -oE '❓')
# echo $IPs
# 192.168.18.195

This minimal reproducible version input just for testing.

en1: flags=8863<UP,BROADCAST,SMART,RUNNING,SIMPLEX,MULTICAST> mtu 1500
        options=6463<RXCSUM,TXCSUM,TSO4,TSO6,CHANNEL_IO,PARTIAL_CSUM,ZEROINVERT_CSUM>
        ether a4:83:e7:91:62:79 
        inet6 fe80::1ca2:3b0a:df9d:465f%en1 prefixlen 64 secured scopeid 0x7 
        inet 192.168.18.195 netmask 0xffffff00 broadcast 192.168.18.255
        inet6 fd80:eae6:1258:0:37:7544:1d1:7b08 prefixlen 64 autoconf secured 
        nd6 options=201<PERFORMNUD,DAD>
        media: autoselect
        status: active

$ ifconfig
# The output is too long, ignore it here, Please see the link below

The full version output of ifconfig link: https://gist.github.com/xgqfrms/a9e98b17835ddbffab07dde84bd3caa5

javascript

This is just used for test ignore the IP 255 works well.

function test(n) {
  let reg = /192\.168\.(1?[0-9][0-9]?|2[0-4][0-9]|25[0-5])\.([2-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-4])$/;
   for (let i = 0; i < n; i++) {
     let result = reg.test(`192.168.18.${i}`);
     if(result) {
       // console.log(`192.168.18.${i} ✅`, i, result)
     } else {
       console.log(`192.168.18.${i} ❌`, i, result)
     }
   }
}

test(256);

192.168.18.0 ❌ 0 false
192.168.18.1 ❌ 1 false
192.168.18.255 ❌ 255 false

      

refs

https://regexper.com/

https://regex101.com/

Answer 1

Since output of your ifconfig contains not of only IPs, but also of text surrounding it, $ is not the only valid marker of IP end.

Simplest way would be to replace it with \b: word boundary. But it will work only if you are using GNU grep.

If your grep lacks support of \b you can use ([[:space:]]|$) or [^0-9] instead. It will capture symbol following the IP if it is (whitespace or end-of-line) or non-digit correspondingly.

Example of command with output of your ifconfig stored in file input_file.txt:

$ ifconfig > input_file.txt

$ cat input_file.txt | grep -oE '192\.168\.(1?[0-9][0-9]?|2[0-4][0-9]|25[0-5])\.([2-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-4])([[:space:]]|$)'
192.168.18.195 

Demo at regex101.

Answer 2

solutions

test environment	result
macOS 13.1	✅
Raspberry Pi OS x64	✅

1. \b

It's very brief and clear. word boundary

#!/usr/bin/env bash

IPs=$(ifconfig | grep -oE '192\.168\.(1?[0-9][0-9]?|2[0-4][0-9]|25[0-5])\.([2-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-4])\b')
echo $IPs
# 192.168.18.195

2. [[:space:]]

This usage is not common, and not easy to understand.

It seems to be called character class keywords in GNU Linux.

#!/usr/bin/env bash

IPs=$(ifconfig | grep -oE '192\.168\.(1?[0-9][0-9]?|2[0-4][0-9]|25[0-5])\.([2-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-4])[[:space:]]')
echo $IPs
# 192.168.18.195

3. grep -w

It only works with grep, but is not compatible with standard regular expressions.

#!/usr/bin/env bash

IPs=$(ifconfig | grep -woE '192\.168\.(1?[0-9][0-9]?|2[0-4][0-9]|25[0-5])\.([2-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-4])')
echo $IPs
# 192.168.18.195

$ man grep

# ...

     -w, --word-regexp
             The expression is searched for as a word (as if surrounded by
             ‘[[:<:]]’ and ‘[[:>:]]’; see re_format(7)).  This option has no
             effect if -x is also specified.
# ... 

refs

https://en.wikipedia.org/wiki/Regular_expression#Character_classes

https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Regular_expressions/Assertions#:~:text=it%20in%20%22eat%22.-,%5Cb,-Matches%20a%20word

https://www.tutorialspoint.com/unix/unix-regular-expressions.htm