On Wed, May 7, 2008 at 9:35 AM, Louis Ryan <[EMAIL PROTECTED]> wrote: > The parsing would need to be pretty simple for performance reasons, so > probably not full HTML parsing with funk tolerance. More likely simple > searches for > src="<url>" and href="<url> in easily recognizable tag patterns and similar > for CSS and JSON structures. If people have suggestions on the > implementation details Id love to hear them.
Doing this server-side is error prone, so plan on failing gracefully. What about client-side solutions? Traversing document.links might work, or walking the DOM.
