From: "Mark J. Reed" <markjreed@...>
Date: 2003-09-24T09:56:51+09:00
Subject: Re: regular expression

On Wed, Sep 24, 2003 at 09:34:12AM +0900, paul vudmaska wrote:
> I'm doing some html replacement using regular
> expressions and i ran into this issue...
> 
> Example reg example => /\[b\](.*)\[\/b\]/
> 
> ...works fine to match [b]bold please[/b] but NOT 
> 
> with [b]bold   
> 
>  please[/b]

That's because '.' doesn't match newline by default.  To get it to,
you need to add the /m flag.  (You also need to have more than the one
line in the string, of course).

However, * is greedy, which means that given this text:

    Here is some [b]bold text[/b], and here is [b]some more bold text[/b]

Your regex  will match all the way from the first [b] to the last [/b],
rather than just to the first [/b].  You can prevent that by using the
non-greedy qualifier ?, which goes after the *.  Also, I would use %r
instead of /.../ just to avoid one of those nasty backslashes. :)

So:

irb(main):001:0> x = "Here is some [b]bold
irb(main):002:0" text[/b], and here is [b]some more 
irb(main):003:0" bold text[/b]"
=> "Here is some [b]bold\ntext[/b], and here is [b]some more\nbold text[/b]"
irb(main):004:0> x =~ %r{\[b\](.*)\[/b\]}
=> nil
irb(main):005:0> x =~ %r{\[b\](.*)\[/b\]}m
=> 13
irb(main):006:0> x.gsub(%r{\[b\](.*)\[/b\]}m, '<b>\1</b>'/)
=> "Here is some <b>bold\ntext[/b], and here is [b]some more\nbold text</b>"
irb(main):007:0> x.gsub(%r{\[b\](.*?)\[/b\]}m, '<b>\1</b>'/)
=> "Here is some <b>bold\ntext</b>, and here is <b>some more\nbold text</b>"

-Mark