From: Robert Klemme <shortcutter@...>
Date: 2008-01-20T05:40:22+09:00
Subject: Re: Greedy and non greedy quantifiers

On 17.01.2008 12:17, Rados?aw Bu?at wrote:
> There are 4 (I hope that didn't miss anything) operators to specify repetitions.
> 
> *, +, ? and {m, n}
> * - matches 0 or more
> + - matches 1 or more
> ? - matches 0 or 1 (same as {0,1} )
> {m,n} matches m..n
> 
> All these operators all greedy (I'll show you examples). To make
> non-greedy operator you must add '?' after operator. So non-greedy
> operators are:
> *?, +?, ??, {m,n}?
> *? - matches 0 or more
> +? - matches 1 or more
> ?? - matches 0 or 1 (same as {0,1}?)
> {m,n} matches m..n
> 
> Maybe '??' looks strange, but all of these are absolutely correct.
> 
> So far no differences.
> 
> Why we have greedy and non-greedy operators? Because they works in
> different way. Greedy operators tries to match as many character
> (precisely: left expression) as can and if during matching it must
> move back it tries to match fewer characters.
> 
> For example:
> puts "aaaa".match(/a*/) # matches whole string
> puts "aaaa" =~ /a*?/ # matches no characters (but it success)

But can match more if it needs to to make the match a success:

irb(main):005:0> /a*?/.match("aaab").to_a
=> [""]
irb(main):006:0> /a*?b/.match("aaab").to_a
=> ["aaab"]

> Another example:
> puts "aaaabaaaab".match(/.*b/) # matches whole string
> puts "aaaabaaaab".match(/.*?b/) # matches "aaab"
> 
> I always recommend Jeffrey Friedl's book "Mastering Regular
> Expressions" (http://regex.info/). After reading this book you will be
> master of regexp ;-).

Definitively worth reading!

Cheers

	robert