Match Pattern In An Infinite Number Of Subpatterns

WeeDie

Hello, here's a problem anyone somewhat experienced with regular expressions probably would have the answer to.

How do you match all occurances of "variable<subpattern>variable" where the <subpattern> itself contains occurances of the full pattern.

Here's an example:

$pattern = "/(.?)/is";

<!-- BEGIN somename1 --><p>here's the subpattern content</p><!-- END somename1-->
<p>blah, blah, blah</p>
<!-- BEGIN somename2 --><p>some other text</p>
<!-- BEGIN somename3 --><p>the subpatterns subpattern</p><!-- END somename3 -->
<p>more text</p>
<!-- END somename2-->

with a preg_match_all that would match

match1.

<!-- BEGIN somename1 --><p>here's the subpattern content</p><!-- END somename1-->

match2.

<!-- BEGIN somename2 --><p>some other text</p>
<!-- BEGIN somename3 --><p>the subpatterns subpattern</p><!-- END somename3 -->
<p>more text</p>
<!-- END somename2-->

this matches the stuff between somename1 and somename2 but how do you get it to also match somename3 and (not illustrated by the example), if you wanted also somename4 inside the BEGIN and END tags of somename3, and inside of somename4 any occurances of END, BEGIN and subpattern of somename5, and so on, and so on.

Weedpacket

Theoretically it's impossible, as regular expressions aren't powerful enough to allow counting. In practice, there is a highly experimental and temperamental extension which I've never used myself so I can't comment on its utility in any way (and it may not even be configured in), in which one uses b[/b] with an expression to represent the entire expression.
http://www.php.net/manual/en/pcre.pattern.syntax.php#regexp.reference.recursive

dave420

Doing it in a single pattern is a pretty tall order, but it's very easy to do, nearly exactly as you said.

All you have to do, is keep matching the pattern on the result until it ceased to match. As long as you record each match along the way, you've successfully matched every pattern, including patterns-within-patterns.

I wrote a templating engine using a similar idea, and you're not limited on the arrangement of elements at all - variables within loops, loops within loops, includes within loops, loops within loops within includes within loops, etc. It all works fine.

Recursion is one of the most powerful tools you can master while programming. It allows you to write code that will run on almost any set of data, no matter the size or complexity. As long as you know the rules of the data, it's yours for the taking :-P

dave