From: James Coglan <jcoglan@...>
Date: 2009-03-28T19:22:06+09:00
Subject: Re: Regular Expression interesting problem

--001636c5a81c6bb62504662b3806
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

2009/3/28 Arun Kumar <arunkumar@innovaturelabs.com>

> Eric Hodel wrote:
> > On Mar 28, 2009, at 02:07, Arun Kumar wrote:
> >> />
> >> interesting problem, but i need help for solving it. Please help me.
> > I suggest you use Nokogiri.
> >
> > Barring that, don't use regular expressions, use something more
> > appropriate like StringScanner from strscan.rb.  `ri StringScanner`
> > will get you started.
> >
> Nokogiri is a good option. But i want to use net/http for my assignment
> and it is compulsory.
>
> > PS: You'll probably want to do something like scan for <, then scan
> > for a tag name, then scan for attributes, then scan for >, etc.
>
> As you said I have to check the tag ie'<link' first and then check for
> the attributes. But still the position of the type attribute is the
> problem.



Looks like you will need to parse in stages -- I can't get String#scan to
capture everything using a single regex, though there's every chance I've
screwed up the expression somehow:

'<link type="application" href="http://google.com" rel="alternate" />'.scan
/<([^\s]+)(?:\s+([^\s]+)="([^"]*)")*\s*\/?>/i
#=> [['link', 'rel', 'alternate']]

--001636c5a81c6bb62504662b3806--