sstrings: a strings(1) replacement

After reading that GNU binutils' strings(1) utility was affected by a buffer overflow vulnerability and finding out that a simple tool like that is more complex than it should be (at least in my opinion) I decided to brush up my C skills on UNIX-like systems and to write another implementation of strings(1). Enter sstrings(1).

sstrings(1) is a Simpler strings(1) (or Stupid strings) that does just what I suppose strings(1) should do: read chunks of binary data from a file or from stdin and output every sequence of 4 or more printable characters ending either with a '\0' or a '\n' to stdout.

I will follow the Open Group specifications as close as I can and you're free to nag me when something's wrong but I warn you: this tool is mostly written for my own amusement and it's not meant to be as good or as complete as GNU strings(1).

For istance it ignores the -a flag and ALWAYS scans the whole file. Since this behaviour is implementation dependent I think I'm allowed to do it.

I'm also not planning to make this utility universally portable, my aim is to have it run on a i386 personal computer running OpenBSD and then see if it runs on Linux too. Probably I will kindly ask you to test it on different UNIX-like operating systems, different architectures and different compilers and then send me some feedback about it but that's not mandatory.

The recommended compiler is GCC 4 and the following line should output no warnings:

gcc --std=c99 -Wall -Wextra -pedantic sstrings.c -o sstrings

The license is the MIT license and you can read it in LICENSE. It boils down to: do what you want with this as long as you give me credit for writing it in the first place and please don't sue me.