HTML Nested Tag Replacement

Groove and I ran into an interesting question over on the PHP forum. There was a member who wanted to replace only CERTAIN span tags (with a particular class name) with <strong> tags, also replacing the corresponding closing span with a closing strong. When there are no nested tags at all, the application is a trivial regex. However, what about nested tags? What about a replaceable span with non-replaceable spans inside of it? What about nested replaceable spans?

Since this forum is pretty dead, I thought I'd posit the question here as well. Check out the various PHP solutions we've come up with. My very last post in that thread solves the problem for everything but a replaceable span inside another replaceable span.

What do you guys think? Is it possible in a regexp?

-Dan

I still plan on solving this damn thing.

Good luck to you! I'm still waiting for requinix to drop a regex bomb on us.

-Dan

Right? He'd be the one to do it, if it can be done. I know it's possible to write recursive regex (as well as look-forward and look-behind)... just don't have enough regex-fu to pull it off myself. Yet.

Hi, I'm new here, but I'll contribute my 2 cents.

In the traditional definition of regex, no, it's not possible. Regular expressions are supposed to be reducible to a finite state automata. The moment you add a register (in this case, a variable to remember what tag you're trying to match), you've cheated.

But that's all just boring textbook talk. Many regex engines let you "cheat". The .NET regex engine is my weapon of choice, and it has a built in construct called a "balancing group definition" that would accomplish exactly what you're trying to do: keep track of nested pairs (matching html tags).

I did a quick google search and didn't see an equivalent in PHP. But maybe you're better versed in PHP than I and can run with this.