From: Robert Klemme <shortcutter@...>
Date: 2006-10-28T17:15:07+09:00
Subject: Re: regex blindness

Josselin wrote:
> I was blind yesterday .. could not find what's wrong in this regex (and 
> posted that to mac os forum)
> 
> /^(?k: ( ?k:0[1-9]|[1-8][0-9]|9[0-8] )  (?k: [0-9]{3} )  )$/
> 
> trying to check zipcodes (french cities :  like 78231)
> 
> got an undefined (?...) sequence: 
> /^(?k:(?k:0[1-9]|[1-8][0-9]|9[0-8])(?k:[0-9]{3}))$/

irb(main):001:0> /(?:x)/
=> /(?:x)/
irb(main):002:0> /(?k:x)/
SyntaxError: compile error
(irb):2: undefined (?...) sequence: /(?k:x)/
         from (irb):2
         from :0
irb(main):003:0> /(?i:x)/
=> /(?i:x)/

Ruby does not allow the "k" flag there.  And I believe Ruby does not 
have that RX flag at all.

> I am also checking email addresses with :
> /^([a-zA-Z0-9_\.\-])+\@(([a-zA-Z0-9\.\-])+\.)+([a-zA-Z0-9]{2,4})+$/
> 
> is there a better regexp for that ...

I guess you'll find out plenty out there.  The answer to "better" 
heavily depends on what your goals are (simpler RX, faster matching), 
what RX engine you are using and what data you are processing.  Example: 
in some cases /[^@\s]+@[^@\s]+/ might be perfectly ok to match an email 
address - in other cases it won't.

Kind regards

	robert