From: David Masover <ninja@...>
Date: 2009-12-14T19:23:16+09:00
Subject: Re: problem with trivial regular expression

On Monday 14 December 2009 04:06:08 am David Villa wrote:
> fajljsfjaosfohttp://www.marca.comjafosjodfahttp://www.as.comjfoaasjofja

Any particular context? Or is it actually that random?

> i want to extract the diferents url  but i try with :
> /(http\:\/\/.+com)/ it returns a long match:
> 
>    1. http://www.marca.comjafosjodfahttp://www.as.com

If you think about it, that is still a valid URL. You're trying to limit it 
not to URLs, but only to http:// followed by a domain, and then only a domain 
ending in .com -- there are MANY urls that this will break.

If you're OK with that, the basic problem is that . is going to match as much 
as it possibly can (greedy), and it matches any character. The simple solution 
is to make it match as few characters as it can (miserly). You do that by 
putting a question mark after the + or *:

/(http\:\/\/.+?com)/

But again, that's not matching .com, that's matching anything ending in com. 
For example, on this URL:

http://www.broadcom.com/

it will only capture http://www.broadcom. So there's an easy solution -- add 
an escaped dot:

/(http\:\/\/.+?\.com)/

That's as much as I want to do with it. I'm guessing what you're trying to do 
is auto-linkify URLs in forum posts, or something like that -- some problem 
that's been solved a million times before, and better, so you should look for 
those solutions. But I won't assume that applies to you...

By the way, if you don't already know:

http://rubular.com/