Our "Fizz Buzz" interview question is to ask a candidate to validate an IPv4 address in Python. We have a 50/50 split of candidates using regex versus left to right scanning.

This is my solution in C which attempts to avoid requiring using a C library. There's a few constraints: IP addresses must be in dotted quad form, and each octet must not have leading zeroes. It also runs in constant time as it bounds each octet substring to at most 3 characters, only checking at most 4 octets.

make
./main

Valid IPv4 addresses:

  -- '0.0.0.0' is valid
  -- '127.0.0.1' is valid
  -- '127.127.0.1' is valid
  -- '127.127.127.127' is valid
  -- '1.0.0.1' is valid

Invalid IPv4 addresses:

  -- '(null)' is invalid
  -- '' is invalid
  -- 'abcd' is invalid
  -- '1.1' is invalid
  -- '1.1.1' is invalid
  -- '1.1.1.1.1' is invalid
  -- '1.1.1.1.' is invalid
  -- '01.0.0.1' is invalid
  -- '256.0.0.1' is invalid
  -- 'a.b.c.d' is invalid
  -- '1024.1.1.1' is invalid
  -- '-0.0.0.0' is invalid
  -- '+0.0.0.0' is invalid
  -- '0 .0.0.0' is invalid
  -- ' 0.0.0.0' is invalid

skeeto has a very compact and impressively readable version of an IPv4 parser that can be found here (only supports dotted quads).

lwIP has an implementation of an IPv4 parser that is very liberal and supports addresses that can usually be passed to the ping command, which is called ip4addr_aton.