From: Daniel DeLorme <dan-ml@...42.com>
Date: 2007-06-20T08:22:51+09:00
Subject: Re: Check for text file

Nobuyoshi Nakada wrote:
> You can use String#count:
> 
>   def File.binary?(path)
>     s = read(path, 4096) and
>     !s.empty? and
>     (/\0/n =~ s or s.count("\t\n -~").to_f/s.size<=0.7)
>   end
> 
> In any case, it doesn't work for non-ascii files.

Pedantic correction: it desn't work for non-western scripts. French uses 
accents here and there but it would pass the test above.

Still, I have to say I was surprised; I didn't know that a hyphen in 
String#count had the same effect as in a regexp character class. Talk 
about an undocumented feature!

Daniel