[Gametes]

In the app I'm writing, the user supplies a web address, which's contents are then loaded into a string. What I'm trying to do is extract links from this string, that is:

<a href="http://www.google.com">Google</a>

becomes:

Google
http://www.google.com

The problem is -- as I didn't know before I adapted Apple's sample NSScanner code (which works fine for this format of link) -- is that there are about a dozen or more different kinds of links, as near as I can tell (that is, it almost never looks like the example above).
Yet every web browser displays links the same way, as a segment of underlined text linking to an address, and omniWeb even has a little utility window which extracts all the links from a page.

My question: before I struggle with reinventing the wheel to deal with each and every kind of link, is there any free html scanning code which I can pass a string and get this information out of?

Thanks. I've been looking on CocoaDev and CDC, with no avail.

[Rickster]

All the code we use to do that (the parsing, not the window displaying) is available as open source.

[Gametes]

Thanks Rick. Actually, I went to your dev site right after I wrote that. I've found OWF, and am currently looking through there for this code. You guys are such philanthropists; this stuff is amazing.
Question: What files are responsible for this activity? Is it even possible for me to extract pieces of the OWF and retain functionality?
They look to be intertwined.

There's little documentation in there, and I'm having trouble deciphering if the tokenizer files are what I want to use. Optimally, I'd like to instance some object that I can pass a webpage source string and get back an array of dictionaries, each with a Title string and an Address string.

I sincerely appreciate your illuminating the OWF's files for me.