From: Chai Muang thong <thai4pental@...>
Date: 2008-12-02T23:35:48+09:00
Subject: Re: regexp problem

Constantine Karnacevych wrote:
> Brian Candler wrote:
> 
> def match_pairs s, l, r
>     d, o, e = 0, [],  '(' + Regexp.escape(l) + '|' + Regexp.escape(r) + 
> ')'
>     while m = s =~ /#{e}/ do
>         $1 == l ? (o << "" if 1 === d += 1) : d -= 1
>         o[o.length - 1] += o.last.empty? ? $1 : $` + $1
>         s = $'
>     end
>     o
> end
> 
> str = "<<p1 <<p2>> <<p3 <<p4>>>>>> <and> <<p5>> or <<p6 <<p7 <<p8>>>>>>"
> p match_pairs str, "<<", ">>"
> 
> => ["<<p1 <<p2>> <<p3 <<p4>>>>>>", "<<p5>>", "<<p6 <<p7 <<p8>>>>>>"]
> 
> p match_pairs str, "<", ">"
> => ["<<p1 <<p2>> <<p3 <<p4>>>>>>", "<and>", "<<p5>>", "<<p6 <<p7 
> <<p8>>>>>>"]
> 
> str = '{p1 {p2} {p3 {p4}}} <and> {p5} or {p6 {p7 {p8}}}'
> p match_pairs str, '{', '}'
> 
> => ["{p1 {p2} {p3 {p4}}}", "{p5}", "{p6 {p7 {p8}}}"]

--------------

Thank you very much Brian. It works fine as long as an incoming string
is valid. In other words, if the string has extra pair to the left, it
will return wrong result. For example a string "Test <tag> or <<not>
correct tag" will return ["<<"] instead of [""]. Anyway, I have solved
the problem by implementing a method to handle this instead of using
only regexp. Thank you again.

Chai
-- 
Posted via http://www.ruby-forum.com/.