real world extended regex examples

[szabgab]

In another post palindrome using regular expressions I asked about another ways to write catch a palindrome but my real interest is seeing examples of real world usage of extended regular expression. So if you have examples where you used any of the following constructs, I would appreciate seeing them.

(?=pattern) - zero-width positive look-ahead
(?!pattern) - zero-width negative look-ahead
(?<=pattern) - zero-width positive look-behind
(?<!pattern) - zero-width negative look-behind
(?{ code }) - zero-width assertion with embedded code
(??{ code }) - postponed subexpression
(?>pattern) - "independent" subexpression
(?(condition)yes-pattern|no-pattern) - conditional expression
(?(condition)yes-pattern)

[GrandFather]

Super Search is your friend (as is oft said). A SoPW only search for nodes containing the words "regex zero width assertion" gets you a pile of examples. Just picking a few from the first page:

• Re: A Quick Regex Question
• Re: substituting a regular expression in a file
• Re: REGEX Help
• Re: reg ex NOT
• Re^2: Regex negative word
• Re: Failed regex: negative look-behinds
• Backreferences in negated character classes

Similar searches for related key words would get a pile more.

[Fletch]

Install Regexp::Common and snoop around in its guts. There's a few of these therein.

[Eimi Metamorphoumai]

You might look at the rules in SpamAssassin, and the custom rules at, say http://www.rulesemporium.com/. Lots of complex patterns, particuarly the (?:posi|nega)tive look-(?:ahea|behin)d.

As a simple example, consider something like /\bviagra\b/i. Now assume you want to match also "\/i@gr@" (very simple obfuscation). If you do /\b(?:\\\/|v)i[@a]gr[@a]\b/i, it doesn't actually work for what you want it to. That's because \b insists on a word boundery, and since \ and @ aren't word characters, they'll only match if adjacent to non-word characters. So what you really want is /(?<!\w)(?:\\\/|v)i[@a]gr[@a](?!\w)/i.

[johngg]

I really like this neat combination of positive looks behind and ahead by davidrw. It immediately sprang to mind when I saw your post but it might not come up in a search as no mention is made in it of extended methods.

Cheers,

JohnGG

[blazar]

On a slightly less serious basis than the other answers you got... the way I explain (how not to misspell) my nick.

[planetscape]

The newly minted Google Code Search might be a start.

The Regular Expression Library is fully searchable, IIRC.

Super Search will doubtless turn up many threads, possibly including The craziest RegExes you ever created, which I'm betting has some interesting examples and links.

HTH,