Matching nested groups with named backreference

Question

I need to use a multiline regex to extract part of a text, but I also have to allow those parts to be nested.

Here's an example of my scenario:

<!--BEGIN DELIMITER(A)-->
    sample text
    <!--BEGIN DELIMITER(B)-->
        another sample text
    <!--END DELIMITER(B)-->
<!--END DELIMITER(A)-->

The expected result would be to extract both blocks, including the begin/end lines as A1 and A2 are required for other things.

What happens is that I only get one match, which is the outer group.

Here's my regex:

<!--BEGIN DELIMITER\((?<expr>(\w+))\)-->(?<content>([\s\S]*?))<!--END DELIMITER\((\k<expr>.*)\)-->

Here's a regex101.com link for a quick visual reference.

Any idea on how to match both groups?

Like [this](https://regex101.com/r/4lep7o/1). Just wrap with a lookahead and a capturing group. — Wiktor Stribiżew, May 09 '17 at 15:52
That is not regular. Thus you can't use **Regular** Expressions to match this. It's effectively a different ["HTML syntax", as far as parsing goes](http://stackoverflow.com/a/1732454/1045510). — This company is turning evil., May 09 '17 at 15:52
You can however build a more sophisticated parser that uses RegEx just for the regular parts (detecting the start and end "tags"), and then build a tree out of it. Then you can select the contents however you want. — This company is turning evil., May 09 '17 at 15:55
@Kroltan, you could, i suppose. Remember that the RegEx engine of .NET supports balancing groups... — , May 09 '17 at 15:55
@elgonzo Yes, but it gets surprisingly messy surprisingly fast, and isn't perfect either. When in the future you want to extend the pattern, it's not going to be any fun. — This company is turning evil., May 09 '17 at 15:57
This link might help you : http://www.rexegg.com/regex-cookbook.html#captureparen2 — Gawil, May 09 '17 at 15:57

Matching nested groups with named backreference

0 Answers0