Regular Expression (PHP/Perl) Backreferences

Question

I want to match strings of the form:

Sections 158, 417, 418, 500, 501, and 1111

Where I get all the numbers as backreferences (replacing them with hyperlinks). I got this far:

$text = preg_replace_callback("/Sections? ([0-9.]+)(?:(?:and|,| |or)+([0-9.]+))+/","hyperlink_me", $text); 

That's PHP (which uses perl-compliant regexes). The PHP part is fine I think, it's the regex that's not doing what I want, but I'm giving the whole line of code for context.

The problem is the only two backreferences I get (other than the whole string) are for the first number ("158") and the last one ("1111"). It seems like the second captured backreference ("([0-9.]+") is just getting written over from the second number onwards. From this page , at "Repetition and Backreferences", I get the idea that this is often an issue, but can't figure out how to solve it in this context. Any regex geniuses out there who can help?

Answer 1

That is simply how capturing-parentheses work, and can't be avoided. But you can create a "helper" callback, hyperlink_us , that does this:

$output = preg_replace_callback("/[^0-9.]+([0-9.]+)/","hyperlink_me", $input);

and then you use it like this:

$text = preg_replace_callback("/(Sections? [0-9.]+(?:(?:and|,| |or)+[0-9.]+)+)/","hyperlink_us", $text);

That way, hyperlink_us passes all the numbers to hyperlink_me , but the caller of hyperlink_us can make sure that numbers are only passed in the appropriate contexts (the "Section(s)..." stuff).

(Disclaimer: I'm not much of a PHP-er. I'm assuming that this function behaves like its analogue in JavaScript. I do know Perl, but it doesn't have a callback-based regex-replacement function.)